Super Micro said it is cooperating with the investigation, suspended Liaw and placed another employee on administrative leave while terminating a contractor, and stressed that the conduct described in the indictment "contravenes the company's policies and compliance controls". The stock fell as much as 27 percent on the news, adding to a long list of reputational shocks that have hit the company since its accounting controversy in 2024, and raises fresh questions for investors about whether the compliance and governance weaknesses that allowed this alleged scheme to reach 2.5 billion dollars in diverted sales are isolated individual failures or something more structural.

What exactly does the US government claim

The indictment accuses Liaw, as well as Ruei-Tsang "Steven" Chang and Ting-Wei "Willy" Sun, of conspiring to smuggle U.S. AI servers into China without the required licenses from the Commerce Department. Super Micro is one of the key manufacturers of servers equipped with Nvidia chips, which are prohibited from being exported to China without government approval for national security reasons.

According to the indictment, the defendants used a combination of:

• false documents claiming the servers were destined for other customers.

• fake "dummy" servers to deceive inspectors during inspections.

• complex transshipment routes through third countries to conceal the true destination of shipments.

The DOJ reports that through this scheme, Super Micro generated at least $2.5 billion in revenue between 2024 and 2025.

Super Micro $SMCIResponse

Super Micro issued a statement emphasizing that the company itself is not a defendant in the indictment and that it is cooperating with authorities. Two employees named in the case have been suspended, and the contractor has been fired.

The company claims that the actions of these individuals violated company policies and compliance systems. However, this is a phrase that investors hear repeatedly with Super Micro - the company also went through a forced suspension of results over accounting irregularities and the threat of delisting from the stock market in recent years before it stabilised its situation.

Why Super Micro is in the spotlight again

Super Micro is one of the largest AI server manufacturers in the United States, and its products power the infrastructure of large hyperscalers and AI companies. It's this role that makes it an attractive target for those looking to bypass US export controls, as its servers equipped with Nvidia chips are among the most powerful platforms available for training and running AI models.

The case fits into a broader trend: the US government has stepped up its fight against illegal AI technology exports to China, and indictments for smuggling Nvidia GPUs are now relatively common. But so far, most of the cases have involved outside players - resellers, shippers and brokers. The Super Micro $SMCI case is more sensitive because the accused are people with direct ties to the manufacturer itself, and one of them is even a co-founder of the company.

Impact on Nvidia and the AI server ecosystem

While Nvidia $NVDA is not charged in the case and is cooperating with the authorities, the case inevitably casts a shadow on it as well. Super Micro is one of its key partners and distribution channels for GPUs into AI servers, and any tightening of controls or shutting Super Micro out of part of the market will affect Nvidia's chip flows.

At the same time, Nvidia itself is under pressure over export controls: CEO Jensen Huang has repeatedly criticized the restrictive policy as economically damaging, estimating that in a single quarter Nvidia lost approximately $2.5 billion in H20 revenue due to the ban on advanced chip sales to China. While the Super Micro case shows that chips are making their way into China despite the controls, it is also a reminder to Nvidia that allegations of "leakage" of its GPUs through third parties may increase political pressure for even tighter regulation of its distribution channels.

Impact on stocks and investment risks

The immediate market reaction - a 14.6% drop in Super Micro - reflects a combination of concerns:

• Legal risk: even if the company as a whole is not charged, investigations may bring additional costs, penalties or restrictions.

• Reputational risk: For a company that has only recently settled an accounting case, each new scandal is a major blow to customer and partner confidence.

• Operational risk: temporary staff suspensions and possible widespread investigations can disrupt business operations.

• Regulatory risk: the US Department of Commerce may tighten the conditions under which Super Micro operates or exports.

Super Micro is in a delicate position: it is a key link in the AI infrastructure supply chain, so any restriction on its operations would affect customers who rely on its servers. If the investigation widens or the company loses certifications and licenses, it could also affect the performance of existing contracts.

Wider context: AI chip smuggling as a growing issue

The Super Micro case is not unique in the US justice space. In the past year, multiple groups have been charged or arrested for illegally exporting Nvidia GPUs to China:

• Operation Gatekeeper in late 2025 uncovered a network of AI chip smugglers and led to the seizure of more than $50 million worth of GPUs.

• In November 2025, the DOJ charged three Chinese nationals with smuggling Nvidia chips through Malaysia and Thailand.

• In August 2025, two Chinese nationals were arrested in Los Angeles for exporting Nvidia GPUs through a network of transshipment companies.

The pattern is similar in all cases: false documents, third-country brokerage firms, and efforts to conceal the final destination of the shipment in China. This suggests that demand for US AI chips in China is so strong that it motivates systematic and well-organized attempts to circumvent controls, and that the US government has intensified its monitoring of these flows.

What to watch next

For investors watching both Super Micro (SMCI) and Nvidia (NVDA), the following questions are key:

• Whether the Justice Department will extend the indictment to the company itself or other employees.

• How Super Micro will operationally manage the loss of two key employees and the potential reputational impact on business relationships.

• whether the case will prompt U.S. authorities to tighten export licensing terms for AI server makers.

• what signal it sends to the entire AI server sector regarding compliance and supply chain controls.

Super Micro has been through several crises over the past two years and has managed to bounce back each time due to strong demand for AI infrastructure. This time, however, the challenge is qualitatively different: it's not just about accounting, but about allegations that people running the company knowingly orchestrated the circumvention of US export laws, with proceeds from the scheme exceeding $2.5 billion. How much this affects the trust of partners, customers and regulators will be a crucial question for the future of Super Micro.

Even after rate normalisation, it has had several years of exceptional results, double-digit EPS, very high margins and is now starting to return capital to shareholders in the form of dividends and other payouts without sacrificing the safety of its balance sheet or the potential for further growth. For the investor looking for a combination of undervaluation, a strong balance sheet and exposure to the mid-size tanker cycle, this is a name that may pleasantly surprise in the years ahead if the tight market holds longer than the market is pricing in today.

Top points of analysis

• The firm's revenues have fallen from a peak of US$1.47bn in 2023 to around US$0.95bn in 2025, reflecting the normalisation of tanker rates - yet net margins remain around 37% and the firm generates strong cash flow even in a 'worse' year.

• The firm trades at a valuation of around 6 times trailing P/E and an EV/EBITDA below 4, with the data suggesting a current fair value of $78.92 per share - the stock is therefore roughly 24% undervalued relative to fundamentals.

• The balance sheet is one of the strongest in the sector: debt-to-equity of just 0.03, debt-to-assets of 0.02 and an Altman Z-score of over 8.7 - the company has virtually no net debt and is sitting on a record cash position, with enterprise value well below market cap.

• The orderbook of mid-size tankers is below 6% of the existing fleet near historic lows, roughly a quarter of the fleet will reach 20 years of age by 2026, and the yards have capacity filled 90% by vessel types other than tankers - creating a structurally tight market for the next few years.

• The dividend yield is around 3% with a very low payout ratio, and the company has room to supplement the base dividend with special payouts or buybacks in good years due to its strong FCF and net balance sheet.

• The firm belongs to the Suezmax and Aframax/LR2 tanker segment, where it has achieved the best spot rates in the last 15 years - averaging USD 26.5k/day for Suezmaxes and USD 38.8k/day for Aframaxes - illustrating the strength of the business strategy and the tightness of the market.

Company performance

Teekay Tankers $TNK was formed in 2007 as a spun-off tanker "pure play" from the broader Teekay Corporation group, which since the 1970s has evolved from a regional shipper to a global carrier of crude oil, LNG and other energy commodities. The aim was to give investors direct exposure to the volatile but potentially highly profitable market for the ocean transportation of crude oil and products without mixing with the group's less cyclical segments. Today, TNK is based in Hamilton, Bermuda, is listed on the NYSE and is a major player in the mid-sized tanker segment.

In terms of fleet, TNK's main focus is on Suezmax and Aframax/LR2 tankers, ships of around 80-160k DWT that are more flexible than giant VLCCs and serve a wide range of routes between key export regions and customers. These vessels carry crude oil and refined products between the US, Latin America, West Africa, Russia, the Middle East, Europe and Asia, and changes in geopolitics and sanctions in recent years have significantly lengthened some routes and increased tonne-miles. Additionally, TNK operates ship-to-ship transfer services, particularly in the U.S. Gulf and Caribbean, which generate fee income and increase fleet utilization.

https://www.youtube.com/embed/4MszJuCJAyw?rel=1

What this business really profits from is primarily the spot and short-term charter market. Teekay Tankers keeps much of its fleet "open" to spot and short contracts, allowing it to quickly capitalize on periods of high rates and regularly report top mid-size spot rates.

In terms of its financial profile, TNK has changed significantly in recent years: it has transformed from a company that carried higher debt in earlier cycles to one with minimal debt, a high equity ratio and record cash. Ratios such as debt-to-assets of around 0.02, debt-to-equity of around 0.03 and Altman Z-score of over 8 indicate an exceptionally strong balance sheet, especially for a sector as cyclical as tankers. Management is leveraging this strength for disciplined capital allocation: instead of aggressively ordering new ships through the cycle, TNK is selectively upgrading the fleet, reducing structural risk while opening up room for shareholder payouts.

The market and the cycle in which TNK plays

The global tanker market is defined by a balance between fleet size and the demand for oil and product transportation in tonne-miles. In recent years, these parameters have swung in a direction that favours tanker companies: demand for oil carriage has stabilised after covide, geopolitics has lengthened some routes, while at the same time the orderbook of new ships has remained very low. Teekay has repeatedly stressed in his market updates that mid-size tanker market fundamentals remain positive, although short-term rate fluctuations are inevitable.

For the mid-size segment, which is TNK's focus, it is key that the orderbook is at a level below 6% of the existing fleet. This means that there will not be a large amount of new capacity entering the market in the next few years, even if companies start ordering new ships today. At the same time, around 11% of the mid-size fleet is over 20 years old and another 14% will reach that age between 2024 and 2026, meaning that almost a quarter of the fleet will soon face either scrapping or costly upgrades due to emissions and safety requirements.

The situation in shipyards is also an important detail. According to data presented by TNK, the global forward cover of shipyards is around 3.5 years, with around 90% of the vessels ordered being non-tanker types. Shipyards thus have orders for containerships, LNG and other segments and do not have unlimited capacity to benefit from any late "rush" for tankers. The combination of a minimal orderbook, an aging fleet and busy yards creates an environment in which even a relatively moderate rate of growth in tonne-mile demand can keep rates above the long-term average for a longer period than is typical of past cyclical periods.

Growth potential - where can it come from and how big can it be

The growth potential for TNK is not in being a classic growth company, but in a combination of three things: a low valuation, a strong market and room for capital returns.

The first source of growth is valuation alone. With earnings per share of around $10 in 2025 and a price that implies roughly 6-7 times trailing P/E, TNK is valued lower than many other cyclical companies with similar or worse balance sheets. If earnings settle in the range of, say, $6-8 per share after rate normalization, and the market is willing to pay 8-10 times earnings for a more stable business, that implies room for tens of percent share price appreciation versus today, purely from a re-pricing multiple.

The second source of growth is the extended mid-size tanker cycle. A low orderbook of under 6% of the fleet and aging ships (roughly a quarter of the fleet in the mid-size segment will reach 20 years of age by 2026) suggest that the favorable environment for rates may last longer than just one or two seasons. Even if rates fall from extreme peaks, they may still remain above levels that correspond to the "old normal", which would allow TNK to hold margins higher than the market currently models.

The third source is capital allocation. A strong balance sheet and cash position give management the ability to combine several actions: continue to renew the fleet at attractive prices, return capital to shareholders through dividends and buybacks, and possibly take advantage of weakening competitors to drive acquisitive growth. If these steps can be taken during a period when valuations are still low, TNK can increase value per share not only through earnings, but also through NAV per share growth and share count reduction.

Management and CEO

The current management of Teekay Tankers is led by Kenneth Hvid, a long-time manager of the entire Teekay group. He has more than two decades in various roles at Teekay - from managing specific segments to the entire group - and thus has experience of several complete tanker cycles, including the high rate period before the 2008 crisis, the subsequent downturn and the current super-cycle.

His approach manifests itself mainly in three areas. The first is debt and risk. After previous years when parts of the group were more leveraged, TNK has moved under his leadership to a structurally low debt position, with debt-to-asset and debt-to-equity ratios well below the sector average. The second area is disciplined capex - the company has not used the recent strong years to aggressively order large numbers of new ships, but rather to selectively renew and modernise its fleet. The third area is the flexibility of capital returns: the dividend was only reinstated when the balance sheet could bear it, and is now set to be extended in the future with special payouts or buybacks if results are good and valuations are low.

Importantly for the investor, this is management that has seen both sides of the cycle and seems to prefer long-term capital return over short-term "chasing" fleet size or dividend yield. In an environment where it is easy to succumb to the temptation to go into debt to order new ships, this is a plus rather than a minus.

The market in which TNK operates - outlook and expectations

Teekay Tankers operates in the global marine transportation market for crude oil and products, with its centre of gravity in the midstream segment. Demand for these services is driven by the volume of crude oil transported, the length of routes and the structure of trade flows. Sanctions on Russian crude oil, growth in exports from the US and Middle East and changes in refining capacity around the world have led to the extension of a number of routes in recent years, thereby increasing tonne-miles.

The medium-term outlook for tanker shipping demand is rather positive, although not linear, according to Teekay and independent analysis. Global oil demand is expected to hover close to current levels in the coming years, possibly rising slightly depending on economic developments and energy policy. At the same time, it will not be easy to scale up the fleet quickly with new ships, as shipyards have contracted production for several years ahead and most slots are occupied by other vessel types.

TNK itself has repeatedly stated in its market updates that the fundamentals for the mid-sized tanker market remain positive. Management anticipates that while rates will remain volatile between quarters, the low orderbook and aging fleet will support solid average rates over the next few years. In addition, in Q4 2025 Teekay indicated that it expects rates to remain strong in Q1 2026 due to seasonal factors and continued tonne-mile demand, which combined with the state of the fleet suggests that it does not view recent strong quarters as a one-off anomaly.

What may surprise the company in the future

The scope for pleasant surprises in the years ahead lies mainly in what TNK can do beyond the baseline scenario.

• It can accelerate capital returns if it maintains high free cash flow and low debt.

  • That could mean a more stable/higher ordinary dividend, more frequent special dividends, or more aggressive buybacks in periods when the P/E remains low.

• Can profitably renew and optimize the fleet.

  • By selling older ships at times when the value of second-hand vessels is elevated, and buying younger units when opportunities arise, TNK can increase value per share even without growth in the number of ships.

• It can take advantage of any weakening of competition.

  • If some of the less disciplined players come under pressure in the next cycle due to debt, Teekay's clean balance sheet gives it the opportunity to buy quality ships or entire portfolios in forced sales.

• It can benefit from the fact that regulation and ESG pressures will accelerate the phasing out of old ships faster than the market now expects.

  • If older units are scrapped earlier due to emissions standards, supply tensions may intensify, which would support rates and profitability for longer.

Results and figures

On a results level, Teekay Tankers has had a very strong couple of years. It has sales of around $1.47 billion in 2023, around $1.23 billion in 2024 and around $0.95 billion in 2025, reflecting the normalisation of rates after an extremely strong period, but still levels well above historical averages. Gross margins are around 27-28%, operating margins around 27% and net margins for 2025 around 37%, which is very attractive for a capital intensive, cyclical business.

Net income for 2025 was about $351 million, compared to $404 million in 2024 and $520 million in 2023. EPS for 2025 is around $10 per share, after two previous years of $11-15, with a relatively stable share count of about 34.5 million. This, coupled with very low debt, leads to a valuation of around 6 times trailing earnings, which remains conservative even with expected profitability normalization.

The cash flow picture is similarly strong. TNK has generated robust operating cash flow in recent years, which, combined with relatively moderate capex, has led to high free cash flow and double-digit FCF growth over time. This has allowed it to virtually eliminate net debt, fund its fleet renewal program, and at the same time begin regular payouts to shareholders. In Q4 2025, the company reported a record cash position and emphasized that its capital strategy will continue to combine prudent fleet investment with disciplined capital returns to shareholders as valuations and market conditions permit.

Dividend and capital allocation

Dividend at TNK is an add-on to the cyclical story, not its main axis. After a period when the priority was clearly on reducing debt and stabilising the balance sheet, the company has restored the dividend in recent years and gradually increased it. It currently pays about $0.25 per share per year, which at the current price represents a dividend yield of about 3%.

The payout ratio is very low given the high EPS, so the base dividend represents only a small portion of earnings, leaving the firm with wide leeway. Moreover, Teekay is not dogmatically focused on just regular quarterly payouts - in periods of exceptionally strong performance, management has the option to supplement the base dividend with special payouts.

Capital allocation is guided by several priorities: maintaining a very strong balance sheet, prioritizing returns over growth at all costs, continuously rejuvenating the fleet, and only then increasing direct payouts to shareholders. This means that in a period of high rates, an investor need not expect an extreme dividend yield of 10-15%, but rather a combination of a reasonable base dividend, potential special payouts and growth in the value of the company through an improving fleet and balance sheet. In a weaker market, TNK is then likely to take the dividend flexibly and possibly adjust it to protect the balance sheet - but the main difference from the past is that today's balance sheet gives it much more room to manoeuvre.

Balance sheet and valuation

Teekay Tankers' balance sheet is one of the strongest in the sector. The debt to assets ratio is around 2%, the debt to equity ratio is only 0.03 and the Altman Z-score of around 8.7 is well above the point where an investor would need to worry about solvency or restructuring. Moreover, the company holds a large amount of cash, so its enterprise value is well below its market capitalization, which in practice means that shareholders have both fleet value and a liquidity "cushion".

From a valuation perspective, the stock trades at roughly 6-7 times trailing P/E, with EV/EBITDA around 3-4 times and P/B near 1.1. Analysts in the "deep value" community point out that with this combination of balance sheet and profitability, it is still an undervalued title, even though some of the good news is already in the price. The analyst consensus sees room for further upside in TNK's price target, though it warns of cyclical risks and the possibility of earnings normalizing to lower levels in the coming years.

The key argument for undervaluation is therefore the difference between how the market values current profits (as highly cyclical and temporary) and how long the fundamentals suggest the favourable conditions of an ageing fleet and low orderbook can last. If it turns out that the "new normal" of rates is higher than in the past, and that TNK retains some of today's margins longer than the market expects, the stock has the potential to combine rerating multiples with continued high cash flow generation.

Investment Scenarios

In a positive scenario, the tanker market for the mid-size segment will remain tight for most of the next decade. The orderbook will remain constrained, an aging fleet will accelerate the scrapping of old ships, tonne-miles will be supported by geopolitical factors and global oil demand will remain close to current levels. In such an environment, Teekay Tankers will maintain a portion of today's rates, continue to generate high FCF, upgrade the fleet and increase payouts to shareholders; the stock has room for a combination of solid dividend yield and significant capital appreciation.

In the base case, rates gradually normalize to levels that are below today's levels but still above the long-term average as the orderbook and aging fleet keep the market in a relatively healthy balance. TNK's EPS will fall from today's double digits to mid-single to low double digits, but the company will retain flexibility, a reasonable dividend, and the ability to continue fleet renewal due to its strong balance sheet. In such a world, the stock could move from "deep value" status to a valuation closer to mid-cycle multiples without an investor having to rely on extreme scenarios.

A downside scenario would combine a steeper decline in oil demand, changes in export flows that shorten routes and reduce tonne-miles, and a potential freeing up of yard capacity leading to a wave of new tanker orders. Rates could fall near or below levels that cover fully loaded cash breakeven, EPS would thin significantly, and the company would have to realign the dividend and pace of investment. A strong balance sheet should allow it to weather such a period in better shape than weaker competitors, but for shareholders it would be a classic cyclical "drawdown" where undervaluation could persist for some time.

What to take away from the article

• Teekay Tankers is a pure play on the mid-size tanker cycle, not a dividend title or a defensive position - investors need to be prepared for cyclical volatility as earnings will move significantly over time as rates move.

• The company is entering the rate normalization phase from an exceptionally strong position: near-zero debt, record cash, an Altman Z-score of over 8, and an enterprise value below market cap are a rare combination for the tanker sector, giving the investor a safety margin that is not the norm for cyclical stocks.

• Valuations around 6 times trailing P/E and EV/EBITDA below 4 may look like a trap as the market factors in rapid earnings normalisation - but the key question is whether an orderbook below 6%, an ageing fleet and busy yards will keep rates above the "old normal" for longer than consensus models.

• According to the data, the stock is currently undervalued by about 24% against a fair value of $78.92, and this discount is not due to poor business quality but cyclical pessimism - a material difference for a value investor with a longer horizon.

• Management, led by Kenneth Hvid, has so far shown that it prefers capital discipline to expansion at any cost - that's a plus, but it's also true that the main "test" will come when rates actually fall and the company has to demonstrate its ability to counter-cyclically allocate capital.

• For an investor who understands the tanker cycle and is willing to follow the evolution of rates, orderbook and capital allocation as closely as quarterly earnings, TNK may be an interesting medium-term bet on a combination of re-rating multiples and continued FCF - but for an investor looking for stable passive income or a defensive position, it will be too cyclical a profile.

Personally, I find this sector quite attractive, but I haven't invested in it yet and I'm looking for a suitable stock. So far I'm most interested in $RKLB and $ASTS. SpaceX could also be interesting once it goes public.

I don't understand this sector very well yet, so I'd appreciate your opinions, tips, or warnings about any potential risks.

The energy market in 2026 is going through one of the most turbulent periods since the oil shock of the 1970s. Coordinated US and Israeli military strikes on Iranian infrastructure in late February triggered an avalanche of events that completely redrew the global oil market map. The closure of the Strait of Hormuz, through which around 20% of the world's oil production normally flows, drove prices to levels that analysts were talking about as an unlikely extreme scenario just six weeks ago.

According to the latest data from the International Energy Agency, Brent futures reached over $114 per barrel in the short term during March 2026, while the EIA revised its March outlook for the average Brent price for the full year 2026 to $78.84 per barrel, up from its original estimate of $57.69. At the end of the trading session on 19 March, Brent was trading around USD 108 and WTI near USD 96 per barrel.

The war premium, which Goldman Sachs $GS estimates at around $18 per barrel, remains firmly anchored in prices and will most likely be present throughout the first half of 2026.

This macroeconomic context creates an exceptional environment for oil producers and refiners outside the Gulf. US energy companies, which are benefiting from higher oil prices without direct exposure to the logistical turmoil in the Strait of Hormuz, are now enjoying increased investor interest.

Occidental Petroleum $OXY

Company profile and market position

Occidental Petroleum is one of the largest oil and gas producers in the United States with a strong concentration in the Permian Basin, the most productive oil basin in the world. At the same time, the company is building a unique strategy combining traditional production with Direct Air Capture (DAC) carbon capture and storage technologies, making it a player with an asymmetric profile towards different energy transition scenarios. Revenues for fiscal 2025 reached $21.6 billion.

The most significant move of 2025 was the sale of OxyChem's chemical division, completed on January 2, 2026, which immediately reduced the company's debt by $5.8 billion. This brought total debt down to $15 billion, down from nearly $40 billion in 2019 after the Anadarko acquisition. This structural shift significantly improved the company's financial resilience and flexibility just as the energy market entered a phase of extreme volatility.

Key financial indicators

Breakeven oil price - the minimum price per barrel of oil where the company is covering all of its costs

Berkshire Hathaway and the geopolitical premium

The presence of Berkshire Hathaway as the largest shareholder remains a key stabilizing factor for investors. Warren Buffett has repeatedly reaffirmed this position, buying up shares of $OXY on every major downturn. This institutional anchor significantly limits downside volatility and signals long-term confidence in the firm's business model even in an environment of price shocks.

In addition, Occidental is benefiting from one of the current oil shocks. While Gulf producers face direct production constraints caused by the closure of the Strait of Hormuz, US production in the Permian Basin continues uninterrupted. Meanwhile, a breakeven price below $40 per barrel means that even with a significant cooling of the geopolitical premium, the company will remain strongly profitable.

The STRATOS project and the long-term outlook

The second pillar of OXY's long-term strategy is CO2 capture technology through the STRATOS project, the world's first commercial DAC station, which is expected to be fully operational by mid-2026. This investment is strategically interesting for two reasons: first, it opens up access to premium-priced carbon neutrality credits from large corporate customers, and second, it builds a platform that can make OXY a relevant player even in the tougher regulatory environment of the future. Analysts point to the launch of STRATOS as a key catalyst for potential share repricing in the second half of the year.

