"Google is raising $80 billion, and you're bearish?" — @cjgustafson222

Why this is in the vault

CJ turns Alphabet's record $80B equity raise into a clean operator-grade lesson on the three capital "drawers" (cash, debt, equity) and cost-of-capital balancing — a fresh datapoint for RDCO's hyperscaler-capex capital-cycle thesis, paired with the recurring Benchmarks-for-Operators metrics section that feeds Sanity Check's SaaS-metrics literacy.

⚠️ Sponsorship

• Primary sponsor: Abacum — top-of-issue block, "proudly powered by Abacum." CJ pitches Abacum as the first FP&A platform live as a Built-for-NetSuite SuiteApp (connected planning/forecasting against live ERP data; name-drops Strava, PostHog, JG Wentworth as customers). Abacum is a KNOWN recurring Mostly Metrics sponsor (FP&A/NetSuite play). The opener is written in CJ's voice (the "data plumber exporting CSVs from NetSuite" lament), so the sponsor framing bleeds into the editorial hook — treat the FP&A-tooling angle as paid, not independent.
• Secondary / affiliate: Koyfin — closes the benchmarks section as the "data partner" with a tracked affiliate link (?via=metrics). All the valuation/efficiency charts are sourced from Koyfin. Affiliate disclosure, not a display sponsor.
• No Mostly Talent (CJ's own recruiting arm) self-promo in this issue.

The core argument

Alphabet announced an ~$80B equity raise to fund 2026 data-center buildout — by CJ's framing the largest-ever public equity offering, ~3x the prior record, and the first ever raised from a position of financial strength (prior mega-raises were distress: Boeing safety crisis, BofA/AIG/Citi in 2008). The piece walks the "three drawers you open when you need cash":

1. Cash — Google's ad engine is "the best cash machine in the world" (~$174B TTM operating cash flow), but last quarter $35.7B of $45.8B OCF went to capex, leaving only ~$10B. Cash alone can't fund the buildout.
2. Debt — raised $85B+ across six currencies in a year, total debt past $100B from $25B a year prior, including a 100-year bond due 2126 at 6.125%. More debt would threaten the pristine AAA/AA+ rating and raise marginal cost of debt.
3. Equity — the rare third drawer. With the stock near an all-time high and a ~$4T market cap, cost of equity is "remarkably cheap," so they issue stock to re-balance WACC and avoid choking the credit rating.

Two deeper points: (a) asset-liability time-horizon matching — you fund a multi-decade, generation-defining industrial infrastructure shift with permanent capital, not rolling working capital or short-term debt; the equity base mirrors the multi-decade lifecycle of the physical data-center grid. (b) Execution without tanking the stock — anchored $10B with Berkshire Hathaway (steady long-term holder) and used mandatory convertibles for a fixed, predictable dilution schedule. Context: management guided 2026 capex to $180–190B, climbing in 2027 — about to outspend what the best business model in the world generates in cash.

Operator lessons: when your valuation is rich, your equity is your cheapest currency on the balance sheet; "scared money don't make money"; don't run an aggressive product roadmap with a timid capital strategy. "Open the expensive drawer while it's full."

The back half is the recurring Benchmarks for Operators data section: weekly valuation/efficiency metrics (Rule of 40 = TTM revenue growth % + TTM adj. EBITDA margin %), NTM revenue multiples, OPEX-as-operating-leverage (COGS / S&M / R&D / G&A buckets, +25% profitability target at scale), plus the refreshed nine-sector public-comp taxonomy with full company lists.

Mapping against Ray Data Co

• Investing / capital-cycle thesis — STRONG reinforcement. This is the demand side of RDCO's hyperscaler-capex thesis, on the record from the buyer's own balance sheet. The $180–190B 2026 capex guide "climbing in 2027," plus a record equity raise specifically earmarked for data-center expansion, is a textbook Phase-2 capital-cycle marker (capacity-announce / capex-funding) — exactly the signal investing-edgar-watch is built to scrape from 10-Qs. It corroborates the memory-cycle and power-cycle theses (more compute buildout → sustained HBM/DRAM and power demand). Pairs directly with the vault's Ben Thompson "Google Capital Company" note on the same raise — CJ gives the FP&A/cost-of-capital mechanics, Thompson gives the "cash is the ultimate commodity" strategic frame.
• SaaS-metrics literacy for Sanity Check — MEDIUM, reusable fuel. Rule of 40, NTM multiples, OPEX/operating-leverage buckets, and the nine-sector taxonomy are clean reference scaffolding for the Sanity Check voice. The taxonomy here matches CJ's earlier "Sector Refresh / New Nine" (Take-Rate Platforms split out from Vertical SaaS; Payments split from Consumer Fintech) — useful for any RDCO comp-set framing.
• Content-as-a-product — voice model, not a topic. The banger headline + self-deprecating asides ("Strategic Backoffice Overhead," "this is my quant") are a model for Sanity Check cadence. Per the no-derivative rule, this is evidence/voice reference, NOT a Sanity Check topic to restate — any SC angle would need an original re-frame (e.g., "when is your equity your cheapest currency" applied to bootstrapped operators).
• Gap / caution: "largest-ever public equity offering" and the 3x-prior-record claim are CJ's editorial framings without a cited league table — treat as directional, not a verified superlative, if reused in an RDCO output.

Related

• [[06-reference/2026-06-02-stratechery-google-capital-company]] — Ben Thompson on the SAME Alphabet $80B raise / Berkshire $10B anchor; the strategic-frame companion to CJ's FP&A mechanics
• [[01-projects/investing/theses/2026-05-17-memory-cycle-v1]] — hyperscaler capex is the load-bearing demand signal; this raise is a Phase-2 capex-funding marker
• [[06-reference/2026-05-17-mostly-metrics-sector-refresh-new-nine]] — the same nine-sector public-comp taxonomy CJ uses in this issue's benchmarks section
• [[06-reference/2026-05-14-mostly-metrics-q1-2026-startup-benchmarks]] — prior Mostly Metrics benchmarks drop (Rule of 40 / CAC / NDR reference set)