Sundar Pichai's Alphabet Has Grown Google Cloud Revenue 82% Year Over Year. Here's Why That Growth Rate Justifies the Company's Capex Bet.

Alphabet’s Google Cloud delivered Q2 revenue of $24.8B (+82% YoY) and more than tripled operating income to $8.8B, expanding segment margin from 20.6% to 35.5%, with $51.4B backlog and TPUs contributing to revenue for the first time. The company raised 2026 AI capex guidance to $195B–$205B (from $180B–$190B), with ~60% of quarterly capex going to servers, while management expects continued capex growth through 2027 despite temporary third-party capacity pressures. Even with free cash flow at -$5.9B in Q2, Alphabet still generated $39.1B operating cash flow and $185.7B over the trailing 12 months, supporting the AI infrastructure build-out.

Analysis

Google Cloud’s acceleration reframes Alphabet’s AI spend from a cost overhang into a funded moat: the key variable is not capex intensity, but whether incremental spend is still clearing hurdle rates above Alphabet’s WACC. If cloud margins hold near current levels, the market should be willing to look through near-term FCF noise and assign a higher terminal multiple to GOOGL/GOOG versus peers that are still buying growth with weaker operating leverage.

The second-order winner is not just Alphabet’s cloud franchise; it is also the ecosystem around proprietary AI infrastructure. That said, the mix matters: more in-house silicon and integrated stack deployment is supportive for total compute demand, but it caps the long-run pass-through to merchant GPU suppliers. Translation: NVDA still benefits from the broader capex wave, but the “every AI dollar = NVDA dollar” trade is increasingly crowded and mechanically less clean as hyperscalers substitute toward custom accelerators.

The main risk is timing mismatch between revenue and cash conversion. Backlog and margin expansion can make the story look self-validating for a quarter or two, while temporary third-party capacity keeps FCF suppressed and forces the market to test whether growth is truly self-funding. If cloud growth decelerates materially or margins roll over from the low-to-mid 30s, the stock could de-rate quickly; if not, this becomes a 6-18 month multiple-expansion story rather than a one-day headline trade. The contrarian view is that the market may be underestimating Alphabet’s compounding advantage, while overestimating the durability of pure-play GPU scarcity.

Key Decisions for Investors

• Long GOOGL on 5-10% pullbacks; 6-12 month horizon. Base case is multiple expansion as the market reclassifies AI capex as a high-ROIC growth engine rather than a cash drain. Falsify if cloud margin slips back below 30% or backlog conversion slows for two straight quarters.
• Pair trade: long GOOGL / short AMZN for 1-3 months. Thesis is relative cloud operating leverage and stronger AI monetization visibility at Google; risk is AWS re-accelerates or Amazon signals a sharper AI workload inflection.
• Trim or underweight NVDA on strength if your thesis is purely hyperscaler capex growth. Custom silicon adoption at Google and other hyperscalers shifts some of the spend away from merchant GPUs over 6-18 months; keep NVDA only as a broader AI-capex beta, not a clean standalone beneficiary.
• Set an alert on Alphabet’s next cloud update: if operating margin stays above ~30% while capex remains elevated, add on dips; if margin compresses and FCF stays negative beyond the next 1-2 quarters, fade the enthusiasm.