Shares of $OXY have already gained 45% this year.

Phillips 66 $PSX

From a single refiner to a diversified energy player

Phillips 66 has undergone one of the most significant transformations in its industry over the past two years. The company, traditionally viewed as a pure refining colossus dependent on volatile crack spreads (the difference between the price of crude oil (input) and the price of refined products like gasoline or diesel (output)), has been systematically repositioning its business model toward a more diversified and stable source of revenue. In the words of CEO Mark Lashier, 2025 was "a year of transformation" and the fourth quarter results confirm this assessment.

Revenue for the full year 2025 was $132.19 billion, with net income up more than 108% year-on-year to $4.39 billion. In Q4 2025, the company reported adjusted earnings per share of $2.47 versus the consensus estimate of $2.14. Shares of $PSX have appreciated approximately 24% since the beginning of 2026 and are now trading at an all-time high of just under $180.

Key Financial Highlights

Refinery recovery and structural improvements

The refining segment, which was the company's biggest weakness a year ago, is undergoing a significant turnaround. Phillips 66 achieved a record net product yield (88%) in Q4 2025 and operated at 99% of refining capacity. A key factor is the ability to process Canadian heavy crude at a significant discount to the WTI benchmark, and the closure of the Strait of Hormuz will widen this differential further in favor of US refiners.

The Midstream segment complements the refining business with steady revenues from NGL transportation and fractionation. The acquisition of a 50% stake in WRB and the takeover of Coastal Bend have strengthened the company's competitive position, and management has signaled EBITDA guidance of $4.5 billion by the end of 2027. Goldman Sachs $GS recently raised the target price to $186.

LA refinery closure as a strategic move

One controversial but strategically logical move was the plan to close the Los Angeles refinery. Although the company booked $239 million of accelerated depreciation in Q4 2025, it is getting rid of one of the most costly and regulatory-intensive assets in the portfolio. The result is a portfolio with higher average margins and lower environmental liabilities, consistent with the strategy to optimize capital allocation to the most profitable segments.

ConocoPhillips $COP

World's largest independent oil and gas producer

ConocoPhillips is the world's largest independent oil and gas company by production and is one of the few players in the industry whose portfolio diversification includes both premium Améric assets (Permian Basin, Eagle Ford, Bakken) and a global presence in Canada, Norway, Qatar and other regions. The Marathon Oil integration, completed in 2024, added high-quality oil reserves and management reports achieving double the originally planned synergies in the acquired regions.

Fourth quarter 2025 delivered mixed results, with revenue of $14.19 billion, adjusted EPS of $1.02 (slightly below consensus of $1.08) and operating cash flow of $4.3 billion. The main negative factor was lower average realized oil prices, USD 42.46/bbl vs. USD 52.37 in the same period in 2024. Production, on the other hand, surprised positively: total output in Q4 2025 reached 2,320 Mboed (thousand barrels per day), up 137 Mboed year-on-year.

Key financial indicators

Cost reduction program and path to free cash flow

Management's most significant commitment for 2026 is to reduce capital expenditures and operating expenses by more than $1 billion compared to 2025. CEO Ryan Lance has repeatedly called this a "margin improvement initiative" and the company has already shown measurable progress. Well productivity in the Permian is up 8% year-over-year in 2025 in terms of oil production per foot of well length, and 7% in the Eagle Ford. These internal efficiencies allow the company to maintain production at a lower CapEx.

Four large projects, Willow in Alaska (scheduled to begin production in 2029), LNG projects in Qatar, the Surmont expansion in Canada and the Permian Basin development, are the backbone of the plan to double free cash flow by the end of the decade. Management estimates incremental free cash flow from these projects of $7 billion by 2029 compared to a 2025 base year. The Willow project, approximately 50% complete, alone should add approximately $4 billion of annual free cash flow.

The current geopolitical premium in oil prices significantly improves the near-term outlook. With Brent prices above USD 90 per barrel, COP is generating significantly stronger cash flow than the original forecasts for 2026. At the same time, the company has one of the lowest breakeven points in the industry, at around USD 40 per barrel, giving it exceptional resilience even in the event of a rapid price decline following de-escalation of the conflict.

Comparison of key indicators

An overview of key metrics for a quick comparison of companies:

Strategic view

All three companies are direct beneficiaries of increased prices due to the current geopolitical shock, but for different reasons and with different risk profiles.

Occidental $OXY offers the purest exposure to oil price movements in the DAC technology space. The drastic reduction in debt eliminates the existential risk that accompanied the company after the Anadarko acquisition and opens up the space to return capital more aggressively to shareholders. An investor who believes in a continued war premium in oil prices or a future regulatory premium for carbon emissions will find an asymmetric risk/reward ratio in $OXY.

Phillips 66 $PSX provides a different type of exposure. As a refining and midstream company, it benefits from a wide differential between the price of heavy oil and the resulting products, and the ability to process Canadian heavy oil from discounted sources is a distinct competitive advantage in the current environment. A diversified model with five segments (refining, midstream, chemicals, marketing and renewable fuels) reduces dependence on a single price driver and stabilizes earnings. The combination of record efficiency, strong cash flow and strict capital discipline makes $PSX an attractive choice for investors seeking energy exposure with lower commodity volatility.

ConocoPhillips $COP represents the "strength and discipline" category. It is the company with the broadest global diversification, the deepest inventory of drilling locations in the United States, and a clear plan to double free cash flow by the end of the decade. The goal of returning 45% of cash flow from operations to shareholders in 2026 while maintaining an ambitious investment program is an exceptional commitment in the industry. Lower near-term sensitivity to oil price movements (versus $OXY) remains a key risk as some production comes from regions with lower realized prices.

What to watch next

Key factors to focus on in the coming months:

• The length and intensity of the Hormuz crisis: each week the Strait is closed means approximately 10 billion barrels less oil transported. The length of the disruption will determine whether oil companies benefit from a short-term shock or a longer-term restructuring of supply chains.

• Launching STRATOS ($OXY): the first DAC station is expected to begin operations in mid-2026. Its operational results will test whether OXY can monetize carbon credits at commercial scale.

• Refining Margins and Crack Spreads ($PSX): Crack spreads (the difference between the price of crude oil and the resulting petroleum products) remain a key indicator for Phillips 66. In an environment of supply disruptions from the Middle East, they could continue to rise.

• Willow development ($COP): the Alaska project represents the company's largest single generator of future cash flow. Any regulatory or environmental complications could rewrite the outlook.

• OPEC+ moves and production easing: the eight OPEC+ members agreed in March 2026 to add 206k b/d from April. If the geopolitical premium subsides faster than the market expects, there could be a rapid correction in oil prices and thus energy company valuations.

• De-escalation of conflict and opening of the Strait of Hormuz: Israeli Prime Minister Netanyahu's March 19 statements about helping to reopen the Strait signal a possible de-escalation that could significantly reduce the war premium in oil prices. Investors should watch these diplomatic negotiations very closely.

Is the era of the oil giants beginning?

The year 2026 reminds investors that the energy sector remains an indispensable part of the global economic system and that geopolitical events could completely redraw the investment map in a matter of weeks. Occidental Petroleum, Phillips 66 and ConocoPhillips are three companies with different business models and risk profiles, but they share one common characteristic: a distinct ability to generate free cash flow even in a lower oil price environment.

A key factor in the future course of all three will be the length and intensity of the geopolitical disruption in the Persian Gulf and the ability of management to maintain capital discipline no matter which way oil prices go.

The backdrop is a sector wide memory crisis. Mobile DRAM prices surged 70 percent quarter on quarter and 151 percent year on year in Q1 2026, while NAND flash costs jumped 80 percent quarter on quarter and 360 percent year on year, forcing OPPO, vivo and other Chinese Android brands to announce price hikes on existing models. Jefferies now forecasts global smartphone shipments to fall 31 percent to 867 million units in 2026, and IDC expects a roughly 13 percent decline, with both firms noting that Apple and Samsung are structurally better positioned than Chinese OEMs to weather the crunch because of their premium pricing power and financial strength.

How Apple is growing against the market in China

Counterpoint data shows that China's smartphone market weakened 4% year-on-year from January to early March 2026, despite new government subsidies earlier this year. Cautious consumers and longer upgrade cycles are holding back demand, especially for cheaper Android brands.

But Apple $AAPL has been able to increase the volume of iPhones sold by roughly 23% in this environment, for several reasons:

• The basic iPhone 17 qualifies for government subsidies.

• E-commerce platforms offered aggressive discounts and promotions, often combined with subsidies.

• Apple's strong supply chain control has not yet forced it to price it across the board, unlike some of its competitors.

Counterpoint's commentary points out that Apple is better equipped to absorb some of the pressure from rising memory chip prices without having to immediately pass the entire cost on to customers, thanks to its bargaining power with suppliers and cost optimization. This allows it to keep prices stable where OPPO and vivo have already announced price increases for some models.

Pressure on competition: price increases and the search for balance

Rising memory prices are already being reflected in the pricing policies of Chinese brands. According to Caixin and SCMP:

• OPPO and vivo have announced the price hikes of selected models, citing the "steep and sustained rise in memory chip costs".

• Memory component costs rose more than 50% quarter-on-quarter in the first quarter for DRAM and over 90% for NAND in some categories.

• This drove total manufacturing costs up by 11-25% for many models, and by $100-150 per unit for premium devices.

Counterpoint also points out that Huawei has a partial cushion due to a larger share of domestic memory suppliers charging lower prices than large global manufacturers. This allows Huawei to more aggressively target the lower and mid-range segment with less pricing, which mainly threatens other Chinese brands in this band.

Global picture: memory shock and 31% drop in shipments

In its report, Jefferies portrays 2026 as a memory crisis for the entire smartphone market. Analyst Edison Lee states that:

• Global smartphone shipments are likely to fall 31% this year to about 867 million units.

• The original estimate was -12%, the revision is in response to the unexpected rate of memory price increases.

• Mobile DRAM LPDDR5 prices rose 70% QoQ and 151% YoY in 1Q.

• NAND prices added 80% QoQ and 360% YoY, and another 50%+ QoQ growth is expected in 2Q.

According to Jefferies, this means that the cost of memory components:

• for the average Android phone will increase roughly 3.6 times YoY.

• For Apple devices, approximately 4.2 times, as Apple uses higher capacities and more demanding configurations.

Manufacturers are thus forced to choose between three unpopular options: get more expensive, cut specifications, or sacrifice margin. In an environment of weaker demand, each of these is risky.

Who should emerge stronger from the crisis?

Jefferies expects the memory shock to favor Samsung and Apple at the expense of Chinese brands:

• Samsung, thanks to its own memory business and guaranteed access to DRAM and NAND

• Apple due to a less price-sensitive customer base and a strong brand.

The bank estimates that these two players will gain around 7 (Samsung) and 5 (Apple) percentage points of market share in the global market, while Chinese OEM volumes will fall significantly. For example:

• Xiaomi could reduce shipments by 55% in 2026, partially offset by a 31% increase in average selling price (ASP), according to Jefferies.

• Other brands such as OPPO, vivo and Transsion are set to face volume declines of 45-52%.

What this means for the investment view of Apple

For the investor who follows Apple, several things can be gleaned from these two reports:

• Relative strength in China: Apple is able to grow volumes in a weak market and rising cost environment, suggesting it can gain share in key regions even in tougher years.

• Brand resilience: even as memory costs rise, Apple has more room to manoeuvre - it can absorb some of the costs, pass some on in pricing, and lean on user loyalty and the ecosystem.

• Structural share shift: if the Jefferies scenario comes to pass, the global market may shrink, but Apple can slice a larger share of it - especially at the expense of Chinese OEMs in the mid- and high-end segment.

⭐️ I’ve been talking about Uber non-stop for almost 2 years as an investment I fully trust, and so far I don’t see any signs of slowdown or disruption to its business model.

🟢 On the contrary, it’s closing exclusive partnerships with companies like $NVDA $LCID $GOOGL $NBIS $AMZN..

..and just a few hours ago announced a NEW partnership with Rivian for $1.25 billion $RIVN 🎉

🚨 A look at $UBER and its “dying business model”:

🟢 41 million rides daily

🟢 202 million monthly users

🟢 70 countries / in 15,000 cities

✅ +30% higher ride frequency via AVs

✅ One of the TOP network effects

✅ Exclusive partnerships

This doesn’t even remotely look like a sinking business. The opposite is true.

On the stock and as a company, I’m still BULLISH on UBER and the fundamentals confirm it:

For investors, the picture is a classic trade off between near term earnings pressure and long term growth. China e‑commerce is focused on stabilising users and monetisation but is seeing margins squeezed by subsidies and quick commerce competition, while cloud and AI are growing fast and could become a much larger share of group revenue by the end of the decade if current trends hold. The quarter therefore looks weak on short term profit metrics, yet it sketches a framework in which heavy spending on instant commerce and AI could pay off over time, provided the industry avoids another price war and Alibaba can translate its AI and cloud leadership into durable, higher margin cash flows.

How was the last quarter?

Total revenue for the quarter was CNY284.843 billion (about USD40.7 billion), up roughly two percent year-on-year; excluding the Sun Art and Intime retail businesses, it would have been roughly nine percent growth on a comparable basis. Operating profit, on the other hand, fell to CNY10.645 billion (about USD1.5 billion), down about 74 per cent year-on-year.

Adjusted operating profit (Adjusted EBITA) fell 57 percent to CNY23.397 billion (about USD3.3 billion) as Alibaba significantly increased investment in quick commerce, customer experience and technology, only partially offset by better cloud performance and savings in other parts of the group. Net profit attributable to shareholders was CNY16.322 billion (about USD2.3 billion), while total net profit was CNY15.631 billion (about USD2.2 billion), down by about two-thirds. Adjusted net profit was CNY16.710 billion (about USD2.4 billion), also down 67 percent year-on-year.

Earnings per share fell accordingly: diluted earnings per American Depository Receipt (ADS) were CNY5.93 (about USD0.85) and adjusted earnings per ADS were CNY7.09 (about USD1.01), down about 67-71 percent from a year ago. Operating cash flow dropped to CNY36.032 billion (about USD5.15 billion) from about CNY70.9 billion previously, while free cash flow fell to CNY11.346 billion (about USD1.6 billion) - mainly due to heavy investment in quick commerce. On the other hand, the group remains very liquid, with cash and liquid investments of around CNY560.2 billion (about USD80.1 billion).

• Full presentation with results.

Segments and AI + cloud

Chinese e-commerce(Alibaba China E-commerce Group) generated sales of about CNY159.347 billion (about USD22.8 billion), up about six percent year-on-year. Of that, customer management (advertising, commissions) brought in CNY102.664 billion (about USD14.7 billion), up just one percent from a year ago, as transaction activity was weaker and the one-time effect of software deployment fees wore off. Quick commerce (mainly renamed Ele.me, now Taobao Instant Commerce), on the other hand, grew rapidly, with sales increasing to roughly CNY20.842 billion (about USD3.0 billion), up about 56 percent year-on-year, while gradually improving unit economics through growth in average order size and more efficient logistics.

Alibaba International Digital Commerce Group (AIDC) had sales of around CNY39.201 billion (about USD5.6 billion), up four percent year-on-year. Retail international e-commerce (AliExpress, Lazada, Trendyol, etc.) grew three percent, wholesale grew ten percent, while AIDC's losses were significantly reduced year-on-year, mainly due to optimized logistics and more efficient use of marketing spend. The AliExpress Choice business improved unit economics and the "Brand+" program accelerated brand recruitment, which boosted sales.

The most important growth driver is the Cloud Intelligence Group. Its sales reached CNY43.284 billion (about USD6.19 billion), up about 36 percent year-on-year. AI product revenue grew at a triple-digit rate for the tenth consecutive quarter, and the Model-as-a-Service (MaaS) platform is becoming the new cloud driver. Alibaba Cloud is also strengthening reputationally: according to Gartner and IDC, it is the leader in databases, generative AI in Asia and financial cloud in China, with a share of around 43 percent in infrastructure for the financial sector.

Management commentary

In the results, CEO Eddie Wu highlighted that Alibaba continues to invest heavily in its two pillars - AI and consumption - and that AI will be one of the main growth drivers. He noted that Cloud Intelligence Group revenue is up 36 percent and that AI products are showing triple-digit growth for the tenth consecutive quarter, with MaaS already acting as a new cloud engine; on the consumer side, he highlighted the integration of user scenes into the Qwen app, which has reached over 300 million monthly active users. The tone is confident on the AI and cloud side, but between the lines he admits that the price for this strategy is significant pressure on current profits.

CFO Toby Xu commented on the rapid growth of the AI+cloud business as a reason why he is not afraid to ramp up investment, while noting that Quick Commerce is scaling while improving unit economics. He emphasized that strong liquidity and resilient cash flow generation give the company room to sustain high strategic investments. The CFO's tone is consistent with a "growth before short-term profit" strategy, emphasizing that the current decline in profitability is a conscious decision, not just the result of weak demand.

Outlook

Alibaba itself did not present a detailed numerical outlook for the quarter ahead in the press release itself, but several important lines can be gleaned from management's words. First, cloud and AI are set to continue to grow faster than the whole - management expects AI products and MaaS to continue to drive double-digit cloud growth and be even more integrated with the e-commerce ecosystem.

Second, quick commerce (Taobao Instant Commerce) will remain a priority - the company expects continued growth in volume and average order value, as well as continued investment, so the group's profitability will remain under pressure in the coming quarters. Thirdly, the Qwen app is becoming a horizontal interface for many services (Taobao, Tmall, Amap, Fliggy, Alipay) and is set to increase user engagement and generate transactions across the platform in the long term - the outlook thus rests on the assumption that AI-assisted shopping will become a common shopping tool.

Long-term results

For the twelve months ending 31 March 2025, Alibaba achieved revenue of CNY996.347 billion (approximately USD139.3 billion, at an exchange rate of around CNY7.15/USD), up nearly six percent year-on-year from CNY941.168 billion (approximately USD131.5 billion) in the previous period. In 2023, sales were about CNY868.687 billion (about USD121.0 billion) and in 2022 about CNY853.062 billion (about USD118.8 billion) - showing that sales growth has been more in the mid-single digits in recent years, although the absolute volume is huge.

Gross profit reached CNY398.062 billion (about USD55.6 billion) in the latest year, a double-digit growth against CNY354.845 billion (about USD49.4 billion) in the previous year. Thus, gross margin is gradually improving as the mix shifts towards higher value-added (cloud, AI, services) while the firm keeps a lid on the growth in cost of goods sold. Operating expenses were CNY257.157 billion (roughly USD35.9 billion), versus CNY241.495 billion (about USD33.6 billion) a year earlier - rising but slower than gross profit, which allowed operating profit to rise to CNY140.905 billion (about USD19.7 billion) from CNY113.350 billion (about USD15.8 billion).

Pre-tax profit jumped to CNY161.421 billion (about USD22.6 billion) from CNY93.861 billion (about USD13.1 billion), mainly due to the improvement in operating profit and other items; after-tax net profit remained CNY130.109 billion (about USD18.1 billion) compared to CNY80.009 billion (about USD11.3 billion). Thus, earnings per share grew faster than revenue in the past year: from roughly CNY31.6 (about USD4.4) to CNY55.1 (about USD7.9), and diluted EPS from CNY31.36 (about USD4.4) to CNY53.6 (about USD7.6). This is helped by a gradual reduction in the number of shares outstanding from roughly 2.69 billion in 2022 to 2.43 billion in 2025.

EBIT increased to CNY147.710 billion (about USD20.6 billion) from CNY119.507 billion (about USD16.7 billion), EBITDA to CNY182.672 billion (about USD25.5 billion) from CNY164.011 billion (about USD22.9 billion). Thus, over the long term, Alibaba is showing solid profitability growth on a full-year basis, but quarterly numbers (including this December one) can fluctuate significantly depending on the investment cycle, restructuring, and how aggressively the company pushes new quick commerce and AI projects.

Shareholder Structure

Alibaba has a very low official insider stake, around 0.01 percent of the stock, while the institution holds approximately 11.9 percent of the stock and free float. This is a relatively small institutional stake compared to U.S. technology titles, reflecting both regulatory and political risks and the fact that some shares are held through various structures and depositories.

The largest institutional investors include JPMorgan Chase (holding around 0.9 per cent), Primecap Management (around 0.8 per cent), UBS Group (around 0.45 per cent) and FMR (Fidelity) with a holding of just over 0.4 per cent. Most of the shares are therefore held by a broad base of smaller investors and various structures outside of traditional funds, which means greater sensitivity to market sentiment and news of regulation in China, but also room for potential future increases in institutional holdings if the risk profile improves.

News and strategic shift

• Integration of Qwen app into the consumer ecosystem (Taobao, Tmall, Taobao Instant Commerce, Amap, Fliggy, Alipay) - AI chat is becoming a "layer" on top of Alibaba's services that can arrange shopping, transportation and travel in one go; the goal is to make Qwen the main gateway to shopping and services.

• Qwen's rapid growth - consumable Qwen has surpassed 300 million monthly active users, the campaign around Chinese New Year has brought tens to hundreds of millions of first-time AI shoppers; AI is thus becoming not only a marketing tool but also a driver of real orders.

• Developing its own AI infrastructure - T-Head' s chip division has mass-produced its own GPUs for training and deploying models, compatible with major AI frameworks; combined with Qwen and the cloud, Alibaba has more control over computing capacity and costs, which is key in the era of expensive GPUs.

• Strengthening global cloud footprint - Alibaba Cloud operates 92 availability zones in 29 regions and holds a leading position in China in financial and hybrid cloud, according to IDC and Gartner; while expanding outside China to offer AI and database services to international clients.

• Growth of quick commerce and rebranding of Ele.me - the service has been renamed Taobao Instant Commerce and more closely integrated with both Taobao and the Qwen app; quick delivery of food and goods is set to become an important growth pillar, even at the cost of weaker margins in the short term.

• Expanding international partnerships - as part of the AIDC, Alibaba is expanding its collaboration with Shinsegae in South Korea and bringing more brands to AliExpress and other platforms through the "Brand+" program; the goal is to make international business build on stronger brands, not just anonymous cheap goods.

Fair Price

The launch of a generative AI assistant right in the One Medical environment changes the game by bringing AI deep into real primary care processes for the first time. It's no longer just a chatbot, but a tool that works with medical records, lab results, prescriptions, and can suggest a type of care, book a patient, and help with follow-up.

Top points of the analysis

• Amazon has launched a generative AI assistant in One Medical that works directly with patient charts, lab results and medications, and can book and explain.

• This shifts the focus from "health chat" to AI connected to patients, which opens up the topic of regulation, HIPAA, privacy, and an algorithmic approach to care.

• If AI assistant takes hold, it can increase physician productivity, reduce time on routine agendas, lift office utilization, and increase revenue per One Medical member.

• Synergies with Amazon Pharmacy and logistics mean AI can feed more prescriptions and drug orders into its own ecosystem over the long term.

• The healthcare industry is regulated and has high barriers to entry, so successfully integrating AI over One Medical can give Amazon a hard-to-imitate position in primary care.

• The main risks are regulatory (when AI crosses the "support" threshold into the diagnosis area), reputational (erroneous or dangerous recommendations) and execution (adoption by patients and physicians, real impact on numbers, not just PR effect).

Amazon's healthcare story

Until now, Amazon's $AMZN healthcare story was made up of several pieces, but together they didn't act as a clear growth engine. The purchase of One Medical brought a network of primary care practices and a subscription model, Amazon Pharmacy added a distribution channel for drugs, but lacked the digital layer to tie it all together and bring a visible advantage over conventional care providers.

The generative AI assistant in One Medical brings exactly that layer. It can read and explain medical records, interpret the results in a clear way, suggest the appropriate type of care (online chat, video, office, urgent care), and offer booking with a specific doctor right away. What this means for the patient is that they get the answer to "what should I do next" in one app and don't have to wait for someone to call them back.

The impact mechanism looks like this: The AI assistant takes over some of the communication and pre-preparation (collecting symptoms, summarizing the history, explaining findings), the doctor gets more structured input and can spend more time making decisions rather than on administration. This can increase the number of patients per physician, reduce waiting times, and at the same time raise the number of interactions per member (more consultations) without a linear increase in staff costs.

Because the AI runs directly within the One Medical environment and is designed to be HIPAA-compliant, Amazon can work with detailed data about the care process, but must also adhere to strict boundaries around security and consent. This differentiates it from freely available symptom checkers, which don't have access to the actual patient chart and don't carry comparable regulatory liability.

