AI Capex Approaching $1 Trillion: What Big Tech Q2 Numbers Really Mean

Aug 3 / Geoff Robinson






Alphabet, Microsoft, Amazon and Meta reported second-quarter 2026 results in late July. The headline numbers were strong across all four. The underlying story, and the one the market is now trading, is capital expenditure. Combined 2026 capex guidance across the four hyperscalers has reached approximately $725 billion, up roughly 77 percent from $410 billion in 2025. Consensus estimates for 2027 point toward $950 billion combined. The scale of AI infrastructure spending is now approaching the entire capital investment cycle at the peak of the late-1990s dot-com build-out as a share of GDP.

Meet Ava - Your AI Investment Coach

What Alphabet Actually Signalled

Alphabet raised its full-year 2026 capex guidance to $195-205 billion on 22 July, up from $180-190 billion the previous quarter. Google Cloud revenue grew 82 percent year-over-year to $24.8 billion, and the cloud backlog jumped $50 billion in one quarter to $514 billion. Finance chief Anat Ashkenazi described the environment as "supply-constrained", meaning Google is not building capacity fast enough to meet demand. Despite the operational strength, the stock fell in extended trading. The market is not rewarding the spending increase. It is questioning whether returns will scale with costs.

Where the Market Is Actually Positioning

Individual Mag 7 performance has diverged sharply in 2026. Microsoft is down approximately 21 percent year-to-date despite planning capex above $190 billion. Meta is down approximately 9.8 percent as investors question the return profile of its $125-145 billion capex programme. Amazon is roughly flat. Markets are rewarding companies where AI-related revenue growth is clearly visible (Alphabet, Nvidia) while penalising companies where the capex-to-revenue conversion is less obvious.

The beneficiaries of hyperscaler spending are outperforming the hyperscalers. Chipmakers, power infrastructure, cooling equipment, and data centre real estate names have caught investor attention as the more direct plays on the capex cycle.

What Analysts Should Actually Watch

Three metrics matter over the next four quarters. The ratio of AI-related revenue growth to capex growth at each hyperscaler. If revenue growth decelerates while capex accelerates, valuation compression continues. Cloud gross margins, which reveal whether AI infrastructure spending is generating attractive incremental economics or being competed away. Updated capex guidance at Q3 and Q4 earnings. Three of the four hyperscalers raised guidance at Q2. Another round would signal continued supply constraint; stabilisation would signal capacity catching up.

Conclusion

The AI capex cycle is now the largest single capital investment programme in the history of the technology sector. Whether it produces returns commensurate with the spend is the question that will define the next phase of Mag 7 performance. Passive investors in the S&P 500 are meaningfully exposed to the outcome whether they realise it or not.

For analysts who want to build the framework for reading capex cycles and their return profiles, the Equity Research and Valuation pathways on TheInvestmentAnalyst.com cover ROIC analysis and capital efficiency frameworks. Log in to start the free trial.

Choose Your Plan

Get 2 months for FREE with our yearly subscription.
No payment details needed - cancel anytime.

InshgtOne Logo

Monthly Subscription

30-day free trial
5,000+ digital assets


£20/month
Billed monthly

Yearly Subscription

30-day free trial
5,000+ digital assets
2 months FREE
£200/year
Billed yearly

For Teams

Continue
Professional-grade investment training for institutions.

£200/year
One-off annual payment
Purchase multiple subscriptions. Distribute, manage and reallocate subscriptions