GMO's Supply-Side Bubble Thesis: SpaceX Lockups + Anthropic/OpenAI IPOs Could Add ~5% to Investable Market Cap — the Break May Come Before AI Demand Disappoints
GMO's current quarterly letter ("A Catalyst for the AI Bubble Break," Ben Inker & John Pease) introduces a bear vector that is mechanically different from every demand-side argument in the capex debate: the catalyst may be equity supply, not disappointing AI revenue.
The argument, in GMO's own words
- "Bubbles need fresh capital, and rising supply often breaks them… in the case of the AI bubble, we think supply may be the catalyst for its eventual demise.1"
- The supply events are dated and specific: "On June 12, 2027, the last batch of approximately $2 trillion in SpaceX shares will be released from its contractual 'lockup.'" And: "IPOs from Anthropic and OpenAI could add roughly 5% to the investable market cap too."
- The historical multiplier: "a 1% increase in the market share of IPOs has historically led to an approximate 4% decrease in prices."
- The implication: "we might anticipate returns approximately 20% lower due to rising equity supply, ignoring additional risks from high valuations and expectations."
- Structural fragility: ~83% of the $30T in active/passive funds focus solely on U.S. equities and only ~10% of retirement accounts change asset mix annually — "even if all funds moved their portfolios to reflect changing opportunity sets, the price dislocations created by flows might remain unmanageable."
Why this matters for the Nvidia capex verdict
Every other bear vector in this topic requires something to go wrong with AI demand or financing (cancellations, depreciation catches up, OpenAI breaks). GMO's vector requires only that the ecosystem's early holders cash out — an event that is already scheduled. The perversity: the very IPOs being watched as validation of the capex story (Anthropic's ~$2T IPO: Mid-November Target, Oct 14 Investor Meetings — and the S-1's 2025 Revenue Baseline Is Now Public via Reuters — Anthropic's up-to-$100B raise at ~$2T; OpenAI's confidential June 8 S-1) are simultaneously supply events that GMO's flow math says pressure the whole market, Nvidia included. SpaceX's lockup ladder adds a second, larger tranche through mid-2027.
This does not contradict the intact verdict on capex substance (Nvidia's AI Capex Cycle Sustainability: Verdict 2026-10-08 — Intact for a Ninth Straight Read; Records on the Supply Leg, and the Debt Funding It Gets Priced Harder) — demand, unit economics, and financing are all still confirming — but it supplies a mechanism by which the equity market could break before any fundamental datapoint does. It is the mirror image of the de-rate: NVDA's multiple can compress on supply/flows alone with no change in orders.
-
An instance of The bear case for AI capex no longer requires demand to disappoint. — GMO's dated equity-supply events — SpaceX lockups and lab IPOs adding roughly 5% to market cap — are a break mechanism requiring no failure of AI demand. ↩︎