OpenAI's trillion-dollar listing drifts toward 2027

OpenAI filed confidential IPO paperwork with US regulators in June and is reported to be targeting a valuation above $1 trillion, but the timeline has slipped: chief financial officer Sarah Friar has described the company becoming public in 2027, and Reuters reported in late June that a listing this year was no longer the plan.

The revenue trajectory explains why waiting is affordable. Reporting puts OpenAI's annualised run rate above $40 billion as of August, from roughly $20 billion at the end of 2025 and $25 billion early this year. A company compounding at that pace re-rates itself faster by staying private than by going public: every quarter of growth makes the trillion-dollar number look less like an aspiration and more like an arithmetic consequence. At $40 billion of run-rate revenue, $1 trillion is a multiple near twenty-five — aggressive by any historical software standard, yet conservative next to several private AI rounds priced this year at more than a hundred times revenue.

What the delay actually reveals is that the private market has stopped imposing a deadline. The traditional reasons to list — capital access and employee liquidity — are both being met privately, through mega-rounds on one side and structured tenders on the other. A company that can raise tens of billions and let staff sell without a prospectus goes public only when it wants the currency or the discipline.

The unresolved question is what the eventual filing will force into daylight. A prospectus would put numbers on the things the AI economy currently argues about from leaks: gross margins net of compute, the durability of consumer subscriptions, concentration among enterprise customers, and the true cost of the capacity commitments OpenAI has signed. The valuation debate is, in the end, a debate about those figures. The longer the listing waits, the larger they grow — in both senses.