Alphabet

Type: Company (GOOGL) — Hyperscaler, demand side

Alphabet is one of the two or three largest single sources of datacenter demand and, for this KB, the cleanest illustration of how fast the capex cycle has re-based.

Q2 2026 HAS PRINTED — this page now carries actuals. Alphabet filed its Q2 2026 8-K Ex-99.1 on 2026-07-22 (SEC accession 0001652044-26-000066). The pre-print "expectation" language below has been superseded by the printed numbers; see Alphabet Q2 2026 Results (Actual Print). Headline actuals: total revenue $119.8B (+24% YoY); Google Cloud $24.8B (+82% YoY) at a 35.6% operating margin; Q2 capex $44.9B; free cash flow NEGATIVE $(5.9)B, funded by a $49.6B June equity raise + $20.3B of senior notes. [primary: 8-K Ex-99.1, 2026-07-22]

The load-bearing print: capex outran cash, and the gap was funded by capital markets. Q2 capex ($44.9B) roughly doubled year-over-year while operating cash flow ($39.1B) did not keep pace, driving quarterly free cash flow negative $(5.9)B. Alphabet plugged the gap with a $49.6B equity raise (June 2026, explicitly earmarked for "capital expenditures to scale AI infrastructure and global compute") plus $20.3B of senior unsecured notes; long-term debt more than doubled to $98.2B in six months. This is direct, on-print confirmation of the debt/equity-financing fragility the KB flagged as open on the Hyperscaler Capex Cycle — the buildout is now financed by the balance sheet, not the ad business's cash.

Two watched signals did NOT resolve on the print. The Ex-99.1 press release carries neither a FY2026 capex dollar guide (the pre-print's watched $180–190B is a call/10-Q item) nor a Google Cloud RPO/backlog figure (the watched ~$462B lives in the 10-Q). Both await the call transcript / Form 10-Q; the KB should not treat either as confirmed or broken yet. The prior $240B-vs-$462B backlog conflict therefore remains open. See Contracted Backlog & RPO.

Sourcing quality is now high. The dossier is anchored to Alphabet's own 8-K Ex-99.1 (first-party, fetched from EDGAR), replacing the paywalled, summary-derived preview that previously drove the "low" confidence rating.

Key Facts

Q2 2026 — Actuals (printed 2026-07-22)

• Total revenue $119,796M (~$119.8B), +24% YoY (+23% constant currency); consolidated operating margin 34%. Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• Google Cloud revenue $24,768M (~$24.8B), +82% YoY — a meaningful acceleration (was $13.6B a year ago, $20.0B in Q1 2026). Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• Google Cloud operating income $8,814M → operating margin 35.6% (up from 20.7% a year ago). The pre-print's watched "~32.9%" was held and expanded. Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• Q2 capex (PP&E purchases) $44,924M (~$44.9B) — roughly double the year-ago $22.4B; H1 $80.6B; TTM $132.4B. Landed right at the pre-print's watched ~$45B. Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• Free cash flow Q2 2026 = NEGATIVE $(5,855)M — capex ($44.9B) exceeded operating cash flow ($39.1B). Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• June 2026 equity raise $49.6B net (Class A/C + 6.25% mandatory convertible preferred), earmarked for AI-infrastructure capex; plus $20.3B net senior unsecured notes; a $40.0B ATM program established (none sold as of Jun 30). Long-term debt doubled to $98.2B in six months. This confirms real, on-print equity/debt funding of the buildout — superseding the unverified "$84.75B June raise" figure the preview carried. Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• Google Cloud now books TPU system sales as product revenue ("product revenues primarily from the sale of TPU systems") — an own-silicon value-capture disclosure. Evidence: strong (first-party 8-K) [primary: 8-K Ex-99.1, 2026-07-22] (GOOGL Q2 2026 print)
• NOT in the print — awaits call/10-Q: the FY2026 capex dollar guide (watched $180–190B) and the Google Cloud RPO/backlog (watched ~$462B; prior $240B-vs-$462B conflict still open). Neither appears in the Ex-99.1. Evidence: n/a (absent from source) [primary: 8-K Ex-99.1, 2026-07-22 — items absent]

Pre-print context (superseded by the actuals above)

• FY2026 capex guidance $180–190B heading into the print — the product of prior upward revisions from an initial $71–73B; a ~2.5x re-guide inside one planning cycle. Its printed status is unresolved (guide absent from the exhibit). Evidence: weak (summary-derived preview) (Alphabet preview); (Futurum)
• [SUPERSEDED — now actual] Q1 2026 Google Cloud revenue $20.0B, +63% YoY, operating margin 32.9% was the base the Q2 print accelerated from. Evidence: weak (web-search-derived preview) (Alphabet preview)

Mentioned In

• Hyperscaler Capex Cycle — the $71–73B → $180–190B re-guide is the cycle's clearest single-name signature
• Contracted Backlog & RPO — the backlog conflict lives here
• Data-Center Construction as a Macro Variable — Alphabet's capex is a measurable share of the national investment number
• Hyperscaler Buyers — group-level treatment

Related Entities

• Hyperscaler Buyers — Alphabet is one of the four
• Oracle — the other name in this KB with a large disclosed backlog/RPO
• TSMC — Alphabet's TPU custom silicon competes for the same advanced-node and packaging allocation

Outstanding Questions

• Does the FY2026 capex dollar guide (watched $180–190B) get raised, held, or cut? Not in the Ex-99.1 — resolve from the call transcript / Form 10-Q. The Q2 run-rate ($44.9B) annualizes to ~$180B.
• Which backlog figure is right — $240B or ~$462B, and did it grow? Still unresolved: no RPO/backlog line in the earnings release. Pull from the 10-Q, with duration, cancellability, and customer concentration.
• Equity/debt funding is now confirmed on-print ($49.6B raise + $20.3B notes + $40B ATM) — how long can FCF stay negative before it constrains the buildout? The financing fragility is no longer hypothetical; the open question is its trajectory.
• How much of Alphabet's capex is TPU custom silicon versus NVIDIA purchases, and how much of the compute spend is recaptured via the new TPU-system product revenue line?
• Next action: ingest Alphabet's Q2 2026 10-Q (RPO/backlog + capex guide) and the call transcript when available to close the two items the Ex-99.1 left open.

Changelog

• 2026-07-20 — Seeded from the paywalled preview + capex-cycle sources (low confidence, summary-derived).
• 2026-07-23 — Compiled the Q2 2026 8-K Ex-99.1 (ingested 2026-07-23): actuals replace the pre-print. Load-bearing spine finding — capex outran cash: Q2 capex $44.9B (calendar-Q2, cash PP&E) > OCF $39.1B → FCF $(5.9)B, funded by a $49.6B June equity raise + $20.3B notes (LT debt doubled to $98.2B). Google Cloud $24.8B (+82% YoY) at 35.6% op margin; Cloud now books TPU-system sales as product revenue. FY2026 capex dollar guide and RPO/backlog absent from the exhibit (call/10-Q items) — marked unresolved-by-the-print, not invented. Confidence low → high (first-party 8-K).