OpenAI seeks $30B at a $1.4T valuation while its IPO slips to 2027 — safety is the stated reason

OpenAI is negotiating a pre-IPO round of at least $30 billion at a valuation near $1.4 trillion, Bloomberg reported on Tuesday and TechCrunch relayed. The company did not respond to TechCrunch's request for comment, so treat the figures as reported, not confirmed. The round would be a bridge to a public debut now expected next year — 2027 — rather than this one.

The arithmetic on the reported numbers is what makes this notable. In March, OpenAI raised $122 billion at an $852 billion valuation, and that was supposed to be its last private raise. Six months later the implied price tag is roughly 64% higher. Run-rate revenue — annualized from the most recent period, not money already banked — reached $40 billion in August, up 70% since July, a jump the report attributes to a strategic refocus on coding after Anthropic momentarily outpaced OpenAI at the start of the year. At $1.4 trillion against $40 billion of run-rate, that is about 35 times revenue: a growth-software multiple applied at a scale no private company has ever carried.

The IPO delay comes with an unusually blunt justification. Altman told Fortune: "I think it is unacceptable to be taking like a 10% chance of killing everybody by the end of the decade," and said safety work takes priority over the listing. We take the statement at face value and attribute it as his. But the structural consequence deserves saying plainly: staying private means fewer disclosure obligations, not more. A company that cites civilizational risk as its reason to wait is also postponing the audited filings, risk factors and quarterly scrutiny that would let outsiders evaluate that risk. If safety is the reason for the delay, the safety evidence should arrive before the prospectus does.

The competitive read matters as much as the financial one. The reported recovery came from coding — the one domain where AI's economic value is demonstrated rather than promised, which is exactly why every frontier lab has converged there. And the same TechCrunch coverage that day lists an office-suite-style launch and an agentic avatar called Dots, which fits the pattern we have tracked all year: the fight has moved from who has the smartest model to who controls distribution, integration and the plumbing agents run on. Capital at this scale is a weapon in that fight, not just a scoreboard.

Our thesis: this round is less a vote on today's product than a bet that the 2027 listing arrives with revenue that justifies the number. That bet is now systemically relevant. When a handful of firms carry balance sheets larger than most national budgets, they set the pace, the standards and — de facto — the rules, which is the concentration-of-power problem we keep flagging as the genuine short-term cost of this boom. A miss would not be contained to one cap table; it would reprice the entire AI capex cycle.

Our read: we are not here to sneer at big numbers. Compute this expensive is what funds protein modeling, diagnostics and materials discovery — the work that eventually cures things and hands people back their time, which is the whole reason we are optimistic about the long arc. But optimism is not credulity. Three things to watch: whether run-rate growth holds once the coding surge normalizes, whether the 2027 filing contains real safety disclosure rather than boilerplate risk language, and whether a 35x multiple survives contact with public-market investors who get to vote with their money. Until then, $1.4 trillion is a price, not a verdict.