Judgment Day for the AI Trade
Beats across the board, but the market has moved from rewarding AI demand to interrogating AI returns.
Yesterday was AI judgment day.
Google, Microsoft, Amazon, and Meta all reported earnings at the same time, giving markets the closest thing we get to a live diagnostic on the AI economy.
The companies that own the cloud, the chips, the data centers, the ad rails, the enterprise relationships, and the consumer distribution all stepped up to the mic at once.
The result? Objectively strong quarters across the board - double-digit growth, accelerating cloud, real AI revenue.
The reaction? A lot more mixed. Google up. Everyone else down.
Ultimately it all seems to come down to narrative - who can tell an AI ROI story that the market will believe. Let’s get into it:
Google: Proving ‘em wrong
Google had the cleanest print. Overall revenue came in at $110B, +22% YoY. Google Cloud revenue grew 63% to $20B driven by AI workloads, TPUs, and enterprise Gemini adoption. GCP is moving from ‘lagging third place’ in the cloud to AI native challenger.
The Core businesses continue to steadily compound. AI was feared to be an existential risk to Search but not only did search not break, usage is growing and monetization is intact.
The most important part: Google is scaling AI without visibly wrecking margins. Operating income came in at roughly $40B, up 30% year over year, with margins around 36%.
Capex guide went higher to $180-190B in 2026 with even higher expected in 2027 as they continue doubling down on the infra buildout but Google got a pass because it showed the market a coherent equation: the core business is holding, cloud is accelerating, AI is contributing, and margins still look excellent.
Big Read: Google is the only company in the group that answered the market’s three questions at once: Is the core business safe? Is AI driving growth? Can you spend aggressively without destroying the P&L? For now, the answer appears to be yes.
Microsoft: Great is just not good enough
Microsoft also reported a very strong quarter, which the market received with the enthusiasm of a senior partner reviewing a very good associate memo and asking for more detail on page 37.
Revenue came in at ~$83B, +18% YoY. Microsoft Cloud hit roughly $55B, up 29%, while Azure grew around 40%, driven by AI workloads. AI revenue is now running at approximately a $37B annualized rate.
Copilot is seeing real adoption with 20M+ paid users across products, but monetization is still early and under scrutiny as customers evaluate ROI on the ~$30/user/month price point.
The big overhang is cost. Capex continues to ramp aggressively (~$30B+ this quarter, ~65% YoY growth) as Microsoft leans into the AI infrastructure buildout. Unlike Google, the market is more sensitive here given less clarity on long-term margin profile.
Big Read: Microsoft is still the leader in AI monetization today - with real revenue across Azure and Copilot - but the narrative is shifting to “prove the returns.” This is an AI-as-a-feature story layered onto the world’s best enterprise distribution engine, but investors are watching closely to see if growth can stay elevated while absorbing massive infrastructure spend.
Amazon: Demand is obvious, returns are not
Amazon’s quarter was also solid. Overall revenue came in at ~$181.5B, +17% YoY with a massive EPS beat ($2.78 vs ~$1.6 expected).
AWS grew 28% to ~$37.6B - its fastest growth in years - driven by AI workloads and enterprise demand. This is a clear re-acceleration and puts AWS firmly back in the conversation as a top-tier AI infra platform
Advertising continues to compound into a second high-margin engine, growing to ~$17B with strong momentum from retail media and AI-powered seller tools.
The core business is firing across the board - e-commerce profitability improving, cloud accelerating, and ads compounding. This is one of the cleanest multi-engine growth stories in Big Tech right now
The issue, again, is spend. Capex hit ~$43B in the quarter with ~$200B expected for 2026, driven by AI infrastructure and data center expansion
Despite the beat, stock reaction was negative as investors focused on lower free cash flow and continued heavy investment ahead of returns
Big Read: This is the most underappreciated setup in Big Tech: multiple high-margin engines (AWS + Ads) funding one of the largest AI infra buildouts in the world. AWS is reaccelerating and custom chips (Trainium/Graviton) give them a credible full-stack story. But the market is still skeptical - not about demand, but about how much they need to spend to win.
Meta: Capex increases decrease confidence
Overall revenue came in at ~$56B, +33% YoY - one of the fastest growth rates across Big Tech and meaningfully above expectations.
Advertising continues to perform extremely well, driven by AI-powered targeting and engagement. Ad impressions and pricing both increased, showing that AI is directly improving monetization.
The issue is forward spend. Meta raised capex guidance to ~$125–145B for 2026, a massive step up driven by AI infra (data centers, compute, chips). This meets lingering skepticism from prior investment cycles (remember Reality Labs?) and triggered a sharp stock reaction, making Meta the worst performer among the Mag 7 this earnings cycle.
Big Read: Meta is proving that AI already works for its core business - improving ads, engagement, and revenue growth. But unlike Google, the market is not giving them a pass on spend. The debate has shifted from “can Meta monetize AI?” to “are they over-investing ahead of returns?” This is the cleanest example of the new market regime: even the best AI ROI story gets punished if capex runs ahead of trust
All 4 companies showed strong growth, real AI demand, clear PMF and yet the market diverged sharply. Why? because we’ve entered a new phase.
2023-2024: Who has AI?
2025: Who can monetize AI?
2026: Who can monetize AI efficiently?
I think this capex cycle is going to look obvious in hindsight but for now, the market will continue to nitpick Azure growth, punish Meta capex, and question Amazon FCF. See you all in three months for another episode of: “Incredible quarter, stock down.”