Strategically, it's important that the AI layer is not built "next to" One Medical, but within. This makes it an integral part of the product that increases the allowance per member (ARPU), improves retention, and may enable more differentiated pricing or packages for employers and payers over time.

What needs to work for this to work

• High enough adoption of AI assistant among members for the improvement to show up in aggregate numbers.

• Demonstrable improvements in physician and nurse productivity, not just shifting work from one spreadsheet to another.

• No major regulatory interventions that would limit key functions.

• Visible synergies with Amazon Pharmacy and other healthcare services that bring in additional revenue.

• A compelling story for employers and insurers that the combination of One Medical + AI delivers better care for similar money.

How it becomes money

1) Higher revenue per One Medical member

One Medical operates on a member fee plus reimbursement model for care provided. AI assistant can increase ARPU by:

• Increasing the number of relevant interactions per member (more consultations, check-ups, preventive inquiries).

• Improves adherence to treatment (patients better understand instructions.

• generates more appropriate office visits instead of unnecessary emergency room visits.

In practice, this means that the same network of physicians can serve more "economically active" contacts over time, with fixed infrastructure costs being budgeted for higher volumes of care.

2) Better use of capacity and margins

Primary care is expensive in terms of human time. AI assistant takes the routine activities out of it (explaining basic findings, asking about symptoms, scheduling) so that doctors and nurses can spend a greater proportion of time on what is truly most medically and economically valuable. If the number of patients per doctor can be increased while wage costs are not rising too fast, the operating margins of this part of the business will also increase.

Moreover, from Amazon's point of view, it is not just about the margin within the healthcare segment. Each member of One Medical is also a deeper part of the Amazon ecosystem and has the potential to generate additional revenue in other segments (pharmacy, possibly other medical products, services tied to Prime).

3) Cross-sell into Amazon Pharmacy and other services

An AI assistant that understands specific medications, dosing, and patient history is an ideal place to tie into Amazon Pharmacy. After explaining a finding or adjusting a medication, the system can offer to order the medication, set reminders, or suggest switching to another form of distribution (such as home delivery).

Each such step increases the likelihood that the patient will fill their prescriptions through Amazon rather than a competitor. For Amazon, this means higher prescription volume, better use of logistics, and the ability to offer more relevant services around adherence and chronic disease over time.

The numbers that support this thesis

Because Amazon's healthcare segment does not break out earnings in full detail, an investor must work with what is available and what can be reasonably estimated. More important than the absolute number is the trend and direction.

• One Medical has a network of dozens of clinics in the U.S. and a membership base in the hundreds of thousands to low millions, and is expanding into other states.

• Amazon has been emphasizing the growth of the healthcare segment in recent years and has restructured Amazon Health Services to have clearer accountability and reporting.

• The Pharmacy business is growing at a double-digit rate and forms a logical channel to cross-sell from One Medical.

• Investments in AI within AWS and its own products suggest that AI will not be a one-off project but a consistent part of the strategy across segments.

From a cash flow perspective, healthcare is still a small player within Amazon as a whole, but it has a characteristic that e-commerce does not: more stable, cycle-independent demand and relatively high barriers to entry. If AI can improve margins, today's "tiny" could become a more visible pillar within a few years.

How much of the market can Amazon capture

The market for digital assistants in healthcare today is in the hundreds of millions of dollars in size and is set to grow to roughly $1.5 billion by 2030 at an annual growth rate of around 14%. The broader digital health market is even more ambitious - roughly $200-300 billion today with a view to more than $500 billion by 2030 according to various estimates.

Amazon is no longer entering this from scratch. It has its own One Medical network, Amazon Pharmacy, and now Health AI, a combination that the pure software players don't have. If Amazon manages to monetize a few percent of the expected market for digital assistants and part of digital primary care within One Medical and its own telehealth alone, this gives a revenue potential of hundreds of millions to units of billions of USD per year over 5-7 years, with a trend of increasing margins due to automation.

Theoretical numbers

To get an idea of the potential, let's build a smaller model:

• Let's assume One Medical serves 2-3 million members (including employer programs and Medicare) over the long term.

• Each member generates X number of visits, telehealth contacts, and prescriptions annually.

• Health AI can:

  • Reduce "dead" administrative work by Y%.

  • Enable physicians to handle 10-20% more clinically relevant contacts.

  • Raise ARPU by a few percentage points.

Even if Health AI only brings 5-10% extra to revenue per member at similar or slightly better margins, we're talking about hundreds of millions of dollars in additional revenue per member per year at One Medical alone. Additionally, the combination of higher productivity and better capacity utilization can gradually lift the operating margin of the healthcare segment by several percentage points, which will translate into a multiple of Amazon's overall EPS once the healthcare portion becomes larger.

Position in the race for "health AI agents"

The race for AI assistants for healthcare is not just big tech companies, but also specialized startups and traditional healthcare companies that are testing their own tools for education and decision support. Amazon has three key advantages in this race:

• Integrated intervention - One Medical (clinical care), Amazon Pharmacy (pharmaceuticals), Health AI (digital layer) and AWS/Bedrock (infrastructure), all under one roof.

• Actionable AI agent - Health AI is not just a conversational chatbot, but an agent capable of taking specific actions (book appointments, suggest urgent care, connect to telehealth, prepare documentation for doctors). This puts it just behind regular symptom-determining doctors and gives Amazon a head start in "real" patient flow automation.

• Rapid scale-out capability - once the model has proven itself within One Medical, Amazon can replicate a similar principle in other environments:

  • Extending the AI assistant beta within the Amazon Health site to non-members.

  • Greater integration with pay-per-visit telehealth services.

  • Partnerships with payers or employers to offer "AI triage + One Medical + pharmacy" as a package.

If Amazon manages to stay ahead in "agent-based" AI for primary care, it may become the de facto standard for a portion of the population (especially in urban areas), much like AWS is the standard for a portion of cloud infrastructure today.

Valuation - what's included and what's not

Today's Amazon valuation is primarily based on two pillars: cloud (AWS) and e-commerce with advertising. The healthcare part is implicitly included in the price, but more as an "option" than as a main driver.

This has two effects:

• if the AI in One Medical and the broader health strategy fails, the market will likely absorb it without a dramatic valuation rewrite because the core story is elsewhere.

• If, on the other hand, it can show that healthcare can make a visible contribution to revenue and margin growth, it will be a positive surprise for valuations.

So for an investor, today's AI move in One Medical is more of an asymmetric option. The upside may add a new growth pillar, the downside for now mainly adds execution and reputational risk, but that doesn't threaten Amazon's core business model.

Macro and the market

Primary care and digital health are areas where three pressures have been cumulating for a long time: an aging population, rising costs, and a shortage of medical staff. This is creating a structural demand for solutions that increase efficiency and allow physicians to handle more work without compromising quality.

Amazon is not alone in trying to take advantage of these trends, but is one of the few to combine a real-world network of clinics, a pharmacy business, logistics, cloud infrastructure and proprietary AI tools. This gives it the chance to cover a bigger chunk of the value chain than is common with pure software or, conversely, pure healthcare players.

Risks

The first major risk is regulatory. Once an AI tool makes a triage decision and a recommendation for the next step of care, regulators may conclude that it is a form of diagnostic tool that should fall into a more stringent category. This would mean additional certifications, longer approvals for changes, and potential limitations on features.

The second risk is reputational. One badly handled case where an AI assistant underestimates a serious symptom can have major PR implications, especially when Amazon as a well-known brand is involved. Reactions could come from patients, regulators or partners.

The third risk is enforcement. Just because AI looks good in a presentation doesn't mean doctors and patients will want to use it. If doctors see the system as a burden rather than a help, or if patients prefer to call the practice directly rather than write AI, the real impact on numbers will be much less than PR suggests.

Investment scenarios

Optimistic scenario

In the optimistic scenario, the AI assistant becomes a standard part of care at One Medical. Adoption is high, patients use it, and physicians see it as an aid. Amazon's healthcare segment grows at a double-digit rate, margins improve, and management gradually starts giving more detail in reporting.

For an investor, this means Amazon is gaining a third visible pillar alongside AWS and e-commerce, which can support valuation growth even if some other parts of the business slow down.

A realistic scenario

In a realistic scenario, AI assistant is useful but not revolutionary. It will improve some processes and increase satisfaction for some patients and doctors, but the impact on numbers is spread out and incremental. Healthcare will remain an interesting addition to the Amazon story, but the market will continue to see it as an option rather than a core.

Investors benefit indirectly through better diversification and less dependence on one sector rather than through dramatic rerating.

The pessimistic scenario

In the pessimistic scenario, the AI assistant faces low adoption, regulatory options, and possibly negative incidents. Amazon has to cut back on some features, there is no real improvement in productivity, development and compliance costs rise, and the healthcare segment remains small and problematic on the books.

The market reacts to this with more relief that the core business is elsewhere, but confidence in Amazon's ability to expand into other sectors may suffer as a result.

What to watch next

• How quickly and widely AI assistant will spread among One Medical members.

• Comments from Amazon executives on the impact of AI on productivity and satisfaction in the health segment.

• Visible signs of synergies with Amazon Pharmacy (prescription volume, new features).

• Regulators' reaction, potential investigations or recommendations on AI in primary care.

• Whether Amazon will add the healthcare segment more prominently to its presentations and investor materials.

• Potential expansion of AI features beyond the One Medical app (e.g., integration into the main Amazon app or voice assistants).

What to take away from the article

• Amazon is shifting AI in healthcare from chatbots to a tool that actually works with the patient's chart and influences the primary care workflow.

• This builds a layer that can increase physician productivity and revenue per One Medical member if high adoption succeeds.

• Synergies with Amazon Pharmacy and other healthcare services are key to making AI not just a "nice to have" feature, but a revenue driver.

• Regulation, reputation and enforcement are the three main risks that can slow the whole story down significantly.

• For Amazon's valuation today, AI is more of an option at One Medical, but in a successful scenario, it can help create a third growth pillar.

• An investor holding Amazon for the long term should watch the healthcare segment not for the short-term effect on EPS, but for how it expands sources of growth and reduces dependence on individual industries.

The company Micron Technology $MU posted results that show how strongly artificial intelligence can transform an entire industry. Revenue nearly tripled year-over-year and the firm significantly beat analysts' expectations.

📊 Results for the most recent quarter

Adjusted earnings per share: $12.20 (estimate $9.31)

Revenue: $23.86 billion (estimate $20.07 billion)

A year ago, revenue was about $8 billion.

📈 What is driving the growth?

The main driver is extreme demand for memory chips related to the artificial intelligence boom.

Each new generation of GPUs—especially from Nvidia—requires more memory, which creates supply pressure. The result is a combination of:

memory shortages (DRAM, NAND)

rising prices

significant margin improvement

The company's gross margin rose to as much as 74%, which is an exceptionally strong level in this segment.

📊 Outlook also well above expectations

Micron expects growth to continue in the coming period:

Revenue estimate: about $33.5 billion

EPS estimate: about $19.15

This significantly exceeds market consensus, confirming that demand for memory remains extremely strong.

📦 Where is the biggest growth?

cloud memory: +160% (to about $7.75 billion)

mobile and client segment: a significant jump from about $2.2 billion to $7.7 billion

At the same time, production capacity is shifting toward HBM (High Bandwidth Memory), which is key for AI chips and offers higher margins.

🏗️ Massive investments in capacity

Micron is responding to demand by aggressively expanding production:

The company announced new plants in Idaho and New York, involving investments in the tens of billions of dollars. Production start-up is expected in 2027–2028. Capital expenditures are set to increase significantly in the coming years.

📈 Stock performance

Micron shares are experiencing extraordinary growth:

Year 2025: more than +300%

Year-to-date: approximately +60%

After the results announcement they are currently down more than 4% in pre-market trading!

What is your view on Micron and its current valuation? Do you think the current growth is a long-term trend or just a cyclical phase?

But for investors, it wasn't the actual rate decision that the market was expecting that was key. Much more attention was drawn to Fed Chairman Jerome Powell's press conference and updated economic projections, which for the first time this year reflected the impact of the war in Iran on the US economy. It was Powell's comments on inflation, the oil shock and the future direction of monetary policy that provoked a strong reaction in the markets and pulled the major US indices into the red.

FOMC decision and new economic projections

The Fed held rates unchanged for the second meeting in a row after four rate cuts. In a statement after the meeting, the committee did not significantly change its view on the economy, but added an important sentence, "The effects of developments in the Middle East on the U.S. economy are uncertain." It is this recognition of the uncertainty associated with the war in Iran and its impact on energy markets that is a key signal to investors.

Summary of Economic Projections (dot plot)

Along with the rate decision, the Fed released the first update to economic projections this year. The median FOMC member continues to expect one rate cut in 2026 and another in 2027, unchanged from the December projection.

But what is important is what has changed in the details. Seven of 19 participants now expect rates to remain unchanged this year, one more than in December. Meanwhile, the range of opinion within the commission is extremely wide. For 2027, one member expects a rate increase, three expect no change, four expect one reduction, six expect two, three expect three and one even five reductions. This dispersion clearly shows how little consensus currently exists within the Fed.

Key changes in the projections include:

• Inflation (PCE) for 2026: raised to 2.7% from December's 2.5%, in both headline and core measures

• GDP growth for 2026: raised slightly to 2.4%, for 2027 to 2.3%

• Unemployment: unchanged at 4.4% at the end of 2026

• Long-term neutral rate: slightly raised to 3.1% from 3.0% in December

Higher inflation projections combined with slightly stronger GDP growth suggest that the Fed is counting on the economy's productivity to partially offset inflationary pressures. Powell confirmed this, saying that the higher growth outlook is related to expected gains in productivity.

Powell's press conference: key messages

The press conference that followed the announcement of the decision was the main catalyst for the market reaction. Powell faced dozens of questions on the oil shock, the inflation outlook and the future direction of rates. His answers painted a picture of a central bank that is fully aware of the complexity of the current situation but deliberately refuses to react prematurely.

Inflation: slower progress than the Fed had hoped for

The strongest signal of the entire press conference was Powell's admission that the Fed was not making as much progress in fighting inflation as he had expected. Headline PCE(Personal Consumption Expenditures) rose 2.8% over the past 12 months and core PCE, which excludes volatile food and energy prices, was up 3.0%. Powell pointed out that the elevated readings mainly reflect inflation in goods, which is being boosted by the impact of tariffs. At the same time, short-term inflation expectations have risen in recent weeks, likely related to the sharp rise in oil prices.

Meanwhile, Powell said that a key factor the Fed is waiting for this year is progress in goods disinflation as the one-off price effects of tariffs gradually work their way through the economy. If that progress does not materialize, the announced rate cut is unlikely to materialize. He literally said that the forecast rate cut is contingent on the evolution of the economy, and if there is no progress in inflation, the cut will not happen.

The oil shock and the war in Iran: maximum uncertainty

Powell's comments on the oil shock were among the most candid expressions of uncertainty we have heard from a Fed chief in recent years. He answered repeated questions from reporters about the impact of the war in Iran on the economy with marked caution. He stressed that no one knows what the economic impact will be, and that it may be smaller but also much larger than anyone expects. Several FOMC members reportedly suggested that if they were ever to skip publishing economic projections, this would be an appropriate meeting, because the degree of uncertainty is so high that any numerical estimates are of limited predictive value.

In terms of inflation, Powell acknowledged that higher energy prices would translate into higher headline inflation in the short run. Brent crude oil was trading near $107 a barrel on Wednesday, up nearly 50% since the conflict began. US gasoline prices averaged $3.84 a gallon, an increase of 92 cents in a single month. These factors will push March inflation significantly higher. But Powell also suggested that the Fed's conventional approach to energy shocks is to treat them as one-off price increases that should be absorbed gradually. The problem is that the Fed is also waiting for the inflationary effects of tariffs to wear off, and the two price pressures now overlap, making it difficult to distinguish between temporary and permanent inflationary factors.

Chart of Brent crude oil prices from 2022 - weekly

Stagflation? Powell rejects this term

One of the most watched moments of the press conference was Powell's response to the question of whether the US economy is approaching stagflation. Powell categorically rejected this term. He recalled that stagflation was a concept from the 1970s, when unemployment was in double digits and inflation was much higher. The current situation, with unemployment close to its natural level and inflation one percentage point above the target, was fundamentally different, he said. Yet analysts at Oxford Economics describe the current oil shock as a stagflationary shock because it has the potential to simultaneously weaken growth and raise inflation. While America is not in stagflation in the classic sense, the combination of a weak labour market, persistently elevated inflation and rising energy prices creates an extremely complicated situation for the Fed.

Dual mandate under pressure

Powell has been very candid in describing the stresses facing the Fed. On the one hand, the risks to the labor market are to the downside, which would justify a rate cut. On the other hand, the risks to inflation are pointing up, which would require keeping rates at current levels or raising them.

February's labour market data showed a surprising 92,000 drop in employment and unemployment rose to 4.4%. Powell, meanwhile, pointed out that the slowdown in job growth partly reflects lower immigration and lower labor force participation, not just weaker demand for labor.

It is this stalemate, with the two goals of the Fed's mandate pushing in opposite directions, that is why the Fed is choosing a wait-and-see strategy. Powell has reiterated several times that monetary policy has no pre-determined course and decisions will come on a data-driven, meeting-by-meeting basis. The bar for a rate cut has been raised, which was confirmed by analysts watching the press conference.

Powell's future at the head of the Fed

Powell also weighed in on the question of his future tenure. President Trump has nominated former Fed Governor Kevin Warsh for his post, but his confirmation by the Senate is complicated by Republican Senator Thom Tillis, who is blocking the nomination pending the resolution of the investigation into the reconstruction of Fed headquarters led by Washington Attorney General Jeanine Pirro.

Federal Judge James Boasberg last week brushed off subpoenas issued as part of that investigation, saying the government has presented virtually no evidence of wrongdoing and that the entire action serves as a tool to pressure Powell to lower rates. Powell has confirmed that he will not leave office until the investigation is concluded. If Warsh is not confirmed by the end of Powell's term on May 15, Powell will remain in office as chair pro tem, or interim chair, until his successor is officially confirmed.

Market reaction: stock sell-off, oil above $107

The markets reacted clearly negatively to Powell's words. The Dow Jones Industrial Average lost 768 points to close at 46,225, hitting a new low for 2026 and falling below its 200-day moving average. The S&P 500 index weakened 1.36% to 6,625 points and the tech-heavy Nasdaq Composite fell 1.46%. Wednesday's producer price index(PPI) report for February, which showed a 0.7% increase, more than double the consensus expectation of 0.3%, also contributed to the deepening sell-off. Year-on-year, the core PPI rose to 3.9%, the highest reading in more than a year.

In the bond market, the yield on the 10-year US Treasury note rose slightly to 4.23%. The two-year yield strengthened nearly 4 basis points to 3.71%. The dollar index gained 0.6% after Powell's comments. Brent crude closed at $107.38 a barrel, its highest level since July 2022, after Israel announced the killing of an Iranian security official and Iran attacked a gas field in the United Arab Emirates.

Traders after the press conference began pricing in zero rate cuts by the end of the year, despite the Fed's median projection still calling for one cut. This divergence between the dot plot and market valuation shows how much investor confidence in a rate cut this year was shaken after Powell's comments.

Micron $MU: A bright spot amid the sell-off

Amid the negative market sentiment, memory chipmaker Micron Technology delivered a positive surprise when it posted fiscal second-quarter results after the markets closed. The company's revenue nearly tripled to $23.86 billion, well ahead of the consensus estimate of $20.07 billion. EPS was $12.2 (32.8% above estimates).

CEO Sanjay Mehrotra said that the entire HBM(High Bandwidth Memory) production capacity for 2026 is already sold out. The board of directors approved a 30% dividend increase, and the third-quarter outlook of $33.5 billion in revenue significantly beat analysts' estimates of $22.5 billion. Still, Micron's shares weakened 4% in the aftermarket.

A strategic view

Yesterday's Fed meeting brings several key implications for investors. The Fed is in wait-and-see mode and the bar for further rate cuts has been raised. Powell has clearly indicated that without visible progress in disinflation, a cut will not happen. This means that a higher for longer rate environment remains the baseline scenario for the remainder of 2026.

This has several implications for equity markets. Growth technology companies with high valuations and low current earnings will continue to be under pressure as higher discount rates reduce the present value of their future cash flows. Conversely, firms with real profits, strong cash flow and pricing power should be relatively more resilient. The energy sector remains a direct winner in the current environment of rising oil prices.

The big unknown remains the conflict in the Middle East and its impact on oil prices. The closure of the Strait of Hormuz is turning transportation disruption into a real global supply loss, as Morgan Stanley has pointed out. If the conflict prolongs or escalates, inflationary pressures could worsen significantly and the Fed will face a scenario in which it will have to choose between fighting inflation and supporting a weak labour market. In such an environment, a rate hike cannot be ruled out, although this is not yet the baseline scenario.

The impending change in Fed leadership is also an interesting factor. Kevin Warsh, whom Trump nominated for the chairman position, has historically preferred lower rates. If he is confirmed and takes over the Fed leadership in the coming months, he could bring a shift in communication and approach to monetary policy. But this also creates another layer of uncertainty for markets.

What to watch next

• Oil price developments and the situation in the Middle East: Any easing of the conflict could significantly reduce inflationary pressures and open up room for rate cuts. Conversely, further escalation would further complicate the Fed's situation.

• March inflation data: headline inflation is expected to rise significantly due to rising energy prices. Core PCE will be key to show whether inflationary pressures are spreading to other categories.

• Labor Market Data: After a surprising drop in employment in February, March's payrolls numbers will be key in assessing whether this is a trend or an anomaly.

• Kevin Warsh Confirmation: If Senator Tillis continues to block the nomination, Powell will remain in office beyond his term, which could reassure markets but also extend the period of policy uncertainty surrounding the Fed.

• Earnings season: Corporate earnings for the first quarter of 2026 will show how rising energy prices and lingering tariffs are affecting corporate margins and outlooks. Micron $MU set the bar high for the entire technology sector.

What the session brought to the markets

The Fed's March meeting confirmed that the US central bank is in one of its most difficult periods since the start of the post-pandemic inflation cycle. The combination of persistently elevated inflation, the oil shock caused by the war in Iran, a weak labour market and political pressure on the Fed's leadership creates an environment in which any decision is risky. Powell has adopted a strategy of maximum caution and waiting, which is probably the most responsible approach in the current situation.

For investors, the main message is that the era of higher rates for longer continues and a rate cut is not certain in the foreseeable future. Markets need to adjust to a reality in which the Fed has no room to act preemptively and where every decision hinges on data that could be fundamentally affected by unpredictable geopolitical developments within weeks. In such an environment, the quality and fundamental strength of the portfolio becomes more important than ever.

Canada thus becomes a gateway: BYD does not officially sell passenger cars in the US today, and high tariffs plus regulations on "smart" cars effectively shut the door there, but it is simultaneously considering either its own factory in Canada or the acquisition of a weakened legacy automaker, which would give it a stronger position across the North American region in the future.

From an investment perspective, Micron still operates in a highly cyclical industry that historically swings between acute shortages and painful oversupply, but the current phase looks like a full blown “supercycle” powered by AI workloads, with most forecasts showing demand exceeding supply through at least 2026. The key question for shareholders is how long the mix of strong AI driven demand and disciplined industry supply can last before new megafabs in the U.S., Korea and Taiwan come online and start to normalise margins and returns.

What was Q2 2026 like?

Revenue in the second fiscal quarter of 2026 was $23.86 billion, up from $13.64 billion in the previous quarter and $8.05 billion a year ago. This puts quarter-on-quarter growth at over 70%, and year-on-year revenue nearly tripled - the result of a combination of sharp growth in memory prices and very strong uptake, particularly in AI-related segments.

Gross profit jumped to $17.76 billion with a gross margin of around 74-75%, compared to less than 37% a year ago. GAAP operating profit was $16.14 billion, which equates to an operating margin of about 68%; adjusted operating profit of about $16.46 billion implies an even slightly higher margin. Meanwhile, a year ago, Micron $MU was around 22-25% operating margin - so this is a significant shift, driven primarily by the extraordinary pricing environment.

GAAP net income was $13.79 billion, adjusted net income was $14.02 billion. Earnings per share were $12.07 (GAAP) and $12.20 (adjusted), compared to $4.60-$4.78 in the prior quarter and only about $1.41-$1.56 a year ago.

Operating cash flow was $11.90 billion, up significantly from $8.41 billion in Q1 and $3.94 billion a year ago, and adjusted free cash flow was $6.9 billion. Capital expenditures after accounting for government incentives are roughly $5 billion, a high number, but in the context of booming demand and plans to expand production, this is consistent with efforts to "lock in" capacity for AI generation. At the end of the quarter, Micron had about $16.7 billion in cash and investments, so net debt is well manageable given the size of earnings and cash flow.

