Alphabet remains the world's largest digital advertising company, but the center of gravity is shifting. Advertising is still ~75% of revenue, yet Cloud is now 21% of revenue and — critically — 21.6% of group operating income, up from 9.0% a year ago. Cloud alone delivered 63% of the group's operating income growth this quarter.
Revenue breakdown (Q2 2026, reported):
| Segment | Q2 Revenue | % of Total | YoY Growth |
|---|---|---|---|
| Google Search & Other | $63.3B | 53% | +17% |
| YouTube Ads | $11.1B | 9% | +13% |
| Google Network (residual) | ~$7.4B | 6% | ~flat |
| Subscriptions, Platforms & Devices | $12.9B | 11% | +15% |
| Google Cloud | $24.8B | 21% | +82% |
| Other Bets | $0.4B | <1% | +2% |
| Total | $119.8B | 100% | +24% |
Growth drivers this quarter: Cloud backlog hit $514B — the demand pipeline is extraordinary. Gemini Enterprise is now used by ~90% of the Fortune 100. The Gemini app has 950M monthly users and API volume hit 22B tokens/minute (up from 16B last quarter). The FIFA World Cup drove all-time-record Search usage and 1.7B unique YouTube viewers. AI ad tools (Performance Max / AI Max) are delivering ~50% more conversions at similar ROAS, keeping the ad flywheel intact.
Advertising remains beautifully diversified — no customer above 1% of revenue. The structural dependencies are unchanged: ~$20B/year to Apple (AAPL) for Safari default search (at risk in the DOJ remedies appeal), smaller deals with Mozilla and Android OEMs.
Cloud is where concentration is quietly rising. Management flagged "very, very large customers" driving the capacity crunch, and Alphabet is renting third-party GPU capacity — including the reported $920M/month deal with SpaceX (SPCX) for Nvidia GPU compute — as a bridge while its own data centers get built. CoreWeave and Nebius shares jumped ~4-5% after hours when CFO Anat Ashkenazi said third-party capacity use will expand in Q3. Expect cloud revenue concentration disclosures to matter more in future 10-Qs.
The supply chain is now the binding constraint on the entire business:
Reported net income of $112.1B (+298% YoY) and EPS of $9.11 are optically absurd — and should be ignored. The driver was a $99.0B gain on equity securities (vs $1.3B a year ago), almost certainly the mark-to-market of the SpaceX stake post-June IPO plus Anthropic revaluation. Strip ~$6.26/share of that out and you get EPS ≈ $2.85 — exactly in line with the $2.86 consensus. The core business performed; it did not explode.
| Year | Revenue | Rev Growth | Net Income | Op Margin | Note |
|---|---|---|---|---|---|
| FY2021 | $257.6B | +41% | $76.0B | 30.6% | |
| FY2022 | $282.8B | +10% | $60.0B | 26.5% | |
| FY2023 | $307.4B | +9% | $73.8B | 27.4% | |
| FY2024 | $350.0B | +14% | $100.1B | 32.1% | |
| FY2025 | $402.8B | +15% | $132.2B | 32.0% | incl. equity gains |
| TTM Jun'26 | $445.9B | +20% | $244.2B | 33.1% | heavily distorted |
| FY2026E | ~$490B | +22% | n/m | ~33% | GAAP EPS not meaningful |
Segment profitability (Q2 2026): Google Services operating income $39.5B (41.8% margin) · Google Cloud $8.8B (35.6% margin, from 20.7%) · Other Bets -$1.8B (loss widened from -$1.2B — Waymo international expansion costs) · corporate ~-$5.7B. Total operating income $40.8B, a 34.0% margin. For every incremental cloud revenue dollar, ~54 cents fell through to operating income — best-in-class operating leverage.
| Metric | Value | Assessment |
|---|---|---|
| Cash & Equivalents (FY25) | $30.7B | Now swelled by Q2 raises |
| Q2 equity raise proceeds | $49.6B | First tranche of $80B program — dilutive |
| Q2 senior notes issued | $20.3B | €9B + C$8.5B + ¥576.5B multi-currency |
| Feb 2026 global bond raise | $31.5B | Incl. rare 100-year sterling bond |
| Debt-to-Equity (FY25) | 0.16 | Rising by design |
| Interest Coverage | >50x | No solvency risk whatsoever |
Assessment: Solvency is not the issue — Alphabet could carry 5x this debt. The issue is the model inversion. For two decades this was a self-funding annuity that returned cash. In 2026 it became a project-finance vehicle: negative FCF, $80B of new equity (the first major issuance in company history), and a century bond. Dilution of roughly 2-3% from the equity program lands precisely when buybacks have stopped — a double hit to per-share math.
