Analysis
Alphabet, Amazon and Meta together plan to spend more than $500 billion on AI infrastructure in 2026, with Alphabet guiding to $195-205 billion, Amazon to roughly $200 billion, and Meta to $125-145 billion -- spending levels Fortune and Yahoo Finance have both tracked climbing with nearly every quarterly earnings report this year.
The market isn't rewarding that spending uniformly. Alphabet shares have risen on its capex guidance even as Microsoft -- down roughly 21% on AI spending concerns despite over $190 billion of its own capex -- and Meta, down nearly 10% after a similar earnings reaction, have drawn investor skepticism on comparable numbers. That split suggests markets are pricing execution and monetization differently by company, not simply penalizing AI spending as a category.
“That split suggests markets are pricing execution and monetization differently by company, not simply penalizing AI spending as a category.”
Nvidia remains the one name collecting cash regardless of which hyperscaler wins the underlying model race: the company posted $102.7 billion of operating cash flow in fiscal 2026 and authorized an $80 billion buyback, per reporting cited by 24/7 Wall St -- the clearest picks-and-shovels beneficiary of the buildout.
What the headline capex numbers miss: a growing share of this spending is financed with debt and structures like Amazon's reported $8 billion sale-and-leaseback of Nvidia chips, rather than purely from retained earnings -- the same leverage dynamic Pulse has tracked straining smaller neocloud borrowers' own debt deals. Announced capex is also not deployed capex; the gap between a capex guidance number and actual installed, revenue-generating capacity is where most of the execution risk investors are now pricing sits.