The company also announced a quarterly dividend of $0.15 per share, payable in April 2026 to shareholders of record at the end of March. While this is a relatively small amount in the context of current earnings, management frames it as a signal of confidence in the sustainability of the cycle, which may be an important psychological factor for some investors.

• Full presentation with results.

Segments.

The Cloud Memory Business Unit generated revenue of roughly $7.75 billion, up from $5.28 billion in Q1 and $2.95 billion a year ago. The segment's gross margin was around 74% and operating margin was around 66% - exceptionally high figures, driven by a combination of industry-leading AI server products and limited capacity availability across the industry.

The "Core" datacenter division had sales of about $5.69 billion, versus $2.38 billion in Q1 and $1.83 billion a year ago, with a gross margin of about 74% and an operating margin of about 67%. This shows that demand for server memory (not just pure AI, but high performance computing in general) is very strong and customers are willing to pay significantly higher prices for reliable supply.

The Mobile and Client Business Unit (MCBU) segment had revenues of approximately $7.71 billion, up from $4.26 billion in Q1 and $2.24 billion a year ago. The gross margin here is around 79% and the operating margin is around 76%, which is a huge turnaround from just one percent operating margin a year ago. This is related to both the recovery in demand for phones and PCs, and the shift to higher density and performance memories where Micron can collect premium prices.

The automotive and embedded segment generated sales of around $2.71 billion, up from $1.72 billion in Q1 and $1.03 billion a year ago. Gross margin was about 68% and operating margin was about 62%, up significantly from about six percent a year ago - the automotive world is rapidly filling up with memory and storage for assisted driving, infotainment and other smart features, and with the current capacity shortage, this is another source of very profitable growth.

Management comment

Sanjay Mehrotra, CEO and Chairman, termed the Q2 results as new records in sales, gross margin, earnings per share and free cash flow and stressed that he expects similar records in the third fiscal quarter as well. The key message is that the combination of a "strong demand environment, tight supply within the industry and good execution" has created an extremely favorable environment in which memory is becoming a strategic asset for customers in the AI era.

Mehrotra also mentions that Micron is investing in its global manufacturing base to meet growing demand, and that the 30 percent dividend increase reflects confidence in a "consistently strong" business. The tone is very confident, but management also refers to the risks in materials for the Securities and Exchange Commission (SEC), noting the traditional cyclicality of the memory business and other potential factors that could slow the current boom.

Outlook

For the third fiscal quarter of 2026, Micron expects revenue of around $33.5 billion, with a tolerance band of plus or minus $750 million. This marks another significant increase from an already record Q2, and confirms the company's confidence in the continuation of an extremely strong demand cycle, particularly from AI datacenters.

Gross margin is expected to be around 81%, even higher than Q2, and operating expenses are expected to be around $1.6 billion on a GAAP basis and $1.4 billion on an adjusted basis. Earnings per share are expected to be approximately $18.90 ± $0.40 on a GAAP basis and $19.15 ± $0.40 on an adjusted basis. At a diluted share count of around 1.14-1.15 billion, this is an extraordinary level of profitability, approaching the levels of the most profitable semiconductor companies in history.

Management notes in its outlook comments that this scenario is based on continued strong demand for AI memories, disciplined supply in the industry, and the assumption that there will be no sudden overheating of customer investments.

Long-term results

Over the past four fiscal years, Micron has gone through the classic memory boom-bust-boom cycle. It had sales of around $30.76 billion in 2022, dropped to around $15.54 billion a year later, recovered to $25.11 billion in 2024, and shot up to $37.38 billion in 2025. Gross profit even went slightly negative in 2023 (a loss at the gross profit level) as memory prices fell below fully allocated costs; by 2024, gross profit was already around $5.61 billion and jumped to $14.87 billion in 2025.

Operating profit was very strong in 2022, fell into a large loss in 2023, returned to a modest profit of around $1.30 billion in 2024, and jumped to $9.77 billion in 2025. Thus, net income went from a negative $5.83 billion in 2023 to $0.78 billion in 2024 to $8.54 billion in 2025, with earnings per share rising from about $5.34 to $0.70 and then to $7.65.

The number of shares has increased slightly over time (around 1.11-1.12 billion diluted shares), so the EPS growth is primarily driven by a turnaround in profitability, not financial engineering. EBITDA was about $16.74 billion in 2022, dropped to about $2.21 billion in 2023, rose to about $8.94 billion in 2024, and reached about $9.77 billion in 2025. Combined with the current quarterly results, it is clear that Micron has entered a new, extremely strong phase of the cycle - but history reminds us that this phase is not permanent.

Shareholder Structure

Insiders (management and directors) hold only a small proportion of shares, around 0.3 percent. The institution owns about 82-83 percent of the shares, with the rest going to retail and other investors. The largest institutional shareholders include Vanguard Group with about 9.5 percent, BlackRock with about 9 percent, Capital World Investors with about 5 percent, State Street with about 4.7 percent, and FMR (Fidelity) with more than 3.5 percent.

Such strong dominance by large funds means that Micron is a "core" holding in many index and sector funds focused on semiconductors. The stock is very liquid and its performance is sensitive to both specific memory market news and overall sentiment towards technology titles. For the retail investor, this means that it rides with the big pension and index players - but also that if sentiment towards AI and chips changes, movement can be very rapid in either direction.

News and strategic moves

• Dividend increase of 30 percent - the board approved a quarterly dividend of $0.15 per share, a significant jump from the previous level and a signal that management believes in a longer duration of the current earnings cycle, not just a short-term "shock."

• Massive expansion of AI manufacturing capacity - Micron confirms that it is significantly ramping up investment in its global manufacturing base (new and expanded factories in the US, Asia, and especially for HBM and advanced DRAM) to meet demand from datacenter and AI platforms; this keeps capex high in the short term, but should ensure long-term revenue from its most profitable products.

• Use of government incentives - the company continues to draw significant support from the US CHIPS Act and other programs (billions for new US fabs), which lowers net investment costs and improves project returns, but also ties the company to meeting capacity and technology commitments to governments.

• Strengthening its position in Asia - In addition to US projects, Micron is doubling down on its bet on Asia (especially Taiwan and Singapore) with new or expanded manufacturing plants, diversifying geographic risk and bringing production closer to key customers, but remaining exposed to geopolitical tensions in the region.

Fair Price

Rather than betting on a new narrative, Citi points to the gradual shift in Netflix’s model from pure subscriber growth to a focus on operating margins, revenue per user and free cash flow. In its updated post Q4 2025 model, the bank raises revenue and EBIT margin assumptions and expects 2026 operating margins to come in above the Street, while remaining conservative on long term advertising revenue and treating ads as an additional upside lever, not the foundation of the thesis.

How Citi views today's Netflix

Citi $C sees Netflix $NFLX as a streamer that has entered a new phase: the pace of user growth is no longer the main driver, but the company finally has the scale to monetize its catalog and platform to the fullest. Crucially, after years of massive investment in content and global expansion , the numbers are starting to work more in favour of margins.

Citi's commentary highlights three pillars:

• The opportunity to raise EBIT guidance for 2026.

• Room for price appreciation in the US in 4Q26.

• And higher returns on capital due to the absence of large acquisitions and strong cash flow.

The bank sees advertising as an important complement, but not the only valuation driver. This is why it can afford to have an advertising scenario more conservative than consensus and still stick with a "buy" recommendation.

3 reasons to buy:

1) Higher profitability than the market expects

The first reason is to look at operating margins and EBIT. Citi expects Netflix to be able to increase its EBIT for fiscal 2026 and that the actual margin will be about 40 basis points above current consensus.

It relies on:

• Stabilizing content costs - Netflix has already made its largest investments and experiments.

• Better catalog utilization across regions.

• Greater efficiency in marketing and distribution, with the company "burning" less on chasing each new user and benefiting more from its existing base.

Higher margins are key to valuation because even a relatively small shift of a few tenths of a percentage point in operating margin translates into billions of extra in operating profit for a company the size of Netflix.

2) U.S. price appreciation as a revenue catalyst

The second catalyst is expected price appreciation in the U.S. in the fourth quarter of 2026. Citi projects that Netflix still has pricing power, especially in the U.S., where:

• It is one of the main "must-have" services.

• Has a high rate of daily usage.

• and offers a broad mix of series, movies and live content.

Slight price increases in a key market:

• goes virtually entirely to revenue.

• only minimally increases variable costs.

• and, with good timing and communication, may not significantly increase churn (subscriber attrition).

For Citi, it's a clear "mechanical" catalyst - even without dramatic subscriber growth, Netflix can move revenue and profitability higher just by asking a little more for a product that users already consider standard.

3) More cash for shareholders

The third reason is capital policy. Citi points out that Netflix hasn't made any big acquisitions recently, which frees up room for higher returns on capital - especially share buybacks.

The bank argues that:

• Netflix'scash flow is strong enough today.

• The company does not have "transformational" M&A (stock splits) on the table.

• and can afford to increase buybacks and other forms of distribution without jeopardizing content and technology investments.

In an environment where investors increasingly value a combination of growth and disciplined capital management in large technology titles, a stronger buyback program is a clear plus. For Netflix, moreover, buybacks are directly accretive to earnings per share if margins and revenues can be kept on an upward trajectory.

Advertising as a risk, not as a core thesis

Citi also points out that the long-term outlook for advertising revenue is a risk on the expectations side. Consensus reckons Netflix will build an advertising business of around $11bn a year by 2030. Citi is more skeptical and sees a more realistic scenario around $9 billion, with annual growth of about $1.5 billion, not $2 billion, from 2027 onward.

One point that is important for investors: Citi is building its positive thesis to hold up even in a weaker ad scenario. The main drivers it is banking on are margins, pricing and capital discipline, not just an optimistic advertising revenue curve.

Because AWS is already deeply embedded across federal systems and operates GovCloud and classified regions, the alliance gives OpenAI rapid access to agencies it could not easily serve on its own without building certified secure infrastructure. OpenAI says it will still choose which models are offered and can insist on extra safeguards for sensitive deployments, while the Pentagon deal explicitly includes technical and contractual limits meant to keep its models away from domestic mass surveillance and fully autonomous lethal weapons, a stance the company argues also reassures large corporate customers watching how its technology is used by governments.

Top points of the analysis

• The company pays a dividend of around 4% with an annual amount of around $1.78 per share and has had several years of gradual dividend growth, albeit from a relatively low base.

• Revenues have grown steadily, from roughly $12.5 billion to $15.1 billion in four years, net income has been around $1.0-1.3 billion annually, and EPS has fluctuated between $0.7-1.9.

• The firm expects peak demand on its grid to grow 45% from roughly 33.5 GW to 48.5 GW by 2035, with virtually all of the growth to come from data centers and large loads.

• The investment plan calls for tens of billions of dollars of investment, including about $1 billion just to connect large loads, and up to 18% annual growth in the transmission base through 2030.

• Leverage is high: debt to equity of about 2.1, net debt/EBITDA of about 11.9 and Altman Z-score of about 0.8, implying significant leverage and sensitivity to financing conditions.

• The story for an investor stands on a combination of: a stable regulated business, a visible dividend yield and structural consumption growth through AI, balanced against balance sheet risk and the vestiges of a past corruption scandal that the company has been gradually closing in recent years.

Company performance

FirstEnergy $FE is a holding company for several regulated electric utilities headquartered in Akron, Ohio, which today serves more than 6 million customers in six states in the Northeast U.S. - primarily Ohio, Pennsylvania, New Jersey, West Virginia, Maryland and parts of New York. Its primary role is to provide transmission and distribution of electricity in industrially important regions that fall within the PJM Interconnection, one of the largest regional transmission systems in the U.S.

The company was formed in 1997 by the merger of Ohio Edison and Centerior Energy, creating one of the largest investor-owned systems in the U.S.; Ohio Edison itself followed a pattern of consolidation of smaller electric companies since the 1930s. In 2002, FirstEnergy acquired GPU, Inc., expanding its footprint into Pennsylvania and New Jersey and significantly increasing the number of customers and the length of the grid. In the following years, the company combined acquisitions and organic growth, but after the bankruptcy of its former subsidiary, FirstEnergy Solutions (a competitive generation company), it strategically refocused on purely regulated transmission and distribution activities and began systematically divesting competitive generation assets.

Today, FirstEnergy's business is divided into three major segments: Distribution, Integrated and Stand-Alone Transmission. The Distribution segment provides electricity to end-use customers in Ohio and Pennsylvania, where earnings depend on retail rates approved by state regulators and the volume of MWh withdrawn. The Integrated segment combines distribution with limited generation in some areas where the company still has its own resources or long-term contracts; however, its share of earnings has become progressively smaller as management has pushed its emphasis toward the more infrastructure-oriented part of the business. A separate transmission segment manages high-voltage lines and other transmission assets subject to federal regulation (FERC) and typically a relatively attractive regulated rate of return - this is where much of the current investment story related to AI data centers and grid upgrades is taking place.

In terms of where FirstEnergy makes the most money today, regulated infrastructure is the key - the return is based on the size of the regulated asset base (rate base) in distribution and transmission and the allowed return approved by regulators at the state and federal level. In recent years, the company has significantly expanded and updated its network investment program - known as Energize365 - which includes large-scale projects to modernize distribution networks, strengthen transmission, digitize, and improve reliability, among other things, specifically to address the skyrocketing energy demands of AI data centers. This strategy has a twofold effect: responding to regulatory and customer pressures (reliability, resilience, RES integration) while building a larger rate base to which future regulated revenue is tied - a major source of profitability improvement.

At first glance, FirstEnergy's products and services are "boring" - electricity supply, grid operations, new customer connections and related customer services; however, the role of infrastructure partner for large customers such as data centres, industrial parks and manufacturers who need high reliability and capacity is increasingly being promoted within the company. This includes specialised interconnection projects for large loads, where the company designs and builds new transmission and distribution lines tailored to the specific project, and these investments are then regulated for value. As part of its long-term strategy, FirstEnergy profiles itself as a "forward-thinking electric utility" that wants to be the "backbone" of the regional economy in an era of electrification and digitalization - the exact trend that AI is accelerating.

CEO

The company has been led by CEO Brian X since 2023 . Tierney, an executive with nearly 30 years of experience in energy and infrastructure. Prior to joining FirstEnergy, he led Portfolio Operations & Asset Management at Blackstone's Infrastructure Group and before that spent 23 years at American Electric Power (AEP), one of America's largest utilities, where he served as Chief Financial Officer (CFO) and later Executive Vice President of Strategy, among other roles for more than 11 years. That means he brings to FirstEnergy a combination of deep utility business knowledge, capital markets experience and a large investment program - just what the company needs as it launches tens of billions of dollars in grid upgrades and forges relationships with AI players and data centers.

Why is the company interesting?

Our analysis builds a central thesis that FirstEnergy is well positioned to benefit from AI-driven electricity demand growth because part of its grid lies in regions where the bases of new hyperscale data centers are concentrated. It mentions that the company expects system peak load to grow by 15 GW by 2035, from roughly 33.5 GW to 48.5 GW, with virtually all of the growth to come from data centers and other large customers.

The second key thesis is the investment plan - FirstEnergy has increased its five-year capex plan from $28 billion to $36 billion, with a focus on transmission and distribution infrastructure, which is subject to regulated payback. If the investments can be made on time, adjusted EPS could grow at a rate of 6-8% per year, plus a dividend of around 4%, for a total return potential of around 10-12% per year.

Where FirstEnergy operates and what market it serves

FirstEnergy serves about six million customers in Ohio, Pennsylvania, New Jersey, West Virginia, Maryland and parts of New York, where it operates a combination of distribution and transmission networks. This is an area that falls within the PJM Interconnection regional transmission system - one of the largest and most liquid power systems in the U.S., where much of the new industrial and data center projects are clustered.

The PJM region has been under pressure in recent years as demand for electricity from data centers, industrial electrification and the shift of manufacturing back to the U.S. has grown faster than originally anticipated. Studies and grid operators warn that without massive investments in transmission, grid reinforcements and new capacity, the grid may approach limits at some nodes - creating an opportunity for companies like FirstEnergy to invest in "bottlenecks" and then receive regulated returns on those assets.

In addition to residential customers, a key segment for FirstEnergy is the wholesale sector - industry, commercial clients and data centres in particular, which have high demand at a single location and require a robust and reliable connection to the network. Specific interconnection projects are being developed for these clients, which include the construction of new lines, substations and reinforcement of existing infrastructure, with a significant portion of the cost ultimately dissolved in tariffs.

AI data centres and demand growth

FirstEnergy forecasts that the peak load on its grid will grow from roughly 33.5 GW to 48.5 GW, or 45% , by 2035, and itself states that AI data centres and other large electricity-intensive projects are the main drivers of this growth. This is a very significant number by utility standards - historically, electricity demand growth in developed regions has tended to be low units of percent per year or stagnant.

In this context, data centers are different from ordinary consumers in that they represent a long-term, relatively predictable demand, but also a concentrated risk - they typically need hundreds of megawatts at a single location, and connecting them is technically and regulatorily challenging. For FirstEnergy, this means significant investment in specific network sections and connection projects, but also the chance to increase returns in the transmission segment, which often carries a higher approved ROE than pure distribution.

From an investor's perspective, it is also important that this demand growth is not just theoretical - the pipeline of data centers in PJM and surrounding regions has grown significantly in recent years, and transmission operators are openly talking about "AI being the new big electricity customer." If this trend is confirmed in the medium term, FirstEnergy will gain a long-term growth vector that is the exception rather than the rule for utilities.

Management's investment plan and outlook

At the Q4 2025 results, FirstEnergy announced that it is increasing its investment plan for the 2026-2030 period to $36 billion from the original $28 billion, with the largest portion going into transmission and distribution. Investment in transmission is expected to increase from about $14 billion to $19 billion, reflecting the need to strengthen the grid specifically for large customers, including data centres.

Management targets this investment wave to lead to rate base growth at around 10% per year through 2030, thereby enabling 6-8% annual growth in adjusted EPS. The key mechanism is a combination of FERC formula rates in transmission (where the return is automatically based on the invested base) and incremental rate cases at the state level in the distribution side of the business.

At the same time, the company has announced that it will not only fund such a large capex with internally generated cash, but also through higher debt and equity issuance to keep the rating in the investment grade range. This is a crucial point for the dividend: the more FirstEnergy relies on external capital, the more important it will be that regulation and AI demand actually enable the projected EPS growth, otherwise the legs below the 4% dividend will start to wobble.

The business and AI-story

At the heart of the business are the wireline networks and distribution systems that carry electricity from generators to homes and businesses. Revenues come primarily from regulated transmission and distribution tariffs, which are set to cover costs and deliver a reasonable return on invested capital.

The AI story comes through datacenters and large loads. FirstEnergy expects peak demand on its grid to grow 45% by 2035, from 33.5 GW to 48.5 GW, with virtually all of the growth to come from datacenters and associated industrial load. This implies not only higher electricity supply, but more importantly the need for massive investment in transmission, strengthening the grid and connecting large consumers. The company itself talks of 18% annual growth in transmission rate base by 2030 and allocates roughly $1 billion of investment just for interconnection projects for large customers.

For investors, this means:

• A long-term visible investment plan.

• Growth in the regulated rate base on which the approved revenue is calculated.

• and the potential for AI datacenters to "overload" transmission system capacity for several years, strengthening the utilities' bargaining position.

Market and addressable potential

Globally, electricity consumption by datacenters is expected to grow 50-165% by the end of the decade and could account for 6-12% of all electricity consumption in the U.S. by 2030. FirstEnergy's network is in regions where much of the industry and data infrastructure is concentrated, making it one of the direct beneficiaries of this trend.

FirstEnergy's advantage is defined by its licensed area and transmission system, not by free competition across the country. Addressable market growth is therefore not about winning customers from competitors, but about load growth (more MWh) and network investment, which the regulator translates into tariffs. If the projection of 45% growth in peak demand by 2035 comes to pass, FirstEnergy may be one of those "filling the cables" for the AI boom, with the yield being approved and relatively predictable.

Dividend: history, level, sustainability

FirstEnergy pays a quarterly dividend of around $0.44-0.47 per share, which gives a dividend yield of roughly 3.8-4.2% based on the current price. The dividend growth rate is moderate: in recent years, the dividend has increased at a rate of around 1.8-4% per year, and the company has had about 4 consecutive years of increases.

The payout ratio is relatively high: around 75% of earnings. This means the dividend is not ultra-safe like some other lower leveraged utilities, but it is sustainable for now, if:

• EPS remains at least in the range of around $1.7-1.8.

• The investment plan will be largely funded by debt and cash flow.

• and there is no significant drop in profitability or increase in financing costs.

The company has had a problematic episode (corruption scandal and downgrade to "junk" rating around 2020), during which the dividend was maintained but further aggressive growth did not take place. In recent years, the profile has stabilized: profitability is growing, and investors are once again looking at the company as a dividend utility with a structural growth story thanks to its datacenters.

Financial performance

Revenues have grown from roughly US$12.5bn to US$15.1bn over the last four years, which equates to average growth of around 3-5% per annum, with the last year above 12% due to higher loads and tariffs. Operating profit has risen from roughly US$1.9bn to US$2.83bn, with operating margins holding around 14-15%.

Net profit is in the range of USD 0.4-1.3 billion, EPS between USD 0.7-1.9, showing some volatility, but the trend is gradually increasing, especially in recent years. Beta around 0.6 confirms that the stock is behaving more defensively than the market.

The free cash flow margin is around 15%, which is a decent level for a utility, given that a significant portion of cash flow goes into grid investments. Combined with a payout ratio of 70-100%, this gives room to maintain the dividend, but less room for aggressive buybacks or rapid debt reduction.

Balance sheet and debt

The balance sheet is the weaker part of the story. Debt to equity around 2.1, net debt/EBITDA around 11.9, Altman Z-score around 0.8 and very low short-term liquidity ratios (current ratio/quick ratio almost at zero due to the structure of short-term liabilities).

On the other hand, for a regulated utility, debt is not perceived in the same way as for a cyclical firm: investments go into long-term infrastructure, where the cost of capital enters into tariffs. The key is to maintain a relationship with regulators and rating agencies so that funding remains available and tariffs reflect the real cost of the network.

A historic corruption scandal (HB6, allegedly bribing political officials in Ohio over energy legislation) led to a downgrade in ratings and fines, but the company has been making deals with the SEC and putting new controls in place in recent years. For the dividend investor, the bottom line is that the biggest reputational blow has already taken place, but the "memory of the market" and ratings has not yet been completely erased.

Key numbers and valuations

FirstEnergy's net income and EPS have been growing in recent years, but not linearly - 2025 adjusted EPS is around $2.55, which was near the high end of the company's guidance and above analyst consensus, and the company has raised guidance for the next period. This supports the narrative that the investment plan and AI-demand are already partly reflected in financial results, not just presentations.

On valuation, FirstEnergy trades with a dividend yield of around 3.7-4%, a P/E in the higher "utility" multiple range, and an EV/EBITDA that reflects both the regulated nature of the business and increased debt and growth expectations. Compared to more defensive utilities without significant AI exposure, the investor is partly paying a premium for a structured growth story - but they are still buying a regulated business, not a pure growth tech name, so valuation remains within the sector, not outside of it.

Importantly, from a dividend perspective, the combination of a yield of around 4% and a targeted 6-8% EPS CAGR is - if it comes to fruition - attractive; but at the same time, the high payout ratio and sizeable capex increase sensitivity to any disappointment in the numbers. FirstEnergy is thus not a classic "safe bond proxy", but rather a utility that trades some safety for higher AI-linked growth potential.

Valuation

At current levels, FirstEnergy trades at a P/E of around 29, EV/EBITDA of over 25 and P/S of around 2, with a dividend yield of around 4%. These are not pure value multiples, but need to be read in context:

• A regulated business with a lower beta.

• Visible long-term investment and growth plan thanks to AI and datacenters.

• And historic penalties due to scandal and balance sheet.

From a pure "dividend yield" perspective, FirstEnergy is not extremely cheap, but the combination of a 4% dividend and potential rate base growth due to AI may be interesting for a long-term investor. Purely for valuation reasons, it's not a "no-brainer", but within regulated utilities with AI exposure, it's one of those that has a pure growth story to tell.

Risks

• Balance sheet risk: high debt, low Z-score, sensitivity to interest rate environment and rating agencies.

• Regulatory risk: the need to continually defend network investments and tariffs so that customers don't pay too much while funding is sustainable.