| Period | Operating CF | Capex | Free Cash Flow | Capex % of OCF |
|---|---|---|---|---|
| FY2021 | $91.7B | $24.6B | $67.0B | 27% |
| FY2022 | $91.5B | $31.5B | $60.0B | 34% |
| FY2023 | $101.7B | $32.3B | $69.5B | 32% |
| FY2024 | $115.8B | $42.9B | $72.9B | 37% |
| FY2025 | $125.3B | $52.0B | $73.3B | 41% |
| Q2 2026 | ~$39.0B | $44.9B | -$5.9B | 115% |
Assessment: Q2 capex doubled YoY and exceeded operating cash flow. Buybacks: $0 (vs $13.2B in Q2 2025). Pivotal Research's full-year 2026 FCF projection was already just $8.2B back in February — with the capex guide now raised to $195-205B, FY2026 FCF will likely be negative. Management's framing ("highly ROI positive over the lifetime of the deal") is credible given 82% cloud growth, but the market no longer gets paid to wait: no buyback, a $0.20/quarter dividend, and dilution.
| Segment | Google Position | Key Competitors | Trend |
|---|---|---|---|
| Search Advertising | Dominant (~90%) | Microsoft/Bing (MSFT), ChatGPT, Perplexity | +17% and record usage — AI hasn't bitten yet |
| Cloud Computing | #3, fastest growing | AWS (AMZN), Azure (MSFT) | 82% growth, taking share |
| Frontier AI Models | Top 3 | OpenAI (GPT-5.6), Anthropic, Meta (META) | Gemini 3.5 Pro delayed; Gemini 4 next |
| AI Coding Tools | Challenger | OpenAI Codex (10M WAU), Anthropic Claude Code, SpaceX/Cursor (SPCX) | Antigravity 2.4M WAU, behind |
| Autonomous Driving | Clear leader | Tesla (TSLA), Zoox (AMZN), Baidu Apollo Go (BIDU) | Only scaled commercial player in the West |
| Video/Streaming | #1 hours | TikTok, Meta Reels (META), Netflix (NFLX) | World Cup: 1.7B unique viewers |
| Ticker | Company | Relationship |
|---|---|---|
| MSFT | Microsoft | #1 rival in cloud and AI; reports July 29 — the margin read-through matters for the whole trade |
| AMZN | Amazon | AWS is the cloud share leader; reports July 30. Also competes in ads and robotaxis (Zoox) |
| SPCX | SpaceX | Alphabet's stake drove the $99B equity gain; now also a $920M/month compute supplier. Stock down ~33% since Q2-end = Q3 EPS headwind |
| NVDA | Nvidia | Supplier of the GPUs Alphabet can't TPUs its way around fast enough; benefits from the $200B capex wave |
| AAPL | Apple | Receives ~$20B/year for Safari default search — the single biggest antitrust exposure in the DOJ remedies |
| META | Meta | Duopoly partner in digital ads; same AI-capex investor scrutiny applies to both |
| BIDU | Baidu | Apollo Go is Waymo's only scaled robotaxi rival globally (27 cities, 3.2M driverless rides in Q1) |
This quarter splits the story in two, and both halves are true at once. The operating story is the best in megacap tech right now: an 82%-growing cloud business at 35%+ margins with half a trillion dollars of backlog, a search franchise still compounding at 17%, and AI products with real user scale. The financial story is the inversion of everything that made GOOGL a compounder: negative FCF, no buybacks, dilutive equity raises, and century bonds.
The market's 4-5% after-hours wobble is the correct instinct, mildly executed. This is no longer a stock you own for capital returns and multiple expansion — it's a leveraged bet that cloud ROI arrives before investor patience runs out. The June report called GOOGL "the cheapest quality on the board" at $344; that framing dies with this quarter's GAAP distortions. On honest operating earnings it's ~30x — you now have to believe, not just calculate.
Watch two things over the next 90 days: Microsoft's and Amazon's cloud margins next week (validation or isolation), and the Q3 print, where a SpaceX-driven GAAP loss will test whether investors have actually learned to read through the equity-marks noise. If the stock sells off hard on that optical loss while cloud backlog keeps building, that's the entry the patient money gets.