The four major cloud-and-search names reported Q1 fiscal results last week. Microsoft, Amazon, and Alphabet all beat consensus on revenue and earnings, with Microsoft's AI business now at $37B annualized (up 123% YoY) and Azure at 40% growth. Alphabet jumped 10% on Thursday and is up 23% YTD. Meta — which reported $19.84B in capex for the quarter and raised full-year guidance to $125–145B — fell more than 8%.
The cleanest read of these prints is that the market is now pricing AI capex on a per-name basis. Capex itself is not the issue — Alphabet is also spending tens of billions and got rewarded for it. The dividing line is whether investors believe the spend converts to revenue this year, or whether it is a multi-year build.
Microsoft and Alphabet have the cleanest stories. Microsoft's $37B AI run rate is an ARR number against a measurable cost base — even if the gross margin profile is weaker than the legacy software business, the unit economics are visible. Alphabet's Cloud growth and Search-with-AI monetization are similarly traceable. Investors can model the conversion.
Amazon's beat ($181.5B revenue, $2.78 EPS vs. $1.63 consensus) was driven more by retail efficiency and AWS holding 40%-ish growth, but the AI piece of AWS is harder to disaggregate from cloud at large. The market shrugged because there was no thesis disturbance.
Meta is the outlier. The $125–145B 2026 capex range is debt-funded. The AI revenue line is buried inside ad-targeting improvements that are hard to attribute. The company is leaning into a longer-payback narrative than the rest of the group. That is a defensible strategic stance — it is also the one the market punished hardest. Mark Zuckerberg has been here before; the playbook is to ride out the noise and let the build pay off in 2027–2028.
A counter-read: the market is wrong about Meta and right about everyone else. The hyperscaler AI revenue numbers are partially recycled — Microsoft's $37B includes a meaningful chunk that is effectively cloud spend by AI customers Microsoft is also investing in, which is more circular than the press releases suggest. Meta's capex, by contrast, is going into infrastructure that produces user-facing AI products at billion-DAU scale. The "longer payback" framing might be honest accounting, and the 8% drop a buying opportunity.
- Total committed 2026 capex across the four: $700B+, with explicit guidance for higher in 2027 — see CNBC's coverage
- Bloomberg framed it as "AI Trade Winners and Losers"
- AEP separately reported a 63GW data-center contract pipeline, raising its capital plan to $78B
If you are allocating to Big Tech right now, the post-earnings prices reflect a market that wants near-term AI revenue visibility, not capacity bets. Names that can show ARR conversion (MSFT, GOOGL) are getting credit; names with longer-payback theses (META) are not. Pick the side of the trade that matches your time horizon.