• Reputational risk: despite dealing with scandal, further legal or political developments may affect the perception of the company.

• AI hype risk: if the datacenter boom develops more slowly or some capacity goes outside the FirstEnergy region, the expected load growth may not materialize fully.

Investment scenarios

In a positive scenario, AI datacenter expectations come to fruition - projects in the region materialize, peak load actually grows towards +45%, regulation remains predictable and allows FirstEnergy to put most of the capex into rate base with attractive returns. In such a world, it is realistic for adjusted EPS to grow near the upper end of the target (around 8% per year), the dividend to continue to grow in the low single-digit percentage range, and the combination of earnings growth and a 4% yield to offer a double-digit total return over the long term.

In a realistic scenario, the AI pipeline slips in part - some datacenter projects are delayed, others reduce parameters, but electricity demand still grows faster than in the past. EPS sticks more to the lower end of the target (around 6% per year), the regulatory framework remains favorable, but the company needs to consider more the timing of investments and financing to keep the balance sheet under control; the dividend runs further and grows only slowly, the total return is more in the higher single digits.

The downside scenario assumes AI data centers don't meet expectations - some projects run into resistance from regulators and the public due to energy and water consumption, some move to other regions or countries, and global electricity demand grows less than current capex projects. Into this could come a tougher stance by regulators on tariff increases or higher financing costs, which would put pressure on margins and EPS growth; in which case FirstEnergy would have to put the brakes on investment or consider a more conservative dividend policy to protect its balance sheet. In an extreme negative scenario, a capex as high as the one planned today would bring more problems than benefits.

What to take away from the dividend perspective

• A yield of around 4% is attractive in an environment of low volatility and regulated business.

• The dividend history is relatively stable, with recent years of modest growth, but the payout ratio is high, so the dividend is not "untouchable."

• AI and datacenters give FirstEnergy a real growth engine that is not typical of any utility, but will be redeemed by higher investment and debt load.

The number nobody wanted to see

February PPI, the US producer price index, came in at +3.4% year-on-year. The market consensus expected +2.9%. A 0.5 percentage-point difference looks like a small detail at first glance. Unfortunately, in today's situation it is not.

PPI is a precursor to inflation. It shows what producers and wholesalers pay for goods today, and what customers will pay on the shelves tomorrow.

What the number actually means

PPI measures the prices that producers receive for their products — costs that have not yet reached the consumer. It acts as a leading indicator for CPI with a lag of roughly one to three months.

+3.4% year-on-year tells us three things at once:

1) Inflationary pressure isn't just demand-driven. It's not that consumers are spending more. The pressure comes from costs like energy or transportation. And that's the kind of inflation central banks find harder to suppress.

2) The Fed doesn't have the case for quick rate cuts. Throughout the first quarter the market had been pricing in one to two cuts in 2026. After today’s number, that bet is much diminished. Every new inflation surprise extends the period of expensive money.

3) The oil shock from the Strait of Hormuz has not yet been fully reflected in the numbers. February's data captures the world before the escalation in the Middle East. The March figures, which will arrive in a month, could be even more worrying.

The market wants cheap money. Reality isn't cooperating.

The entire year of 2025 was built on one narrative in the markets: inflation is falling, the Fed will cut soon, cheap capital will return. Valuations rose. Sentiment was optimistic.

PPI at +3.4% is a sobering splash. Not a total catastrophe, but a reminder that inflation is not a linear story with a guaranteed happy ending. It's something that returns. And each time it returns, it costs investors more than before, because valuations have meanwhile risen on the assumption that it wouldn't come back.

The whole story around export restrictions therefore looks more like a politically controlled faucet than a definitive stop — freeze for a while, redirect capacity, then allow "limited volumes" with conditions (including that the US wants a share of the revenues) while the queue of Chinese companies that urgently need these chips for their own AI grows. For NVDA the stock-wise implications are another piece of the puzzle: China is shifting from a simple growth engine to a regulation-filtered market, but it doesn’t yet look like Nvidia has completely lost it as a source of demand — trade will likely be more complicated, slower and possibly more profitable per unit due to pressure on subscriptions and special terms.

Micron has gained 62 % year-to-date, Sandisk is up 203 %...

Sure, AI is unstoppable and, thanks to massive CapEx investments by the biggest companies and the never-ending demand, which this week was again confirmed by $NVDA CEO Jensen Huang himself, it looks like these companies will enjoy high margins for a long time.

But we mustn't forget that we're operating in a very cyclical market. Similar, sometimes even larger, rallies were seen in Micron $MU back in 2000 or 2018. A sobering correction always followed.

How do you view this segment of the market now? Are you buying? I personally don't want to jump on this speeding train anymore; if I already owned the shares, I would hold them, but from my perspective opening new positions seems rather risky.

Stock markets are inherently volatile, but the degree of volatility varies dramatically from one title to the next. While some companies move more or less in line with the broader market, others react to each index move with significantly greater amplitude. It is this characteristic that is measured by the Beta indicator, which is one of the fundamental tools of modern investment analysis and an integral part of the Capital Asset Pricing Model(CAPM).

What is Beta and why is it important

Beta expresses the sensitivity of a stock to movements in the broader market, with the S&P 500 Index, whose beta is exactly 1.0, serving as a reference point. If a stock has a beta of 1.5, it means that its price has historically moved 50% more strongly than the index itself. So if the S&P 500 has risen by 10%, that stock has appreciated by 15% on average. But the same principle applies in the other direction. When the index falls by 10%, such a stock has typically weakened by 15%.

Thus, stocks with a beta above 1.0 are more volatile than the market. The higher the beta, the more pronounced the price fluctuations of the stock relative to the benchmark. In contrast, stocks with a beta below 1.0 exhibit lower volatility. Typically, these are defensive sectors such as utilities, healthcare, or consumer staples.

It is important to understand that beta only measures systematic risk, that is, the risk associated with overall market movements. It does not capture firm-specific risks such as product failure, regulatory intervention or management change. Therefore, investors always use it in combination with other metrics such as standard deviation or maximum drawdown.

The beta calculation is based on historical data, typically over a five-year period (5Y Monthly Beta). This is the methodology used by most analytics platforms. However, it is important to keep in mind that historical beta does not necessarily predict future volatility, especially if a company's business model or market environment fundamentally changes.

The following table summarizes the 11 U.S. stocks with the highest beta, ranked from most volatile

Nvidia $NVDA

Why Nvidia is the most volatile large-cap stock on the market

Nvidia clearly has the highest beta of any title at 2.37, making it one of the most volatile mega-cap stocks in the entire U.S. market. This extreme value is primarily due to the company's position at the epicenter of the AI revolution. Nvidia is a dominant supplier of GPU accelerators for data centers, training large language models, and the entire AI infrastructure. Any change in sentiment towards AI investing, whether positive or negative, is immediately and significantly reflected in the stock price.

What drives extreme swings

Nvidia's volatility is amplified by a combination of several factors. The company exhibits extremely high valuation multiples, with a P/E of around 37 and a market capitalization in excess of $4.4 trillion. At such multiples, even a relatively small change in growth expectations can trigger massive price movements. At the same time, Nvidia is exposed to geopolitical risks associated with export restrictions on chips going to China, which adds another layer of uncertainty to the price. Thus, Nvidia's high beta primarily reflects the fact that it is a stock around which a huge amount of speculative and institutional capital is concentrated.

Tesla $TSLA

A narrative-driven stock

Tesla, with a beta of 2.19, has long been one of the most volatile stocks in the US market. Unlike Nvidia, where volatility is primarily driven by the AI cycle, a combination of multiple factors play a key role in Tesla. The company operates at the intersection of electric mobility, autonomous driving, energy and now, increasingly, robotics. Each of these segments brings its own set of risks and opportunities that overlap in share price.

Politics, Musk and sentiment

Tesla CEO Elon Musk plays a significant role in Tesla's volatility, with his public activities and political engagement directly impacting investor sentiment. Thus, Tesla shares historically react not only to financial results, but also to tweets, controversies, and regulatory decisions. In an environment of high media attention, Tesla has become a stock where daily trading volume is often well above average, amplifying short-term price swings. For investors with a higher risk tolerance, Tesla presents an opportunity, but it requires significantly stronger nerves than most other mega-cap titles.

AMD $AMD

Nvidia's eternal challenger

Advanced Micro Devices, with a beta of 2.02, is the second most volatile semiconductor stock. AMD has established itself in recent years as a key competitor to Nvidia in the AI accelerator segment, as well as Intel in the server processor market. It is this position as a challenger on two fronts that is the source of the high volatility. Each new order, partnership with a hyperscaler or, conversely, loss of market share will have a much stronger impact on the price than for companies with a dominant position.

AMD recently signed significant contracts with both Meta Platforms and OpenAI to supply its Instinct GPUs, and EPYC processors hold approximately 24% of the server x86 chip market. Still, AMD's market capitalization is significantly smaller than Nvidia's, which means the percentage impact of each news item on the stock price is proportionately larger. Investors who believe in a continued AI boom but are looking for an alternative to Nvidia with potentially higher upside often head to AMD.

Lam Research $LRCX

The cyclical nature of semiconductor equipment

Lam Research, with a beta of 1.78, is one of the world's largest manufacturers of semiconductor chip fabrication equipment, specifically etch and deposition systems. The high beta here is a direct result of the extreme cyclical nature of the entire semiconductor wafer fabrication equipment (WFE) segment. Chipmakers' investments in new manufacturing capacity tend to come in waves, and it is these waves that are strongly reflected in the earnings and valuations of companies like Lam Research.

Currently, Lam Research is benefiting from a massive investment cycle driven by demand for HBM memory and advanced chip packaging for AI infrastructure. Revenues for the latest quarter are up 22% year-over-year and the stock has appreciated more than 185% over the past 12 months. But it's growth this fast that also increases sensitivity to any negative news, be it tightening export restrictions or signals of a slowdown in the investment cycle.

Netflix $NFLX

Streaming giant with surprising volatility

Netflix, with a beta of 1.71, may surprise with its presence in this review. It is, after all, an established streaming platform with a predictable subscription-based revenue model. But Netflix's volatility stems from other sources. The stock has historically reacted extremely strongly to quarterly results, particularly new subscriber figures. All it takes is for a number to deviate from consensus by units of one percent, and a stock price can move 10 to 20% in a single trading day.

In recent quarters, Netflix has been expanding its portfolio to include an advertising model, live streaming of sporting events and gaming. While these new segments diversify revenue, they also bring uncertainty about the success of their monetization. With a market capitalization in excess of $400 billion and a P/E of around 37, Netflix is valued at a level where the company is expected to continue to grow, and any disappointment could lead to a sharp decline.

But the biggest volatility of recent weeks has clearly been brought to this stock by the scramble to buy Warner Bros Discovery $WBD. The stock rallied sharply after Netflix didn't outbid a competitor's highest bid and abandoned the acquisition.

Palantir Technologies $PLTR

With a beta of 1.69, Palantir is one of the most followed stocks of the current AI boom. The firm supplies analytics and AI platforms to government institutions and the commercial sector, and its AIP platform has seen a significant acceleration in usage in recent quarters. Palantir's volatility stems primarily from the extreme disconnect between its high growth momentum and valuation, which is one of the highest in the entire market. The trailing P/E exceeds 200 and the forward P/E is above 100.

At such a tight valuation, every quarterly result becomes a binary event. If the company beats expectations, the stock could rise sharply. If results fall only slightly short, the market reacts by selling off. Moreover, Palantir generates a significant portion of its revenue from government contracts, which are subject to political decisions and budget cycles, adding another layer of unpredictability.

PS: Shares started shorting Michael Burry this year and are down 18% since his announcement.

Applied Materials $AMAT

A key supplier for chip manufacturing

Applied Materials, with a beta of 1.68, is the world's largest materials engineering firm for the semiconductor industry. Similar to Lam Research $LRCX, Applied Materials is subject to the cyclicality of the WFE market, but exhibits slightly lower volatility due to its broader product portfolio that includes deposition, etching, metrology and services. The company is benefiting from the increasing complexity of its manufacturing processes, where AI chips require significantly more manufacturing steps per wafer than previous generations.

Exposure to the Chinese market remains a risk for Applied Materials, which has exceeded 25% of total revenue in recent quarters. Tightening export controls by the US government may significantly impact this revenue stream. It is the uncertainty regarding Chinese export regulations that is keeping the Applied Materials beta above 1.5, despite being a stable and profitable company from a fundamental perspective.

Oracle $ORCL

Cloud transformation increases volatility

With a beta of 1.65, Oracle is undergoing one of the most ambitious transformations in its segment. The company is investing massively in building cloud infrastructure (OCI) to compete with AWS, Azure and Google Cloud in the rapidly growing AI cloud services market. In the past 12 months alone, Oracle has invested over $48 billion in capital expenditures, far exceeding the company's free cash flow and generating negative FCF.

It is this aggressive investment strategy that increases the stock's volatility. Investors are continually flip-flopping between an optimistic scenario where Oracle gains significant share in the AI cloud market and a pessimistic scenario where massive CapEx fails to deliver sufficient returns. The stock has already weakened 55.3% since its 52-week high of around $345, illustrating how quickly sentiment towards transformation stories can change.

Caterpillar $CAT

An industrial giant with a surprisingly high beta

Caterpillar, with a beta of 1.56, is the only truly "non-technology" company in this review. As the world's largest manufacturer of construction and mining equipment, Caterpillar is directly exposed to the global economic cycle. When the economy grows, demand for heavy equipment rises and the stock appreciates significantly faster than the market. Conversely, in recessionary periods, Caterpillar is one of the first victims of the selling wave.

Recently, Caterpillar's beta has increased due to several factors. The company is benefiting from massive infrastructure investments in the U.S. funded by federal programs, from rising demand for mining equipment driven by the commodity boom, and from the recovery of the industrial cycle. At the same time, it faces risks from geopolitical instability, a potential slowdown in the Chinese economy and rising materials costs. The stock has gained nearly 110% over the past 52 weeks. Thanks to this bull run, which has clearly been one of the strongest in the company's history, the stock's volatility has also risen significantly.

Micron Technology $MU

Memory supercycle and cyclical volatility

Micron Technology, with a beta of 1.54, is a prime example of an extremely cyclical company. The memory industry has historically been one of the most volatile segments of the semiconductor market, where DRAM and NAND prices fluctuate based on the balance of supply and demand. Currently, Micron is in one of the strongest supercycles in its history, driven by demand for HBM (High Bandwidth Memory) for AI accelerators.

The stock is up more than 360% in the past 52 weeks and the company has sold out its entire HBM production for 2026 under contract. Gross margins are around 68%, which is an extraordinary level for a memory company. Historically, however, it is when the memory cycle seems to be strongest that the risk of a turnaround is highest. If competitors start aggressively ramping up capacity or demand from AI infrastructure slows, Micron's beta could quickly backfire.

ASML Holding $ASML

Monopoly position mitigates volatility

ASML, with a beta of 1.43, has the lowest relative value of all the titles tracked, yet is well above the market average. The firm holds a de facto monopoly on the production of EUV lithography systems, without which the most advanced semiconductor chips cannot be produced. This unique position provides ASML with significant pricing power and high visibility of future revenues through multi-year contracts.

ASML's volatility stems primarily from geopolitical factors. The company is based in the Netherlands and its technology has become subject to international export restrictions, particularly in relation to China. Any new regulatory decision by European or US authorities could significantly affect the company's outlook. Yet it is precisely the monopoly position and the irreplaceability of EUV technology that ensure that ASML exhibits considerably lower volatility than other companies in the semiconductor equipment segment.

Volatility comparison across sectors

Looking at the whole picture, it is clear that high beta is not the preserve of any one sector. Semiconductor firms dominate the top of the list (Nvidia, AMD, Lam Research, ASML, Applied Materials, Micron), but we also find streaming (Netflix), enterprise software (Oracle, Palantir), automotive (Tesla), and heavy engineering (Caterpillar).

The common denominator is rather the nature of the business: high beta firms typically operate in industries with rapid technological change, strong cyclicality or high valuation multiples.

It is also worth noting the distinction between "high-quality" and "dangerous" volatility. For companies like Nvidia or ASML, high beta reflects a strong position in a growing market where volatility acts as a tax on extreme growth. Conversely, for some firms undergoing transformation, such as Oracle, a high beta may signal genuine uncertainty about the future direction of the firm.

A strategic view

For active investors, high beta stocks can be a tool to boost portfolio returns in a bull market. If an investor expects continued positive sentiment and rising demand for AI infrastructure, an allocation to titles like Nvidia, AMD or Micron can allow a portfolio to significantly outperform the benchmark. In a bear market, however, these same stocks generate deep losses and investors must be prepared for drawdowns in excess of 30% to 50% from the highs.

We have seen this happen several times with Micron $MU, for example. Like in 2000 or 2018.

More conservative investors can use volatile titles selectively, for example by buying on significant corrections when valuations become more attractive. Many experienced investors also combine high betas with hedging strategies such as protective put options or other strategies that limit potential losses in exchange for a reduction in maximum profits. But this is how the really professional traders or large funds work with volatility, not the gentle stock investor.

The key is to view beta not as an absolute measure of quality or risk, but as one tool that helps an investor understand how a given stock is likely to behave in different market regimes. A high beta is not in itself a reason to buy or sell. It always depends on the context of fundamentals, valuation and the investor's position within the overall portfolio.

What to watch next

• Evolution of the AI investment cycle: any signs of a slowdown in CapEx spending by hyperscalers could significantly impact the semiconductor segment

• Geopolitical tensions and export regulations: in particular, decisions regarding chip restrictions against China directly impact ASML, Lam Research, Applied Materials and Nvidia

• Fed rates and discount factors: higher rates are pushing up valuations of growth companies with high multiples, which increases volatility of titles like Palantir or Tesla(Fed meeting will take place tonight - but the market does not expect a rate cut).

• The memory cycle: the balance of supply and demand in DRAM and HBM memory will determine the sustainability of the supercycle for Micron and other memory companies

• Quarterly results: for companies with high beta, earnings reports tend to generate above-average price movements, so it is advisable to monitor the earnings calendar and adjust positions ahead of the report if necessary

So is beta bad or good?

High beta is a double-edged weapon. In a rising market it increases gains, in a falling market it deepens losses. The eleven stocks in this review are among the most volatile titles on the U.S. stock market, yet in most cases they are companies with real competitive advantages, strong fundamentals, and exposure to the key megatrends of the day. Thus, volatility in these titles is not a sign of weakness, but a reflection of how much opportunity and, at the same time, how much risk the market sees in these firms.

Combined with thorough fundamental analysis, an understanding of cyclical patterns and responsible position size management, high beta can be an asset rather than a threat.

Because AWS is already deeply embedded across federal systems and operates GovCloud and classified regions, the alliance gives OpenAI rapid access to agencies it could not easily serve on its own without building certified secure infrastructure. OpenAI says it will still choose which models are offered and can insist on extra safeguards for sensitive deployments, while the Pentagon deal explicitly includes technical and contractual limits meant to keep its models away from domestic mass surveillance and fully autonomous lethal weapons, a stance the company argues also reassures large corporate customers watching how its technology is used by governments.

How OpenAI replaced Anthropic

Anthropic won a contract with the Pentagon for up to $200 million last year, supplying the Claude model for classified military and intelligence systems through Palantir $PLTR and AWS $AMZN. But the relationship broke down this February. Anthropic refused to remove restrictions on select deployments - particularly domestic surveillance and potential use for autonomous weapons - and the Pentagon formally designated the firm a "supply chain risk" after a series of warnings and decided to phase out the collaboration over six months.

Shortly thereafter, OpenAI announced that it had won a contract to supply ChatGPT and other models to roughly 3 million DoD employees for both classified and unclassified work. According to The Information, the contract is expected to bring in "only" a few million dollars over 15 months, a fraction of OpenAI's estimated $30 billion in 2026 revenue. But the strategic value is elsewhere: OpenAI has entered the nation's critical infrastructure where a competitor has been a major player.

Why the AWS deal is key for OpenAI

The partnership with AWS moves OpenAI beyond the Pentagon itself. Amazon's cloud has deep roots in many federal agencies, and AWS has agreed to offer OpenAI products across its government clientele, according to sources. So OpenAI is gaining.

• Access to a broad base of agencies through existing AWS contracts.

• The ability to deliver AI for both classified and unclassified environments without having to build its own certified infrastructure.

• a strong signal of confidence to large enterprise customers who often see government deployments as a test of reliability.

The deal is also made possible by an adjustment in the relationship with Microsoft. Following OpenAI's move to a for-profit structure, the contract was updated in 2025 so that Microsoft would not have a pre-emption right on all compute for OpenAI and so that OpenAI could provide APIs to government and security customers regardless of the cloud. So OpenAI's first products continue to run on Azure, but the company is now free to use other providers - including AWS - for specific segments such as national security.

The impact on Amazon and the "cloud war" on government

For Amazon, the partnership with OpenAI is a welcome counterpunch at a time when Microsoft $MSFT had a strong marketing argument for Azure thanks to its exclusive relationship with OpenAI. AWS can now offer OpenAI models to governments alongside its own services and those of other partners, strengthening its position as an AI "neutral marketplace".

Moreover, the cloud is strategically important to government:

• It generates long-term stable revenue from multi-year contracts.

• Increases reputational capital with corporate customers.

• gives providers deeper insight into specific security and regulatory requirements, which they can then monetize in the commercial sphere.

This gives AWS the argument that it can be a platform for a variety of top models - from custom to Anthropic (where allowed) to OpenAI - and that it can respond quickly to changes in Pentagon or other agency policies.

Risks and open questions

On the Anthropic side, the dispute with the Pentagon may spill over into the legal arena. The company has previously indicated that it will challenge the DoD's actions, and the outcome of the dispute may affect future rules for AI purchases across federal agencies. At the same time, President Donald Trump's decision to have authorities phase out other models to replace Claude within six months creates room for OpenAI, xAI and other competitors that have already tapped into classified networks.

For OpenAI and Amazon, the main risk is political and reputational. The national security segment is lucrative but also sensitive - work on classified projects, questions around surveillance and autonomous weapons, and public debate about where AI has limits in defense. If controversial deployments emerge, pressure to regulate and limit cooperation with some contractors could quickly increase.

Financially, the new contracts do not yet represent a major part of OpenAI's revenue - a Pentagon contract of "a few million dollars" over 15 months is more of a pilot project than a major source of profit at the scale of a company with an expected $30 billion in revenue in 2026. More important is how quickly it can expand AI services to more agencies through AWS and other channels, and what volume of contracts it can pick up after Anthropic's final exit.

What to watch next

In the coming months, the key:

• What specific agencies other than the Pentagon will begin using OpenAI over AWS.

• How Anthropic's dispute with the Pentagon will play out, and whether it will lead to new rules for AI in government.

• whether OpenAI will begin to list larger government contracts as a significant revenue segment, or whether national security will remain a strategic rather than financial pillar.

• how Microsoft, Google and others in the government AI solutions space will respond.

Warren argues that record profits, very low effective tax rates and big layoff waves are “another example of unrestrained corporate greed” made easier by the current tax code. Meta $META stands out in her criticism: analysis from the Institute on Taxation and Economic Policy estimates its effective federal tax rate in 2025 at just over 3.5 percent, while reports say the company is weighing cuts of up to 20 percent of its workforce to offset surging AI infrastructure costs. Similar moves at Amazon and UPS, which have announced plans to cut around 16,000 and 30,000 jobs respectively, turn the link between tax relief, layoffs and public reputation into a new political risk investors will have to watch.

What exactly is Warren calling for

In her letters, Warren asks that individual companies provide:

• An accounting of the tax breaks and other benefits they received in 2025 thanks to the One Big Beautiful Bill Act.

• Information on whether they expect to receive a refund of tariffs or other one-time benefits.

• a summary of the number of layoffs in recent months, broken down by segment and location.

• An explanation of how layoffs relate to investments in AI, automation and other productivity measures.

Meta is cited in the letter as an example of a company that generated a record $79 billion in U.S. revenue in 2025 while paying only about $2.8 billion in federal tax, an effective rate of just over 3.5%, an all-time low since going public in 2012. At the same time, according to Reuters, the company is considering laying off up to a fifth of its workforce to offset the cost of massive investments in AI infrastructure and to prepare for higher productivity through automation.

At Amazon, Warren is following up on earlier criticism of billions of dollars in tax breaks and simultaneous layoffs, with the company cutting red tape and slimming corporate structures while leaning on the growth of e-commerce, cloud and AI services. UPS is mentioned in the letter in the context of shedding 30,000 positions in a year when it is reshaping its relationship with Amazon and the company is optimizing delivery volumes.

