CapEx Is Exploding as Alphabet Goes On a Spending Spree

Alphabet reported Q2 revenue up 24% YoY to $119.8B and Google Cloud revenue up 82% YoY to $24.8B (≈$8.8B operating income, ~35% operating margin), with cloud backlog rising to $514B from ~$460B three months earlier. However, CFO said capex is increasing by another $15B, lifting full-year capex to a range of $195B–$205B (from ~$190B), driving free cash flow to $-5.9B despite $39.1B operating cash flow—the first negative quarterly free cash flow since going public. Shares fell ~4% in premarket as investors weighed strong cloud momentum against near-term cash flow pressure through 2026–2027.

Analysis

The market is likely over-indexing on free-cash-flow optics and underpricing the strategic asymmetry: GOOG can fund a multi-year buildout from internally generated cash, while competitors have to defend cloud share under the same capex inflation. That matters because the next 12 months are less about reported earnings power and more about who can keep bidding for AI compute without forcing a sharper slowdown in buybacks or a leverage uptick. For AMZN and MSFT, the risk is not insolvency; it is multiple compression if investors conclude the hyperscaler model is entering a lower incremental ROIC phase.

The second-order effect is that AI spending is becoming a tax on index-level capital returns. If the largest names keep redirecting cash to servers and power, the support they normally provide to the Nasdaq via repurchases fades into 2026, making broad tech benchmarks more vulnerable than single-name headlines imply. The more interesting winners are not the clouds themselves but the ecosystem with contracted demand and short payback periods: compute suppliers, networking, and power infrastructure should still see order visibility improve even if mega-cap equities wobble.

Contrarian view: the consensus may be treating negative FCF as a structural impairment when it may simply be a temporary accounting outcome of front-loaded capacity creation. The key falsifier is not one quarter of negative FCF, but a 2-4 quarter slowdown in cloud backlog conversion or cloud margin compression despite rising spend. If backlog keeps compounding and inference costs fall fast enough, today’s capex surge can look like a moat-building phase rather than wasted spending.

Key Decisions for Investors

• Buy GOOG/GOOGL on further 3-5% post-earnings weakness; 3-6 month horizon. Thesis: cash-rich self-funding makes it the cleanest way to own the AI buildout. Falsify if cloud backlog growth slows or capex is raised above the current range without offsetting revenue acceleration.
• Pair trade: long GOOG/GOOGL vs short AMZN or MSFT over the next 1-3 months. GOOG has clearer backlog visibility and better cash-generation support, while the short leg absorbs the valuation hit from rising capex intensity and lower buyback capacity.
• Use a 6-12 month call spread on GOOG instead of outright equity if you want convexity with less drawdown risk. The stock can stay weak near-term while FCF is pressured, but the trade works if the market re-prices backlog and margin durability before 2026.
• Do not chase broad Nasdaq beta here; hedge QQQ exposure into the next cloud earnings season. The risk is index-level multiple compression if other hyperscalers echo the same capex escalation without proportional backlog growth.