Amazon, Microsoft, and Google-parent Alphabet have spent weeks convincing investors that their AI buildout will pay off, but the market’s patience is thinning as 2026 capital-spending plans climb.
Amazon helped spark the latest wobble after flagging about $200 billion of 2026 capex, while Alphabet projected $175 billion to $185 billion, figures that revived worries about near-term cash flow and margins.
In the background, a different trade is working: buying the suppliers and landlords of AI infrastructure.
Heavy AI capex dampens mega-cap stocks
The pressure on hyperscalers is not about demand fading. It is about the price tag of meeting it.
Amazon’s forecast for roughly $200 billion of 2026 capital spending, aimed largely at data centers and AI-related infrastructure, surprised investors and weighed on the shares.
Alphabet’s outlook was similarly jarring: it told investors to expect 2026 capex of $175 billion to $185 billion, a step-up that briefly pushed the stock down as much as 3% in after-hours trading, despite a strong quarter.
Investors are making a classic trade-off calculation.
Heavy capex can boost long-term capacity, but it also hits free cash flow in the near term, money that could have gone to buybacks, dividends, or simply kept margins higher.
The investors are simply concerned about Big Tech’s AI spending, which may approach roughly $700 billion this year, and the “blow to cash” is starting to raise red flags.
That skepticism is also feeding a rotation dynamic.
When large platforms tell the market they need to spend aggressively just to keep up, their stocks can de-rate (meaning investors are willing to pay a lower valuation multiple) even if revenue holds up.
The flip side is that every dollar of hyperscaler capex becomes potential revenue for companies selling chips, manufacturing capacity, and data-center infrastructure.
Read More: Anthropic lands $30B at $380B valuation as AI funding hits new extreme
AI’s quiet winners
As per The Motley Fool analysis, one clear beneficiary has been Taiwan Semiconductor Manufacturing Company.
TSMC has been posting record results and pointing to robust AI-related demand for advanced chips, while planning to keep investing heavily to expand leading-edge capacity.
In January, TSMC’s quarterly profit jumped 35% as AI-driven orders for advanced chips continued to dominate, and the company outlined 2026 capex of roughly $52 billion to $56 billion.
Separate market coverage also tied a stock rally to strong January sales and AI-chip strength, reinforcing the view that TSMC’s fabs remain central to the AI supply chain.
Nvidia has become the purest “picks and shovels” expression of the AI buildout.
The stock jumped about 7.8% on Feb. 6 for its best day since April, as investors bought the companies most directly positioned to benefit from the same hyperscaler spending that weighed on the platforms themselves.
Nvidia CEO Jensen Huang argued that the roughly $660 billion capex buildout by major customers is sustainable, a message that helped steady sentiment after a volatile stretch for AI-linked equities.
Applied Digital is a smaller name, but it sits in a sweet spot of the same trend, providing specialized data-center capacity for high-performance computing workloads.
Applied Digital secured long-term lease agreements with CoreWeave, expected to generate about $7 billion in revenue over the term, effectively positioning the company as a capacity provider to an AI hyperscaler.
The takeaway for 2026 is straightforward.
As hyperscalers absorb the capex burden, the market is rewarding the firms with cleaner, more direct exposure to AI infrastructure demand.
The post These 3 stocks are quietly soaring as Microsoft, Google slump on AI spending appeared first on Invezz