Why the topic is sensitive right now

On aggregate numbers, some may argue that the layoff rate is low and the economy is holding up relatively well after the pandemic. But the situation is different for an individual layoff.

• Few new jobs are being created.

• People are less likely to leave voluntarily because they are afraid they won't find anything better.

• Experienced professionals are also applying for entry-level and mid-level positions, so competition is tougher.

Against this backdrop, a series of big layoff announcements - often justified by "efficiency" or "better productivity through AI" - look much worse in public when accompanied by information about extremely low tax rates and record profits. Politically, this opens up space for the argument that tax breaks for big business do not lead to job retention, but rather to enhanced margins and shareholder rewards.

What this means for Meta, Microsoft, Amazon and others

In the short term, Warren's letters may mainly mean reputational and political pressure. The questions themselves do not change the tax regime or dictate to companies how to manage headcount. But when combined with media coverage and other campaigns (e.g., unions, think tanks like the Institute on Taxation and Economic Policy), the issue can become the basis for.

• Proposals to limit some tax breaks.

• making future incentives conditional on employment or investment commitments.

• Stricter monitoring of AI-conditional layoffs in policy debates.

Meta is a typical symbol here. The company is investing tens of billions in AI infrastructure while planning perhaps the largest layoffs since the "year of efficiency" 2022-2023. From an investment perspective, this may improve margins in the short term, but politically it becomes an easy target: low taxes, massive buybacks, high AI CAPEX, and thousands of people out of work are powerful ammunition for Senator Warren and other critics.

Microsoft and Amazon, meanwhile, feature less specifically in the text, but they are in the same category: benefiting from tax changes, investing heavy billions in cloud AI while downsizing in some segments. This may attract more attention in the future, especially if pressures for more "efficiency rounds" increase.

Risks for investors

In the short term, the main concern is reputational. Warren's "declaration of war" on large companies in her letters will not in itself change the numbers in the next quarter, but may add to stock volatility on any further news of tax changes or large layoffs.

In the medium and longer term, it is more important whether initiatives like these turn into concrete legislative proposals:

• Revisions to portions of the One Big Beautiful Bill Act related to tax breaks for large technology companies.

• Conditioning tax benefits on maintaining a certain level of employment in the US.

• New obligations regarding transparency of tax rates and planned AI-contingent layoffs.

What this could mean for shareholders:

• Higher effective tax rates for companies like Meta, Amazon, Microsoft.

• Less room for aggressive buybacks and other capital operations.

• Pressure for slower or more sensitively communicated layoffs if they are associated with AI efficiencies.

What to watch next

For investors holding or considering titles like Meta, Microsoft, Amazon, UPS, Target, the key to watch will be:

• Whether the companies respond publicly to the Warren letters at all, or let everything happen quietly.

• Whether the issue makes it to the broader congressional hearings (finance, labor, technology committees).

• how the media and analysts are handling numbers like the 3.5% effective rate of the Met and the planned 20% layoffs.

• whether similar criticism will spread to other big tech stocks that benefit greatly from AI and Trump's tax changes.

For Alphabet shareholders, the important point is that the company is not trying to build its own humanoid plants, but to position itself as a kind of Android for robots, licensing a standard intelligence layer into machines built by partners like Boston Dynamics and, via Intrinsic, a wider set of industrial manufacturers. If this strategy works, Google could tap into data flows from fleets of robots to train better Gemini models and capture high margin software and cloud revenue from robotics without the capital burden of hardware, though the scale of that opportunity and the speed of adoption remain open questions that this article will unpack in more detail.

Top points of the analysis

• Google DeepMind integrates Gemini Robotics AI into Humanoid Atlas and other Boston Dynamics platforms as a "brain" layer that translates vision and language into specific motion commands.

• In doing so, Alphabet is testing its "intelligence without hardware" strategy - it wants to become the horizontal standard for robotic AI across manufacturers (Boston Dynamics, Apptronik, Agility and others), much like Android in the mobile world.

• Data from pilot deployments of Atlas in Hyundai factories and other industrial plants will be tuned to the real world by Gemini Robotics, which can create a strong data advantage over competitors that don't have such deep access to physical robots.

• Monetization can come through licensing and "usage-based" models of Gemini Robotics that will run through Google Cloud, expanding the addressable cloud market and adding a new type of enterprise customer in robotics.

• If Gemini Robotics establishes itself as the default intelligence for multiple robotics platforms, Alphabet gets network effects - every robot in the world that uses Gemini improves the model for others.

• The risk is dependence on partners who must actually scale the hardware, and uncertainty about how quickly companies will be willing to pay for AI layers in robotics, which is still in its early commercial phase.

What has changed (from Google's perspective)

$GOOG has been doing robotics research for years ( Everyday Robots, for example), but standalone hardware projects have ended without much commercial impact. In the latest wave, DeepMind decided on a different strategy: instead of building its own robots, it's focusing on creating generic Gemini Robotics models and deploying them on partners' hardware.

Gemini Robotics 1.5 and Gemini Robotics-ER are models that combine vision, language, and action - combining the ability to understand a visual scene, plan multiple steps ahead, and generate motor commands for specific robots. Boston Dynamics brings "athletic intelligence," the ability of robots like Atlas or Spot to navigate challenging environments, but until now has lacked a sufficiently versatile cognitive layer for adaptive tasks.

The partnership announced at CES 2026 therefore seems like a logical move for Google: instead of making its own humanoids, the company will "piggyback" on the brand and engineering of Boston Dynamics and Hyundai and deliver what it does best - a foundation model and orchestration of how the robot should behave. At the same time, Google stresses on the blog and in Gemini Robotics materials that the same intelligence is to run on other partners like Apptronik, Agility and others, confirming its ambition to become the "Android for robotics".

What this means for Alphabet

• Every Boston Dynamics robot with Gemini is a live data sensor for model improvement.

• Google does not have to bear the capital-intensive risk of a manufacturing company, but collects licensing and cloud fees.

• If Boston Dynamics and other partners succeed, Google is "in" on every robot sold, regardless of the specific brand.

How this becomes money for Alphabet

1) Licensing and usage-based revenue from Gemini Robotics

Gemini Robotics is designed as a foundation model delivered via API and cloud. For Alphabet, this means multiple layers of monetization:

• A licensing fee for each robot that uses Gemini (per-device license).

• Usage-based fees for the computing resources in Google Cloud that the robot consumes in perception, planning, and decision-making.

• Alternatively, higher enterprise tariffs for large customers (Hyundai, logistics companies, manufacturers) who deploy larger fleets of robots.

The market for humanoid robots themselves will be relatively small in the first phase (hundreds to thousands of units per year), but the key point is that Gemini Robotics is also intended to run on quadruped robots (Spot), mobile platforms, and other types of robots - so potentially tens of thousands to hundreds of thousands of devices over several years.

Even if Google generates "only" the low hundreds of dollars per year in net AI fees per robot in the early years (a combination of license and compute usage), with tens of thousands of robots, that already means revenues in the low hundreds of millions of dollars per year, which, moreover, will likely run through a high-margin cloud business.

2) Expanding the addressable market for Google Cloud

Every robot with Gemini Robotics is also a cloud customer. Physical robotics therefore expands the addressable market for Google Cloud to include a segment that has been virtually non-existent until now - industrial robots, logistics, manufacturing and service robotics.

These businesses typically need:

• Access to models.

• Storage for data and logs from robots.

• Analytics tools and MLOps for custom debugging and monitoring.

This is exactly where Google Cloud is looking to grow and where it can leverage synergies with AI - a similar story to what we see today with Microsoft (Azure + Copilot) or AWS (Bedrock + customer models). This makes Gemini Robotics not just a product in itself, but an acquisition channel for new cloud customers in the industry.

3) Data advantage and "moat" in embodied AI

The biggest value for Google in the short term may not be licensing revenue, but data. Gemini Robotics-ER is designed as an embodied reasoning model that orchestrates robot actions based on environmental perception and planning. Every Atlas or Spot that runs at a Hyundai factory or other customer generates real data:

• How the robot reacts to unexpected situations.

• What mistakes it makes.

• what strategies lead to successful task completion.

This data may be anonymized and aggregated, but for training and debugging Gemini Robotics, it represents a powerful data source that competitors without similar partnerships won't have. This is similar logic to mobile - Android generates usage data (under privacy rules) that helps Google tune its products, and thus build further leads.

Growth potential: what can be in it for Alphabet $GOOG

1) Robotics as an extension of AI TAM, not a new core business

Analyst estimates for the "digital assistants in healthcare" market put the size at around $1.5 billion by 2030, but the market for digital robotics and humanoids will be orders of magnitude larger - combining hardware, software, and services in industry, logistics, and perhaps even the home.

For Alphabet, the point is not to own the hardware, but to get the AI part of the value chain. If it can make even "just" $1-3 billion a year in licensing and cloud usage associated with robotics (across partners, not just Boston Dynamics) over 5-10 years, at typical cloud margins, this would represent an additional high-margin segment alongside the existing Google Cloud and AI services.

This is a small part compared to Alphabet's revenues today, but important from a valuation perspective - the market today values mainly Search, Cloud and general AI. Robotics can be seen as a new growth vector that is not directly dependent on advertising and has a tangible "physical" impact.

2) The "Android model" in robotics

DeepMind's key strategic bet is that Gemini Robotics will become the horizontal standard for the "brain" of robots:

• A partnership with Boston Dynamics (Atlas, Spot).

• Apptronik (Apollo).

• Agility Robotics (Digit) and others.

• And the ability to offer Gemini Robotics via API to other manufacturers.

For the investor, this means an asymmetric bet - if one robot manufacturer succeeds, Google is there; if more succeed, network effects and standardization will further strengthen its position.

3) Branding and defending against competition

The partnership with Boston Dynamics also has a symbolic level. Boston Dynamics is seen as one of the icons of robotics, Hyundai as a major industry player. With Gemini and Google DeepMind branding sticking to their humanoid Atlas, it reinforces the perception of Alphabet as a leader in "embodied AI" as well, not just in text and images.

In a competitive landscape where Microsoft is betting on Copilot and OpenAI, Tesla on its own Optimus, and Nvidia on an AI stack for automotive and robotics, these partnerships help Google present itself as a company that has a clear role in the world of physical AI systems as well.

Alphabet vs Tesla vs others in robotics

Alphabet / Google DeepMind

• Strategy: be the "brain" of robots via Gemini Robotics - a universal AI layer for various hardware partners (Boston Dynamics, Apptronik, Agility, etc.).

• Strengths: high-end models (Gemini), cloud, and ability to learn from data across partners without bearing the cost of making robots.

• Weakness: dependence on partners to actually manage to produce and sell robots in bulk; Alphabet does not have its own "flagship" robot brand.

Tesla $TSLA

• Strategy: end-to-end approach with humanoid Optimus - Tesla does hardware, software, manufacturing, data and deployment in its own factories, all "in-house."

• Strengths: extreme integration (AI from autopilot, custom chips, automotive manufacturing experience), clear opportunity to deploy tens of thousands of robots internally first.

• Weakness: high capital requirements, technology and security risk, dependence on one company - if Optimus fails, there is no "plan B" on other partners.

Others (Figure, Agility, others)

• Strategy: build a "Tesla-like" approach on a smaller scale - own humanoid + own AI, often with support from large partners (e.g. Amazon at Agility).

• Strengths: agility, single product focus, rapid experimentation with business models (robot-as-a-service, specific verticals).

• Weakness: limited capital, less access to data, high risk of ending up as an acquisition target or being rebuilt by big players.

What this means for an investor in Alphabet

• Alphabet is betting on a horizontal role - it doesn't want to be the "Tesla of robotics" but rather the Android and Google Cloud for the robot world. This reduces capital risk and increases the chance of being with multiple winners at once.

• If Tesla or one of the "small" ones succeeds but takes another stack as the AI brain, Alphabet will lose some of the opportunity - which is why partnerships like Boston Dynamics are key to securing a "seat at the table".

• For Alphabet's valuation, robotics is small for now, but strategically important - it can show that Gemini and Google Cloud can make money outside of traditional software and advertising.

Risks for Alphabet

• Dependence on partners - unless Boston Dynamics and other manufacturers can scale robots, Gemini Robotics will remain marginal in robotics.

• Pace of adoption - enterprises may be cautious about deploying humanoids and AI agents in mission-critical operations, which will slow monetization.

• Competing AI stacks - Microsoft, Nvidia and others may offer alternative solutions for robotic AI, reducing the chance of an "Android moment" in robotics.

• Regulation and Safety - Incidents with AI-driven robots may lead to regulations that will require costly adjustments to models and systems.

Investment scenarios for Alphabet

Optimistic scenario

Gemini Robotics will become the de facto standard for multiple robotics platforms. Alphabet earns hundreds of millions to units of billions of USD annually from licensing and cloud usage in high-margin robotics. The market will start to reflect this as another growth pillar alongside cloud and general AI, which will support rerating multiples, especially if data advantage and network effects also come into play.

A realistic scenario

Gemini Robotics establishes itself with several major partners (Boston Dynamics, Apptronik, Agility), but the market will remain fragmented. Alphabet makes a "nice" hundred million USD a year in robotics, which is supported by Google Cloud, but it's not a game changer for the group. Robotics is helping to strengthen the perception of technical leadership and diversification rather than dramatically changing the financial profile.

Pessimistic scenario

Robotics turns out to be a smaller and slower-growing business than the videos and presentations suggest today. Hardware partners struggle with costs, regulation and low demand. Gemini Robotics will remain a cool technology with limited commercial impact. For Alphabet shareholders, this is more of a cosmetic issue - the core AI and cloud story remains, but the "third pillar in robotics" will not materialize.

What an investor in Alphabet should take away

• The partnership with Boston Dynamics gives Gemini Robotics a physical body and access to real industry data without Alphabet having to build its own hardware.

• Google is looking to replicate the Android model in robotics: be the standard AI layer on top of various robots, collect data and charge for cloud and intelligence.

• In the short term, these will be more like small contributions to revenue, but in the long term, robotics may become another high-margin segment beyond Search and cloud.

• The risk to shareholders is limited - the core business remains elsewhere - but the potential upside if the horizontal strategy in robotics succeeds is intriguing.

CrowdStrike announced a global partnership with Nebius aimed at integrating the Falcon platform directly into the Nebius AI Cloud and bringing unified enterprise cybersecurity to new, high-performance AI environments. The joint solution offers unified visibility and AI-powered detection and response across infrastructure and runtime environments, enabling companies to scale AI without disrupting existing security architecture.

The integration is intended to allow customers to extend existing security policies and response procedures to AI workloads running in Nebius. For investors, this means strengthening CrowdStrike’s position in the rapidly growing AI infrastructure security segment, with the potential for access to new customers and greater adoption of the Falcon platform.

The company of billionaire Bernard Arnault is trading 47% below its 2023 high.

The decline is due to a combination of factors. LVMH reported 2025 revenue of €80.8 billion, a 5% year-on-year decrease. The key Fashion & Leather Goods division, which includes brands like Louis Vuitton and Dior, recorded an organic 5% drop in revenue and operating profit fell by 13%. The group's net profit plunged 13% to €10.9 billion.

Demand in Asia remains weak. This is compounded by consumer fatigue from repeated price increases and a shift of younger customers toward brands like Prada or Miu Miu.

The positive news is growth in selective distribution thanks to Sephora and stabilization in the watches and jewelry segment.

CEO Arnault, however, warned that 2026 will be challenging.

Do you hold shares of this fashion giant in your portfolio, or is it a no-go for you?

The S&P 500 index, which reached its all-time high of 7,014 points on 28 January 2026, has been losing ground since then. As of 17 March, it is trading around 6 700 points, a decline of approximately 4.7% from its peak. On the face of it, this is a standard correction that normally takes place in the markets. However, a closer look at both technical and fundamental indicators shows that the current situation may herald a much deeper decline than most investors currently expect.

In fact, several warning signs that have repeatedly appeared in history before significant declines are meeting in the market at the same time. From extremely stretched valuation indicators to technical patterns on the charts to the geopolitical shock of the Iranian conflict that has driven oil prices above $100 per barrel. It is this combination that is creating an environment in which a 30% decline scenario, while still a less likely option, is certainly not unrealistic.

The upper trend line of the S&P 500: a technical signal that cannot be ignored

One of the most obvious warning signals is the behaviour of the S&P 500 index against the long-term upper trend line of the rising channel that has been forming since 2018. The index touched resistance in the region of around 7,000 points in January 2026, which corresponds to the upper limit of this channel. Historically, the market has repeatedly bounced down from this line and significant corrections have followed.

Chart of the S&P 500 index - weekly

Source.

What the historical bounces from the upper trend line say

Technical analysis shows that the S&P 500 has been respecting a rising channel pattern since 2018. Each time the index has approached the upper trend line of this channel, a decline has followed. In January 2022, contact with the upper boundary preceded a bear market in which the index wrote off over 25%. A similar scenario has played out in previous cycles. Currently, the index is in a zone where selling pressure has historically outweighed buying interest.

Moreover, the index has fallen below the lower boundary of the medium-term uptrend channel, which technical analyses interpret as a signal of slowing growth momentum. Key support lies around the 6,130-point level, while resistance is located at 7,000 points. At the same time, the RSI (Relative Strength Index or RSI) indicator shows a downtrend, which tends to be an early precursor of a price decline.

Buffett indicator: record 230%

One of the most followed valuation indicators, the so-called Buffett Indicator, is currently at 230%. This indicator measures the total market capitalisation of US stocks relative to US gross domestic product. The long-term average is around 80% to 100%, with values above 115% traditionally indicating an overvaluation of the market. Warren Buffett himself has stated in the past that this is probably the best single metric for assessing market valuations.

The current level of 230% is a record high in the history of this metric and far exceeds the values achieved before the bursting of the tech bubble in 2000 or the bear market in 2022. The key point is that in all three previous instances where the Buffett Indicator was within two standard deviations of its historical trend, the S&P 500 Index followed a decline of at least 25%. The 1968 decline was over 30%, the 2000 tech bubble was followed by stocks losing over 50%, and the 2022 bear market followed in 2021.

Source: Current Market Valuation

Shiller CAPE ratio: second highest in history

Another warning sign is the Shiller CAPE ratio, invented by Nobel Prize winner Robert Shiller. This ratio measures the current price of the S&P 500 index relative to average inflation-adjusted earnings over the past 10 years. Currently, the CAPE ratio is around 37 to 40, while the long-term average is about 17. The only time in history that this ratio has been at a comparable or higher level was during the technology bubble at the turn of the millennium.

Historical data show that after reaching a CAPE above 39, the S&P 500 index declined by an average of 20% in the following two years and 30% in the following three years. Moreover, the index has never generated a positive three-year return in such a situation. This does not mean that the downturn has to come immediately, but the probability of below-average returns over several years is very high from a historical perspective.

Source: Multpl

A critical look at valuation indicators

It should be added, however, that valuation indicators have their limits. Some analysts point out that structural changes in the economy, such as the expansion of high-margin technology companies, share buybacks or the globalization of corporate earnings, may justify consistently higher valuations than in the past. According to these arguments, indicators such as the CAPE ratio or the Buffett indicator systematically overshoot in a pessimistic direction because they do not take into account changes in the profitability of U.S. firms over the past 30 years. Even so, current valuations are at extreme levels even after accounting for these structural shifts.

Midterm election year: historically the weakest phase of the presidential cycle

The year 2026 is a midterm (midterm) election year for the US Congress, and historical data shows that these are among the most volatile years from a stock market perspective. There have been 17 midterm cycles since the inception of the S&P 500 index in 1957. In 12 of them, approximately 70% of the time, the index has fallen into correction territory, i.e. down 10% or more. The average maximum intra-year decline was approximately 18%.

Midterm elections create uncertainty because the party in power almost always loses seats in Congress. Investors do not know whether current policies on taxes, regulation or trade will continue. Financial markets react negatively to uncertainty. In 2026, this effect is amplified by President Trump's tariff measures and geopolitical tensions in the Middle East, which add additional layers of risk to the standard cyclical pattern.

On the other hand, historical data also show that the six months following the midterm elections, from November to April, are among the strongest periods of the entire four-year presidential cycle. The S&P 500 Index added an average of 14% during that period. But the key is to weather the volatility until then, and that is where the main risk for investors lies.

Geopolitical shock: war with Iran and oil shock

Adding to the fundamental and technical warning signs since the end of February 2026 has been the acute geopolitical factor. US and Israeli military operations against Iran have led to the closure of the Strait of Hormuz, through which around 20% of the world's oil supply passes under normal conditions. Brent oil prices have exceeded USD 100 per barrel for the first time since 2022.

The oil shock fundamentally complicates the Federal Reserve's position. Higher energy prices increase inflationary pressures while slowing economic growth, a combination known as stagflation. Stagflation was a major factor behind the dramatic stock market declines of the 1970s, when the S&P 500 index lost over 40% during the OPEC oil crisis of 1973. The Fed thus finds itself in a situation where it cannot simply cut rates to support the economy, because to do so would risk further unleashing inflation.

We will learn more at tomorrow's Fed meeting and in the commentary that follows.

Why the market hasn't reacted more strongly so far

Despite the closure of the Strait of Hormuz (which is now only open to select countries) and the sharp rise in oil prices, the S&P 500 has only fallen about 4.7% so far since its January high. Investors seem relatively calm and are betting that the conflict will be short-lived.

Moreover, the US is one of the world's largest oil producers, which reduces its direct dependence on imports. However, it is important to remember that oil prices are global and any increase in them will be reflected in the costs to companies and consumers, regardless of the level of domestic production. If the conflict persists longer than the market expects, the current relative calm could quickly turn into panic.

Market concentration and extreme sentiment

Another factor that increases the risk of a significant downturn is the extreme concentration of the U.S. stock market. Just five companies, Nvidia $NVDA, Microsoft $MSFT, Apple $AAPL, Google $GOOG and Amazon $AMZN, account for almost 30% of the entire S&P 500 index. This level of concentration is reminiscent of the situation before the tech bubble burst, when a narrow group of stocks pulled the entire index up while the rest of the market lagged.

Moreover, investor sentiment is extremely bullish. The NAAIM survey shows that institutions are allocated to stocks at levels corresponding to the 78th to 96th percentile of all historical values. Retail investors hold minimal cash and are heavily invested, according to the AAII survey.

It is this combination of high allocation and low cash that historically has repeatedly occurred prior to major market peaks. A Natixis survey of 515 institutional investors managing nearly $30 trillion in assets showed that 79% expect a market correction in 2026, with 49% predicting a 10% to 20% decline and 20% anticipating an even deeper decline.

Forecast comparison: from a slight correction to a 40% decline

Analyst houses differ significantly in their estimates of the potential downturn. On one side of the spectrum stands Goldman Sachs $GS, which estimates a 25% probability of a recession with a potential 15% drop in the index. On the other side, BCA Research considers a 60% probability of a recession and projects a drop in the S&P 500 to the 4,200 to 4,500 point range, which would represent a drop of around 35% to 40% from current levels.

The options market data adds further perspective. S&P 500 put option prices as of the end of 2025 imply an 8% probability that the index will fall 30% or more during 2026. This may seem low, but for context it should be added that this is well above average and includes a risk premium to reflect market uncertainty.

Overview of warning signs

Strategy

The current market situation resembles a classic scenario in which multiple risk factors meet at the same moment. Extreme valuations, geopolitical shock, the cyclical pattern of the mid-year and high investor sentiment create an environment in which the scope for further upside is limited and the risk of a significant downside is above average.

For investors, several practical conclusions follow. Building a cash position in a portfolio can provide room to buy in the event of a sell-off. Holding only high-quality stocks with strong fundamentals reduces the risk of permanent capital loss. Diversifying outside of U.S. stocks into areas that are not as extremely valued can help mitigate the impact of a potential correction.

At the same time, it is important to remember that the mere existence of warning signs does not mean that a 30% decline is inevitable. Valuation indicators have a limited ability to predict the exact timing of corrections and markets may remain overvalued for longer than most analysts would expect. Should the Iranian conflict be resolved quickly and oil prices fall back, one of the main catalysts for risk could weaken significantly. Similarly, if US corporate profitability remains strong, high valuations could remain justified at current levels.

What to watch next

• Developments in the conflict with Iran and oil prices: if Brent exceeds $120 per barrel and stays above that level, the likelihood of a stagflation scenario increases significantly

• Fed interest rate decision: the market currently expects the first rate cut in September 2026 at the earliest, but the oil shock could push this date even further out

• Earnings season for Q1 2026: if companies start lowering their outlook due to higher energy costs, it could trigger a revaluation

• Market breadth: if the downturn starts to spread from tech giants to the broader market, this is a strong signal of a deeper correction

• VIX index: current readings around 25 to 30 do not yet indicate panic, but a rapid rise above 35 would signal a significant increase in market stress

• Midterm elections in November 2026: historically, markets stabilize after political clarity, and the six months following the midterm elections are among the strongest of the cycle

Possible future developments

The scenario of a 30% decline in the S&P 500 index is not the baseline scenario of most analysts, but it cannot be described as unrealistic. The combination of historically extreme valuations, a technical rebound from the upper trend line, geopolitical shock in the form of the Iranian conflict, a cyclical pattern of mid-years, and extremely bullish sentiment creates an environment that has repeatedly preceded significant corrections throughout history.

The key for investors is to remain rational and not succumb to either over-optimism or panic. Markets have always recovered from downturns and investors who have maintained a long-term perspective and had sufficient liquidity to buy during sell-offs have typically emerged from crises stronger. However, the current environment clearly requires greater caution and a more thoughtful approach to portfolio management than the last three years of unprecedented growth.

Moreover, GTC is turning from a conference show into something like an "annual reckoning for AI"—watched not only by investors but also by companies building their businesses on Nvidia hardware, from cloud providers to telcos and biotech. Every new chip and platform (Blackwell, Rubin, DGX, new AI servers) is tied to specific projects and contracts with major players, so it's not just a "story" but a pipeline of real investments into datacenters with a multi-year horizon.

The question for an investor now isn't just "is it a bubble?", but rather: will those who already own Nvidia keep buying more based on successive bullish theses from GTC, or is the current growth pace and valuation more a reason to "just hold" and wait to see that trillion-dollar outlook translate into revenue and EPS numbers in the coming years? How about you—are you more in buy-on-catalyst mode, adding on these catalysts, or are you just holding and watching to make sure it doesn't become too large a weight in your portfolio?

On the commercial side, Nvidia is deepening ties with hyperscale buyers. AI cloud provider Nebius signed a five year deal with Meta worth up to 27 billion dollars, built around one of the first large scale deployments of the Vera Rubin platform, with 12 billion dollars of dedicated capacity and up to 15 billion in additional compute Meta can claim if Nebius does not sell it to other customers. At the same time, reports suggest Meta is weighing layoffs of around 20 percent of its workforce to help fund an AI infrastructure budget that could reach 600 billion dollars by 2028, underlining the double edged nature of this boom: huge orders for Nvidia and its partners, but intense efficiency pressure on the buyers footing the bill.

Groq 3 and five new racks: how Nvidia is putting together an AI datacentre

A key product innovation is Groq 3, a dedicated chip for inference, i.e. running AI models in production. Nvidia's $NVDA follows it up with an acqui-hire: it struck a licensing deal with Groq and poached founder Jonathan Ross, president Sunny Madra and other key people in a roughly $20 billion package. Groq 3 complements Nvidia's classic GPUs so that the company has a dedicated inference chip for a time when the market's center of gravity is shifting from training to deployment models.

In parallel, Nvidia unveiled five new server racks based on the Rubin platform. The latter combines six key chips - Vera CPU, Rubin GPU, NVLink 6, ConnectX, BlueField and Spectrum 6 - to reduce training and inference costs by up to a tenth compared to the previous Blackwell generation, thanks to greater efficiency and fewer GPUs per system. From an investor's perspective, this means Nvidia wants to be not just a supplier of individual chips, but the architect of an entire AI datacenter, which strengthens its negotiating position with both clouds and hyperscalers.

Space as the next frontier: the Vera Rubin Space Module

GTC also brought a symbolic innovation: Vera Rubin Space Module, a modular platform for on-orbit datacenters, geospatial intelligence and autonomous operations in space. Nvidia is building on the success of startup Starcloud, which brought the H100 GPU into orbit in 2025 and launched Google Gemini and NanoGPT-based models on it.

According to Nvidia, the Ruby GPU is expected to offer up to 25 times more AI performance for "space-based inference" compared to the H100, paving the way for real-world processing of large volumes of data directly on satellites. The Vera Rubin Space Module platform will also be complemented by IGX Thor and Jetson Orin systems for other types of orbital tasks. From a business perspective, this is not a high-volume segment like traditional datacenters in the short term, but a high-margin, technologically prestigious showcase that reinforces Nvidia's brand in the space and geointelligence industry.

AI agents, NemoClaw and the fight for "desktop"

At the software level, Nvidia announced NemoClaw, a security and control layer for the OpenClaw platform that enables AI agents to operate on users' desktops. Originally listed as Clawd, then renamed Moltbot and finally OpenClaw, OpenClaw allows agents to operate over various AI models and act on behalf of the user in WhatsApp, Discord, Slack and other applications.

The problem with OpenClaw is privacy and security concerns, as the agent can control the computer and access personal data. NemoClaw aims to address these concerns by adding a set of tools for rights management, auditing and security sandboxes. Nvidia is also positioning its GeForce RTX, RTX Pro Station, DGX Station and DGX Spark platforms as the preferred hardware to run these agents, bridging the consumer and professional segments.

Uber, Lyft and others: AI chips as the brains of robotaxi networks

There was also news from the autonomous mobility sector at GTC. Nvidia announced that Uber will begin deploying a fleet ofLevel 4 autonomy vehicles based on its Drive Hyperion platform in Los Angeles and San Francisco in 2027. This is a follow-up to a previous agreement to acquire up to 100,000 vehicles on the platform, this time with specific timelines and locations.

In addition to Uber, Lyft, Bolt and Grab are also using Nvidia's systems for their self-driving projects. From Nvidia's perspective, it's further confirmation that its chips and software are making inroads into autonomous transportation, diversifying revenue beyond traditional datacenters while creating a long-term chip business directly in the automotive sector.

Nebius - Meta: 27 billion contract as an advertisement for Vera Rubin

The biggest financial number around GTC came not directly from Nvidia, but from its ecosystem. Cloud partner Nebius announced a five-year deal with Meta Platforms for up to $27 billion to deliver AI infrastructure based on Vera Rubin. The structure of the contract is two-tiered:

• $12 billion of firmly contracted dedicated capacity across multiple sites.

• Up to $15 billion of optional capacity, which Nebius initially offers to other customers and Meta will buy if it remains unused.

Nvidia has announced a $2 billion investment in Nebius, underscoring its interest in developing the partner and using it as a showcase for Ruby. At the same time, Meta's AI strategies talk about possible AI CAPEX of up to $135 billion in 2026, so the Nebius deal is just part of a broader move to an "asset-light" model of buying capacity. What's important to Nvidia investors is that Rubin is getting into large production deployments in the first wave, and that Nvidia is participating in this both through chip supply and through an equity stake in Nebius.

How big can Nvidia's contribution to growth be

Huang's new outlook of $1 trillion in AI chip demand by 2027 builds on last year's figure of $500 billion for the period through 2026. The company also revealed in its Q4 results for fiscal year 2026 that data center revenue reached $62.3 billion in the quarter and accounted for over 90% of total revenue. GTC 2026 is meant to signal to investors that growth is not just about one generation of GPUs, but the entire product and software layer from data center to automotive to edge and orbit.

The contribution of each news item can be summarized as follows:

• Groq 3 and Ruby ensure that Nvidia has the hardware for the next phase of AI (agent systems, massive inference) and can continue to push the price per token down.

• NemoClaw and OpenClaw open up a new area of software revenue and strengthen the link to end users and businesses.

• Uber and other transportation partners are expanding Nvidia's presence in the auto sector.

• Vera Rubin Space Module shows new segments with potentially high added value.

• Nebius - Meta contract demonstrates that big players are willing to tie tens of billions of dollars to infrastructure based on Nvidia's new platform.

For investors, Terafab marks a sharp turn from a fabless model into deeper vertical integration, closer to an integrated device manufacturer in a field dominated by specialists like TSMC. Bringing part of AI5 production under Tesla’s roof could reduce supplier risk and give more control over a core technology, but it also adds a new layer of capital intensity and execution risk, with estimates of tens of billions of dollars required to build a 2 nanometer capable facility and the challenge of ramping a world class fab on top of an already ambitious EV and robotics roadmap.

What we know about Terafab and AI5 today

The official information so far suggests a few certain things. Terafab is supposed to be a project aimed at making AI chips for Tesla $TSLA, Musk describes it as a "gigafactory, only much bigger", and the project's launch is expected to come within a week of the announcement. Tesla hasn't commented on the details - where the fab will stand, what process it will use, the specific investment volume and timeline.

The fifth-generation AI5 chip, according to earlier reports, is designed to be a significant leap over current HW4 hardware, with an emphasis not only on raw computing power, but also on memory and efficiency for the functionality of modern AI models. Musk has previously confirmed that AI5 will be produced simultaneously by both TSMC and Samsung to reduce Tesla's reliance on a single foundry and ensure a "surplus of AI5 chips" for both cars and data centers. This makes it all the more clear that Terafab will not be the sole source of chips from the start, but rather a complement to external capacity.

Musk has also mentioned the possibility of working with Intel $INTC in the past and has talked about the value of having serious discussions with Intel. However, there is no confirmed contract yet and everything remains at the level of speculation.

Why Tesla is pushing for its own chipset

From Musk's perspective, the key driver is a lack of capacity at suppliers. He said last year that even if he takes the best-case scenario of supply from TSMC $TSM and Samsung $SSNLF, "it's still not enough" given the projected volumes for cars, robotaxi fleets and Optimus robots.

The main reasons why Terafab makes sense from Tesla's perspective:

• Control over a key input - AI chips are as strategic to Tesla today as batteries.

• Long-term demand - if Tesla is to really roll out robotaxis and physical AI in a big way, it will need millions of chips a year, not just for cars, but for data centers.

• geopolitics and supply risk - diversifying beyond a pure Asian foundry model may be a hedge against tensions around Taiwan or other supply shocks.

On the other hand, building your own chip fab means entering an industry where barriers to entry are extremely high: technically, financially and organisationally. Even established players like Intel have struggled to catch up with TSMC, despite decades of know-how.

The impact on Tesla's business model

If Terafab succeeds, Tesla would move from the role of a buyer to an integrated player in the AI chip chain. This would have several implications:

• Some of Foundry's margin would remain within Tesla.

• The company would gain more flexibility in how quickly it scales Robotaxi, FSD and Optimus.

• it could eventually offer its own chips or manufacturing capacity to third parties (this is purely hypothetical for now, but model-possible).

But in the short term, it will mainly be a source of cost and CAPEX. Tesla is already planning to significantly increase capex due to the transition from a pure EV story to "physical AI". The addition of a giant chip plant will further inflate this bill and may deepen the period of negative free cash flow that some analysts are already talking about for 2026.

Crucially for investors, this increases the sensitivity of Tesla's story to execution. It's not enough to just deliver cars and software, it adds another critical layer - high-end chip manufacturing - in which there is world-class competition and where mistakes are very expensive.

The risks

Short-term risks:

• Project ambiguity - Tesla doesn't comment on details, so panels on how much Terafab will cost, where it will be, and what process it will use are still guesses. The vagueness may increase nervousness in the market.

• Investment overlap - simultaneously running investments in robotaxi, Optimus, capacity expansion at existing plants and now Terafab. This may exacerbate the perception of cash flow risk.

Medium-term risks:

• Technology slippage - approaching the technology level of TSMC and Samsung is extremely challenging. If Terafab falls one or two process generations behind, it may limit its usefulness for the most advanced AI chips.

• Cost - in-house fabs can have higher unit costs than outsourcing to high-end foundries, especially in the early years.

Long-term risk:

• Either Terafab will become a functional part of Tesla's vertically integrated AI platform and support its lead in autonomy.

• or it turns into a capital-intensive experiment that can't keep up with the foundry market leaders and remains permanently dependent on TSMC and Samsung while investors bear the cost of building it.

What to watch next

From an investor perspective, the following triggers will be key:

• First specific information about Terafab' s location, technology and budget - whether it is a full-fledged leading-edge project or more of a complementary capacity.

• How Tesla will align the timing of Terafab with the start of AI5 mass production at TSMC and Samsung - whether the fab will be a follow-on or just complementary to the supply already running.

• How the new CAPEX will play out in free cash flow projections and whether Tesla will communicate a clear ROI framework.

• what role Terafab will get in the Tesla AI Day / Investor Day communications - as a side project or as one of the pillars of the physical AI strategy.

One of the sector’s leaders now finds itself exactly between these forces. It combines solid revenue growth, a double digit operating margin and roughly a one fifth share of the global immunoglobulin market with higher debt and lingering reputation damage after a short seller attack questioned its numbers. The stock trades for less than one times annual sales and at a single digit multiple of operating profit, which makes it look like a value story inside healthcare, but it is not a simple safe haven for anyone who is not willing to monitor leverage, cash generation and governance very closely.

Top points of the analysis

• The firm's revenues have grown from roughly €4.9bn to €7.2bn over four years, which equates to average growth of around 6% a year with a few stronger years.

• Operating margin is around 18%, gross margin around 39%, but net margin is only around 5%, reflecting high interest costs and tax burden.

• The company is one of the top 3 global producers of plasma drugs and holds a market share of around 20% of the global market in immunoglobulins.

• For valuation, we are looking at a combined P/E of around 12-13, EV/EBITDA of around 9-10 and P/S of around 0.8, which is significantly lower than major competitors like CSL or Takeda.

• The plasma market as a whole is growing at a rate of around 7% per annum due to the growing demand for treatment of immunodeficiencies and other complex diseases, with Grifols controlling the entire chain from plasma procurement to final drugs through vertical integration.

• The investment thesis is that new management can tame debt, stabilize margins and restore confidence after the brief Gotham City assault, while the market still values the company at a significant discount to its market position.

Company performance

Grifols $GRFS is a Spanish healthcare company based in Barcelona that specializes in plasma therapeutics and related diagnostics. At its core is the so-called vertically integrated plasma chain: the company operates its own networks of collection centres, processes plasma in large fractionation plants and produces protein products such as immunoglobulins and albumin from it for the treatment of rare and chronic diseases.

According to the available materials, Grifols operates in four main segments: Biopharma (plasma drugs), Diagnostics (diagnostic tests and devices for transfusion stations and hospitals), Bio Supplies (biological materials and services for the pharmaceutical and biotechnology industries) and Healthcare Solutions (solutions for hospitals and clinics). However, the vast majority of sales and profits come from plasma pharmaceuticals, where the company is one of the leaders.

Geographically, Grifols is a global player, but the key market remains the United States, where it operates an extensive network of plasma centers and where the plasma drug market is the largest and most profitable. Europe, Latin America and Asia are then served by a combination of its own facilities and distribution partnerships. Thus, despite its European origins, the company's economics are heavily "dollarized" and dependent on the U.S. healthcare system.

Business and products

The core of Grifols' business is the collection of plasma from donors and its conversion into life-saving products. Key products include:

• Immunoglobulins (powder or liquid form for intravenous or subcutaneous administration) for the treatment of primary and secondary immunodeficiencies and certain autoimmune diseases.

• Albumin used in the treatment of shock, burns, cirrhosis and other conditions where there is a need to replenish volume and protein in the bloodstream.

• Specialized hyperimmunoglobulins for specific indications such as Rh incompatibility, cytomegalovirus or tetanus.

Through vertical integration, Grifols controls the entire process from donor to final product: it operates the collection centers, logistics, laboratory testing, fractionation (separation of individual proteins from plasma) and final packaging and distribution. This model increases the gross margin (around 39%) while helping to ensure a stable supply in an industry where quality and safety are essential.

The other leg is diagnostics, where Grifols supplies instruments and reagents to blood banks and hospitals: systems for donor testing, blood grouping, virological screening and other tests. This segment has a lower share of revenues than Biopharma, but it brings in steady recurring revenues and strengthens the company's position with blood centres, which feeds back into the plasma business.

Market and addressable potential

The market for plasma drugs is large globally and is growing at a rate above that of conventional pharmaceuticals. According to available studies, the plasma therapeutics market is worth tens of billions of dollars and is expected to grow at a rate of around 7% per year through 2030, driven by increased diagnosis of primary immunodeficiencies, an aging population, and expanding indications. In addition, the albumin segment has a growth projection of around 7.7% per year, driven by higher incidence of major diseases and new, more practical product forms.

Grifols is estimated to be among the top three global manufacturers of plasma drugs alongside CSL and Takeda, and is estimated to have around 19% share of the plasma raw material market and around 20% share of the immunoglobulin market, the most lucrative segment, in 2024. This means that the company is not a "small player" but a structural part of the market that cannot do without its capacity.

The addressable market for Grifols is thus not a question of "if it exists", but rather how much share it can maintain and expand, what the pricing discipline in the sector will look like, and how much of the growth will translate into margins. In a realistic scenario, Grifols can benefit from market growth in the 5-7% per annum range and a slight shift towards higher value-added (specialised immunoglobulins, new indications) if it maintains its position in the top 3.

Competition and market position

The global plasma market is dominated by three major players: Grifols, CSL (CSL Limited) and Takeda $TAK, all of which operate extensive networks of collection centres and large fractionation plants, competing for donors, distribution channels and reimbursement contracts.

• CSL tends to be seen as a quality and efficiency assured company, with very high margins and a more conservative balance sheet.

• Takeda combines its plasma portfolio with its broader pharmaceutical business, giving it diversification but also a different capital profile.

• Grifols has a strong position in immunoglobulins and high chain integration, but also higher leverage and historically more aggressive M&A, which has weighed on the balance sheet.

The difference is not whether they can make the products, but how efficiently they do it. Grifols has the advantage of a large plasma base and global access, but pays a higher price in debt and lower ROIC (ROIC around 4.8%). CSL-type competitors generally have higher margins and better debt-to-earnings ratios, which is reflected in higher valuation multiples.

Grifols wins where its vertical integration, ability to secure supply and offer a broad portfolio of plasma drugs come into play. It may lose out in an environment where the market focuses more on balance sheet quality and return on capital than just sales size and market share.

Management and CEO

As of 2024, the company is headed by José Ignacio (Nacho) Abia Buenache, who came from Olympus (medical devices), where he led the US division as President and CEO for more than a decade, while also holding operational and strategic director positions in the global group. That said, the new CEO brings experience of running a large medtech firm, an emphasis on operational efficiency and capital discipline.

His joining Grifols in 2024 is a clear signal that the firm is looking to strengthen management after a period dominated by family influence and a more aggressive expansion strategy. The new management's priority, according to investor communications, is a combination:

• Deleveraging (selling assets such as the stake in Shanghai RAAS).

• Improving accounting and reporting transparency after the short-seller attack.

• And an emphasis on more profitable growth, not just volume expansion.

Ownership structure remains partly family, but a significant stake is held by institutional investors, who are overseeing the company after several tense episodes (Gotham City, debt refinancing). For investors, it is significant that a manager accustomed to "public company" discipline is at the helm and that the pressure to improve the balance sheet is explicit.

Financial performance

Grifols' revenues have grown from roughly €4.93bn to €7.21bn over the past four years, a solid increase of over €2.2bn, with the strongest year immediately following the covenant. Gross profit rose from around EUR 2.0bn to EUR 2.8bn, with gross margin holding around 39-40%.

Operating profit increased from around EUR 595 million to EUR 1.19 billion, almost doubling, and the operating margin is around 18%. This reflects both the scaling of the plasma business and better control of operating costs, although 2023 was a transitional year with a slight stagnation in operating profit.

Net profit is significantly more volatile. In recent years it has ranged between around €40m and €190m, with 2024 bringing a significant jump back to a higher level of profitability after a weak 2023. Part of the volatility is due to finance costs and taxes, part to one-off effects (share sales, revaluations, restructuring costs). This is reflected in EPS, which has higher volatility than revenue itself.

Cash flow and capital discipline

From an investor perspective, cash flow from operations and free cash flow are key. Grifols generates decent operating cash flow, which is consistent with it being a capital intensive but profitable business. However, after accounting for investments in capacity, plant upgrades and research, there is relatively limited free cash flow, which is reflected in the debt statistics.

The company pays a modest dividend of around $0.14 per share per year, a conservative amount given its debt. The main channel for capital allocation is investment and debt reduction, not high payouts to shareholders. This is logical in the current situation: as long as net debt is around 5.9 times EBITDA, it makes more sense to strengthen the balance sheet than to aggressively raise the dividend or buybacks.

Capital discipline has also been in the spotlight in recent years due to criticism of the accounting and debt structure from short-sellers, particularly Gotham City, who have pointed to the complexity of the structure, interconnected transactions and the risk of undervaluing debt. The company's response has been to take several steps to divest assets (such as a significant stake in Shanghai RAAS), prepare for refinancing and seek to improve communication with investors.

Balance sheet and debt

The balance sheet is Grifols' main red flag. The debt-to-asset ratio is around 0.49, and the debt-to-equity ratio is around 1.87, which means that the company is highly leveraged. The net debt to EBITDA of around 5.88 is well above the level that a healthcare investor would ideally like to see, and is more reminiscent of the profile of a "leveraged" business (financed by debt) than a conservative pharmaceutical player.

Interest coverage of around 2.6 times shows that operating earnings cover interest, but not with much cushion. This is acceptable in an environment of stable rates and rising sales, but increases sensitivity to any margin pressure or increase in financing costs. An Altman Z-score of around 1.1 already signals elevated credit risk: the company is not in an acute crisis, but it is certainly not in a safe zone where debt is a marginal issue.

Adding to this is the reputational dimension: an attack by Gotham City short-sellers in early 2024, who questioned the reporting of certain debts and transactions, caused the stock to plummet and brought questions over accounting transparency to the surface. Grifols responded by selling its stake in Shanghai RAAS for about $1.8 billion and preparing its first major bond financing since then to refinance maturing debt and confirm access to the capital market.

Valuation and valuation interpretation

At today's levels, Grifols is trading at around 12-13 times earnings, EV/EBITDA of around 9.7 and P/S of around 0.8, with a price-to-book value ratio of around 1.1. This means the market is valuing the company at a significant discount to its size and position in the plasma sector, especially compared to CSL or Takeda, which typically trade at double-digit EBITDA multiples and higher P/B.

Grifols' valuation thus looks attractive to an investor willing to accept credit risk and more complex accounting. On pure multiples, the company is cheap given that it has double-digit operating margins, a global footprint and is part of the growing plasma market.

For a "cheap" valuation to turn into an "expensive" one, there would have to be a significant deterioration: a decline in sales, margin compression, and refinancing pressure that would fool investors into believing the firm could handle the debt load. Conversely, for the discount to the sector to narrow, Grifols must show a combination in the coming years:

• Steady revenue growth in the range of at least 5-7%.

• A gradual deleveraging (net debt/EBITDA towards 3-4 times).

• maintaining or slightly improving the EBITDA margin

• and transparent accounting without new controversies.

Growth catalysts and outlook

Positive catalysts include:

• Continued growth in the plasma drugs market, particularly immunoglobulins and albumin, at a rate of around 7% per annum.

• Better utilization of the extensive network of plasma centers.

• and potential price improvements in an environment of higher demand and limited capacity.

On the structural side, there is also the possibility of the company monetizing other non-core assets and using the funds to reduce debt, thereby improving debt ratios and reducing credit risk. Another catalyst may be stabilization after a short-term attack if rating agencies and debt markets confirm confidence in refinancing obligations after 2025.

The industry outlook is relatively clear: demand for plasma drugs will grow, barriers to entry (center network, licensing, regulation) are high, and technologies in plasma processing and logistics are delivering incremental efficiency improvements. The question is not whether the market will grow, but in what proportion and quality Grifols will participate in this growth.

Risks

• Debt and refinancing: high net debt and low Z-scores mean that any deterioration in debt market conditions or weaker years in terms of margins could put the company under pressure. Signals: deterioration in interest coverage, difficulty in placing new bonds, worse credit conditions.

• Accounting and reputational risk: the criticism of Gotham City and subsequent stock swings have shown that the market is sensitive to any doubts about transparency. Further questionable transactions or differences in debt interpretation could further damage reputation.

• Competition: both CSL and Takeda have better balance sheets and often higher margins, giving them greater flexibility to invest in capacity and research. If there were more aggressive pricing pressure, Grifols could have less room to manoeuvre due to debt.

Investment scenarios

Optimistic scenario

Revenues grow 6-8% per year, EBITDA margin stabilizes and improves slightly, net debt/EBITDA declines gradually to 3-4 times due to a combination of growth and debt repayment. EPS grows at a double-digit rate due to operating leverage and lower interest costs. In this scenario, EV/EBITDA could move into the 11-12 range, P/E into the 15-18 area, which at current levels would imply solid price upside in addition to the potential for a modestly rising dividend.

A realistic scenario

Revenues grow 4-6% annually, margins remain stable, debt declines more slowly but remains in a manageable range. EV/EBITDA stays around 9-10, P/E in the 10-14 range depending on the year. The stock in this scenario works more like a gradual "re-rating" bet - the investor gets a return through a combination of gradually rising EPS and a slight narrowing of the discount to the sector.

Negative scenario

Revenues slow below 4% per annum, margins are under pressure (competition, pricing, cost centres) and debt refinancing takes place on worse terms. Net debt/EBITDA does not fall, or even rises, and rating agencies worsen the outlook. In such a scenario, the market could drive EV/EBITDA down to 7-8, P/E loses its telling power due to low or volatile earnings. The stock would become more of a "credit" story than a classic value bet.

What to watch next

• Net debt/EBITDA and refinancing calendar (especially maturities around 2025-2026).

• EBITDA margin and the evolution of operating margin over time.

• Revenues and volumes in key plasma products (immunoglobulins, albumin).

• Steps taken by new CEO Nacho Abii in restructuring, asset sales and increasing transparency.

• Further market and regulatory reaction to Gotham City's accounting wrangles and potential new debt news.

What to take away from the article

• Grifols is a structurally strong player in the plasma drug market with a larger share and double-digit operating margins, trading at a significant discount to major competitors.

• The main cost of this discount is high debt, a low Z-score and reputational scars from the short-seller attack, making Grifols a "value" story with a credit tail, not a pure defensive play.

• A new CEO with a medtech background gives a chance for better capital discipline and a gradually improving balance sheet, but the proof will have to come in numbers, not just presentations.

• For an investor who understands the plasma markets and is willing to monitor debt and refinancing as closely as earnings, Grifols could be an interesting medium-term re-rating bet.

Software sector under fire: what's behind the sell-off of 2026

The year 2026 has brought a very pessimistic AI-related mood to the software sector. Investors who just a year ago were paying premium multiples for companies with fast-growing revenues and predictable subscription revenues are now reassessing those same companies with an unprecedentedly skeptical eye. This turnaround is not primarily driven by poor financial results. Behind it is the fundamental question that AI has posed to the entire sector: will traditional software business models survive the rise of AI agents that can automate the very processes that thousands of businesses have been paying annual subscription fees for?

This existential query hit the sector at a time when it was already carrying macroeconomic burdens. Enterprises are tightening IT spending and lengthening purchasing cycles. And investors, after years of tolerating low profits with high growth, now demand proof of real profitability. The result is an almost toxic combination of valuation overvaluation, cyclical slowdown and structural uncertainty that is dragging the entire sector down.

A telling barometer of this movement is the iShares Expanded Tech-Software Sector ETF$IGV), which has depreciated more than 23 percent since the beginning of 2026. It's already down 28 percent since its peak last fall. That he in direct contrast to the relatively steady performance of the broader S&P 500 index. And it is in this context that the moves of the following five specific companies, whose stocks have been among the biggest disappointments of the entire index this year, must be understood.

Gartner $IT

The advisory model in the AI era: existential question or overhyped panic?

The firm that has dominated the technology research and consulting market for businesses for decades has seen its market value collapse dramatically in 2026. The stock has plunged more than 66% since its 52-week high, and the decline since the start of 2026 is one of the largest in the entire S&P 500. Yet as recently as February 2026, the company reported results that beat analyst consensus in both earnings and revenue.

The immediate trigger for the selloff was the earnings release on February 3, 2026. Gartner reported quarterly revenue of $1.75 billion, up 2.2 percent year-over-year, and adjusted EPS of $3.94, which beat the consensus of $3.51. Still, the stock fell more than 20 percent in a single day.

The reason was the outlook. Management projected full-year 2026 revenue in the range of $6.455 billion to $6.5 billion, while the market was expecting $6.7 billion. The signal was clear. Gartner itself does not believe the outlook of its traditional consulting model is certain in an AI environment.

The structural challenge: when will it stop making sense to pay for analytics reports?

Gartner operates three segments:

• Insights (research and advisory on a subscription basis),

• Conferences

• Consulting.

It is the Insights segment that faces a direct question about its future place in the corporate economy. Companies that used to pay tens of thousands of dollars a year for access to analytics databases and expert recommendations are now testing whether AI tools can provide comparable insights cheaper and faster.

A recent earnings report

Quarterly results confirm this push in the numbers. The Consulting segment reported a 12.8 percent decline in revenue. Global Contract Value, a key leading indicator of future revenue, grew only 0.8 percent year-over-year to $5.2 billion. Outside of the U.S. federal government segment, which is itself experiencing cutbacks, growth was much healthier, at 4 percent, but even that wasn't enough to convince investors that Gartner has the situation under control.

The firm is responding by deploying AskGartner, an AI system that allows customers to more easily search its analytics database. Management reports that users of the tool are reporting significantly higher contract renewal rates. If this trend is confirmed in the coming quarters, AskGartner could be proof that Gartner can integrate AI in a way that enhances, not undermines, the value of its platform. But for now, the market is waiting for more concrete evidence.

The analyst consensus based on the Fair Price Index at Bulios is around a target price of $245, or about 47% upside from current levels, with the valuation pushing the P/E to about 13.8 times earnings, a historically very low level for the company.

AppLovin $APP

From absolute star of 2025 to biggest loser

The company entered the new year as one of the most admired stories in the entire market. In December 2025, the stock hit an all-time high of over $733, adjusted EBITDA margins were at an extraordinary 84 percent, and quarterly revenue grew at a 66 percent year-over-year rate. Since that peak, the stock has depreciated more than 38 percent through mid-March 2026, with a single-day drop of more than 20 percent in February after releasing results that beat analysts' estimates.

Two waves of negative sentiment are primarily behind the sell-off. In January 2026 , the CapitalWatch fund published a report accusing AppLovin' s major shareholder of ties to money laundering operations. The company immediately dismissed the accusations as false and misleading. In February, CapitalWatch retracted the key allegations, issued a formal apology, and the stock briefly rebounded. Still, the incident left a visible mark on institutional investor confidence.

AI as both a threat and an opportunity

A deeper cause of the pressure on AppLovin's valuation is concerns about AI competition. At the heart of the company's business is the AXON advertising platform, which optimizes ad serving in the mobile gaming ecosystem through machine learning. Investors are questioning whether $META Platforms and its increasingly sophisticated AI ad tools could gradually erode the advantage on which AppLovin has built its extraordinary growth. If ad systems with a larger data base can outperform AXON in ad allocation efficiency, the firm's margins could converge toward the sector average under pressure.

The paradox of the situation is that the firm's fundamentals remain exceptional. The fourth quarter of 2025 delivered record revenues, with free cash flow for the full year exceeding $3.95 billion. The company is expanding from mobile gaming into the e-commerce advertising and web space, where management identifies another big opportunity.

Analyst firms like UBS and Jefferies have maintained price targets in the $668 to $860 range despite the sell-off, and the consensus remains at Strong Buy. According to the Fair Price Index, however, the stock is still significantly overvalued despite this year's price declines.

Intuit $INTU

Intuit is one of the companies whose decline in 2026 is most surprising because it stems not from poor results, but from an overvaluation of the market premium. Intuit is a dominant player in tax returns through its TurboTax product and QuickBooks branded small business software. Both platforms have held monopolistic or oligopolistic positions in their verticals for decades, and it is this strength that has historically justified premium valuation multiples.

Shares have depreciated more than 33 percent since the beginning of 2026. In February 2026, they hit a low of around $349, erasing the gains of the previous few years and getting to the same levels they were at in 2022. Yet the company reported second fiscal quarter revenue of $4.65 billion, up 17 percent year-over-year, solidly above consensus, and operating profit up 44 percent year-over-year. Management also reaffirmed full-year guidance for fiscal 2026.

Valuation compression without fundamental failure

The real trigger for the downturn came not from results, but from fear of the future. The market in 2026 is intensely debating whether AI agents capable of processing tax returns or categorizing corporate transactions will threaten the very essence of business. CEO Sasan Goodarzi directly confronts this question, arguing that AI is a catalyst for business, not a threat. As evidence, he cites over 2.8 million customers using AI agents and over 237 million transactions categorized automatically in January 2026 alone.

Wells Fargo downgraded its outlook on Intuit stock to Equal Weight. Goldman Sachs $GS initiated coverage with a Neutral rating and a target price of $519. The valuation compression pushed the forward P/E below 30 times, pushing the stock to its lowest relative valuation in several years. Intuit continues to face competition from H&R Block, which has posted better-than-expected results following rapid expansion of its assisted tax segment and is an increasingly aggressive player in the estimated $37 billion assisted tax preparation market.

Thanks to a drop of as much as 56% from ATH, $INTU stock has come very close to its fair price as indicated by the Fair Price Index on Bulios, which calculates it from DCF and relative valuation. According to this calculation, Intuit stock is slightly undervalued at the moment.

CoStar Group $CSGP

Real estate database champion in uncertainty over return on investment

CoStar Group is a company that has no comparable competition in its primary market segment. CoStar operates the most extensive commercial real estate databases in the world and dominates the online real estate information ecosystem through its portfolio of digital marketplaces such as Apartments.com, Homes.com and LoopNet. Yet the stock has lost more than 33 percent in 2026 and is down more than 53 percent from its yearly high.

A combination of structural and sentiment factors are behind this development. On a structural level, the company has long shown strong revenue growth with very low profitability. Annual sales for 2025 were $3.25 billion, up 19 percent year-over-year, but net income was just $7 million.

The company has been investing massively in building the Homes.com platform as a competitor to Zillow $Z in the residential market, dooming it to minimal accounting profits for years. Investors who were still paying a premium for this strategy in 2025 are now rejecting it in an environment where capital is expensive and patience is significantly lower.

Activist pressure and a key strategic bet

CoStar Group's situation has been further complicated by activist investor Daniel Loeb of the Third Point fund, who has publicly expressed disappointment with the stock's long-term underperformance and capital allocation. Third Point criticized the massive advertising and marketing spend for Homes.com, the return on which remains questionable while shareholders bear the costs. This public confrontation increased pressure on the company's management at a time when the stock was already experiencing a sell-off.

The company reported fourth-quarter 2025 revenue of $900 million, up 27 percent year-over-year. For 2026, management reaffirmed revenue guidance of $3.78 billion to $3.82 billion and adjusted EBITDA in the range of $740 million to $800 million. It also completed a $500 million share buyback and approved a new $700 million repurchase authorization.

The consensus of 19 analysts covering the stock remains at Moderate Buy with an average target price of around $65. However, this goes against the FPI, which shows that $CSGP stock is still overvalued.

GoDaddy $GDDY

The pioneer site for small business owners battling AI

For years, GoDaddy has dominated the market for small business owners looking for domain names or web hosting. In 2024, GoDaddy stock was still experiencing a strong rally driven by optimism about customers switching to higher-margin app and e-commerce products. This made the entry into 2026 all the more painful. The stock has lost about 34 percent this year and is down more than 62 percent from its 2025 peak.

The immediate trigger for this year's decline was the earnings release on February 25, 2026, after which the stock plunged 14 percent in a single trading session, taking the title of the biggest one-day drop in the entire S&P 500 index that day.

Meanwhile, fourth-quarter results themselves beat estimates: EPS came in at $1.80 versus consensus of $1.58. The problem was the outlook. The company was projecting 2026 revenue in the range of $5.195 billion to $5.275 billion, an increase of about 6 percent, while consensus was projecting $5.28 billion. At the same time, bookings in the Applications & Commerce segment slowed from 14 percent to 11 percent growth due to a new go-to-market strategy with promotional pricing on one-year contracts.

AI as a double-edged weapon for web hosting platform

GoDaddy faces a specific form of AI pressure in 2026. While Intuit $INTU or Gartner $IT worry that AI will replace their products, GoDaddy faces the threat that AI will dramatically lower the barriers to entry for competitors. Wix, Squarespace and new AI-native web builders can now generate a professional website for any entrepreneur in minutes, while AI assistants also reduce the need to pay for the consulting and premium tools that have historically been GoDaddy's strength.

The company itself is actively integrating AI into its products, primarily as a tool for customers to create web content and online stores. But analysts question whether this approach is enough to retain customers in an environment where competitors are offering increasingly advanced AI features as a default part of the core product. Barclays, Morgan Stanley $MS, RBC and JPMorgan $JPM all cut their target prices significantly after the results, with RBC slashing it from $200 to $100 and JPMorgan from $200 to $167.

The analyst consensus still maintained a Buy rating with an average target price of around $137. The firm's strong free cash flow of $1.54 billion over the past twelve months remains one of the few unchallenged positive metrics.

However, the Fair Price Index at Bulios is pricing in an even bigger drop to bring the company to its intrinsic value.

Comparison to the market

The five titles described above share several key characteristics. First, they all operate in verticals where AI poses a legitimate and well-grasped business model disruption risk. Gartner sells knowledge that AI can replicate more cheaply. Intuit sells software for processes that AI can automate. CoStar sells database access that AI can make cheaper. GoDaddy sells web tools that AI can significantly democratize. AppLovin is building on advertising AI that customers can replace with a competitor with a larger data base.

Second, all companies entered 2026 with valuations that do not tolerate any slowdown or negative outlook. In an environment where the market has paid premium multiples for software firms, even a slight disappointment in guidance can trigger a revaluation in the tens of percent range.

Third, and perhaps most importantly, all businesses are reporting solid cash flow and revenues continue to grow. None of the companies is in an existential crisis. The stock price collapse is primarily a story of re-rating valuation multiples, not a story of business collapse.

A strategic view

The positive scenario for these companies rests on the argument that AI will be an asset rather than a destroyer for existing software leaders. Historically, platforms that have been able to integrate disruptive technology earlier than competitors emerge stronger from structural change.

The negative scenario, on the other hand, warns that the fundamental redistribution of value in the software stack is just beginning. New AI platforms may replace entire categories of vertical software in ways that are difficult to predict today. In such a scenario, today's valuations are still too high even after significant corrections. The most dangerous traps hide companies whose premium valuations still implicitly assume the continued dominance of a business model that AI can dismantle slowly but systematically.

What to watch next

What to watch next

At Gartner ($IT):

• Global Contract Value Trends in Q1 2026, especially the pace of contract renewals at AskGartner customers

• Ability to stem the decline in the Consulting segment and stabilize the outlook for the second half of the year

• Management's outlook for Analyst Day, if the firm holds one in the first half of the year

At AppLovin ($APP):

• Pace of expansion into the e-commerce advertising segment and specific financial metrics for this division

• Developing relationships with regulators and results of ongoing investigations

• Ability to maintain EBITDA margins above 80 percent even as we diversify into new verticals

For Intuit ($INTU):

• 2026 tax season results and TurboTax market share in the assisted tax segment

• Adoption of AI agents by QuickBooks customers and impact on customer retention metrics

• Potential revisions to management's outlook for the second half of fiscal year 2026

At CoStar Group ($CSGP):

• Monetization of the Homes.com platform in the form of specific revenue and customer adoption metrics

• Developing dialogue with activist investor Third Point and potential changes in capital allocation

• Pace of margin improvement as revenue grows

At GoDaddy ($GDDY):

• Results of new go-to-market strategy with promotional pricing on one-year contracts in Q1 2026

• Applications & Commerce segment growth rate and bookings recovery

• Ability to retain customer base in an environment of increasing AI-native competition

2026 has not started positively for software

The software sector in 2026 is going through one of the most painful valuation corrections of the last decade. The five companies described in this review illustrate the different faces of this phenomenon.

The one thing all companies have in common is not business weakness, but sentiment weakness combined with high entry valuations. Now more than ever, investors need to watch each of these companies to see whether AI will actually erode their competitive advantage or whether today's sell-offs create an opportunity to buy quality platforms at a discount.

Do you think laying off 20% of employees is a good move? Do you own $META shares?

At the same time JD is building an offline presence in Europe – the acquisition of Ceconomy (MediaMarkt, Saturn) for €2.2 billion opens up more than 1,000 brick-and-mortar stores across several European countries, which it can logistically connect with Joybuy and use as warehouses, pickup points and service centers. The whole move thus fits into a long-term strategy: China is saturated, e‑commerce competition is fierce, so JD is seeking further growth in Europe, where it wants to combine local warehouses and physical retail chains to offer the speed and service customers are used to from JD’s home market.

Disney says early tests of Verts on Disney+ and ESPN in August led to measurable increases in user engagement, which it attributes largely to a new recommendation algorithm that personalizes the clips each viewer sees. For now the feed focuses on promotional snippets from existing content, but the plan is to add creator-driven videos, new formats and more personalized experiences, which could turn Verts into a meaningful way to monetize time in the app through ads, partnerships and future bundled offers rather than just another discovery gimmick.

What Verts can bring to Disney+ in numbers

How Verts works: instead of static previews, it launches a stream of short clips straight away. This has several direct effects.

• Increases time spent in the app (more swiping, more clips before leaving).

• improves the conversion from "browse" to "watch" by instantly going from clip to full playback.

• Increases depth of catalog usage by pulling out older and less visible titles.

Disney has already reported from internal tests that Verts leads to additional engagement(user engagement rate) on both Disney+ and ESPN, suggesting a higher number of running titles or longer watches per user. In an environment where the streaming business is fighting for every minute of attention, and where ARPU is stagnant, such an increase in in-app time is essential for any additional monetization.

Monetization: from retention to advertising to licensing

In the short term, the most important thing is that Verts can boost subscriber retention. A user who opens the app for just a few minutes a day and "proscrolls" a few clips is more likely to remain a subscriber than one who only turns on the service occasionally for a single series. Higher retention directly reduces churn and increases customer value without having to increase the price.

The second level is advertising monetization. Disney+ already has an advertising tariff in the US and short vertical videos are an ideal format for:

• Short, unskippable spots between clips.

• Sponsored vertical "moments" (linking brands to specific IP, e.g. Marvel, Star Wars).

• Dynamic product placement and interactive elements in the future.

Once Verts collects enough data on clip viewing, Disney $DIS can add targeting by fandom: a different mix of content and ads for the Marvel fan, another for families, another for the sports audience via a link to ESPN. This brings Disney closer to the social networking model, where the feed is the primary ad inventory.

The third level is licensing and cross-promo. Verts can serve as a storefront not only for titles on Disney+, but potentially for:

• cinema releases (trailers, "first look" clips).

• Parks and experience products (short spots from Disneyland and Disney World).

• merchandising (viral moments promoting sales of toys and other products).

This allows Disney to shift some of the marketing budget to its own ecosystem, rather than paying external platforms like TikTok or Instagram to distribute clips.

Vertical video as a ticket to a younger audience

Disney+ isn't the only streamer testing a vertical feed, Netflix $NFLX is following a similar approach , but Verts is significantly more inspired by the TikTok model. The goal is clear: to reach younger users who typically spend time in the short clip feed and have a higher barrier to entry for classic long formats.

Disney has a major advantage in the strength of its brands. A short vertical clip with brand certainty (Star Wars, Marvel, Pixar, Disney Animation) has a much higher chance of grabbing attention than anonymous content. If Verts can:

• Turn a short clip into a follow-up to an entire movie or series.

• While keeping the experience safe and predictable for families.

It can become an addictive "elevator" to the catalog of movies and series for a generation that otherwise lives in foreign apps. This is an indirect but very important way to increase lifetime value among younger generations who may become long-term subscribers.

How big a contribution can Disney make to growth

There is no direct number yet on how much Verts will add to Disney's revenue. But the logic of the effect looks like this:

• Higher engagement per user increases the likelihood of long-term subscriptions.

• Higher retention reduces marketing costs to acquire new customers.

• more in-app time increases advertising inventory on Disney+ with ad rates.

In an ideal scenario, Verts can add a few percentage points to the average time spent on Disney+, which in turn translates into higher ARPU per user with an ad tariff and lower churn across the base. For a company the size of Disney, this is not a revolution in one quarter, but a structural shift that will manifest itself over a period of years in the form of a more stable and better monetized streaming business.

In addition, Verts is creating the technology and data foundation for other forms of monetization:

• Personalized fan experiences.

• Connections to ESPN sports content and future packages.

• testing short original formats that can live only in the feed without the high cost of full-length series.

Risks and limits

The main risk is user fatigue with vertical feeds. Unless Verts is significantly different from TikTok or Reels, users may prefer to open "original" and Disney+ will remain primarily a long-form content app. In that case, the benefit to growth will be limited.

Another risk is the balance between short and long content. If a feed of short clips starts to "cannibalize" the time users spend watching full movies and series, this could negatively impact the perceived value of subscriptions. Thus, Disney needs to carefully manage the algorithm so that Verts acts as a ramp-up, not a substitute for long-form content.

From a monetization perspective, how quickly and how aggressively Disney integrates advertising into the feed will be key. Too early or intrusive formats can harm the user experience, while an overly cautious approach will reduce the financial benefit.

What to watch next

The following signals are important to investors:

• Whether Disney starts sharing specific engagement metrics (time in feed, transitions to full titles) with Verts.

• How quickly and in what form Verts will expand outside the U.S. and to other devices.

• when and how ad formats and possibly sponsored content will appear in the feed.

• what place Verts will have in Disney's communication to the market (whether it will remain a "feature" or become one of the main pillars of the streaming strategy).

The most important variable is whether Verts will become a true attention hub within Disney+, picking up time and recommending content in a structured way, or just another add-on in the already crowded world of vertical videos. If the former scenario succeeds, it could contribute to more stable ARPU growth and reduced churn over the next few years, which is key to valuing the streaming part of Disney today.

After decades of stagnation, the Japanese stock market is returning to the spotlight for investors. The Nikkei 225 index has strengthened significantly in recent years, and more global funds are beginning to increase their exposure to Japan.

The reason is not just cyclical economic growth. Behind it lies a combination of structural changes, the return of inflation, and also a new political dynamic following the rise of Prime Minister Sanae Takaichi.

Political catalyst: the Takaichi government

The election of Takaichi means relative continuity in economic reforms for the markets, but also a greater emphasis on industrial and technology policy.

Her economic agenda rests on several key points:

•support for the domestic capital market

•investment in technologies and semiconductors

•strengthening the defense sector

•pressure for more efficient operation of Japanese companies

It is important for investors that the government continues corporate governance reforms that force companies to work better with capital and to think more about shareholders.

Why Japanese stocks are interesting again

Several factors are creating a relatively strong environment for the Japanese market.

1. Return of inflation After long years of deflation, inflation in Japan has stabilized around a few percent. This supports wage growth, consumption, and corporate revenues.

2. Improvement in corporate governance The Tokyo Stock Exchange is pushing companies to increase return on capital. The result is more dividends and buybacks.

3. Weak yen A weaker currency significantly helps exporters. Companies like Toyota thus generate strong foreign revenues.

4. Inflow of foreign capital Large global investors have begun to view Japan again as an alternative to the U.S. market, which is significantly more expensive today.

Key sectors of the Japanese market

Technology and semiconductors

Japan plays an important role in the global supply chain of the chip industry. Companies produce machines, materials, and testing technologies for semiconductor manufacturing. The growth of AI and data centers is creating strong demand for these products. ($TOELY, $ADTTF)

Automotive industry

Automakers remain one of the pillars of the Japanese economy. Hybrid technology and global exports still give Japanese companies a competitive advantage, although pressure from Chinese electric vehicle manufacturers is gradually rising. ($TM, $HMC)

Defense sector

Rising geopolitical tensions in Asia are leading Japan to increase military spending. This can support domestic industrial giants and defense technology manufacturers in the long term. ($MHVYF, $KWHIY)

Risks to the investment thesis

The Japanese story of course also has weaker aspects. The biggest include:

•energy dependence – the country imports most raw materials

•aging population

•sensitivity to the global trade cycle. The export-driven economy is strongly dependent on worldwide demand.

Investment conclusion

•Japan today offers a fairly unique combination of factors:

•political stability

•structural corporate reforms

•rising corporate profits

•relatively lower valuations than in the USA

Do you still think the Japanese market is undervalued, or has the Nikkei already priced in most of the positive scenario?

The company $BRK-B still holds a record amount of cash. By the end of 2025 it was roughly $373 billion. The money is mostly held in U.S. Treasury bills, but personally I think the company could already start to reduce some of that cash.

Some stocks are at attractive prices, and in my view Berkshire could start buying now.

We’ll see whether the strategy will at least slightly change now that the company has new leadership and Warren Buffett has left.