To keep $1trn valuation alive, OpenAI might delay IPO

OpenAI's bankers, at a meeting some time in the past week, laid out two doors. Behind the first was a public offering next year at the trillion-dollar valuation Sam Altman has been pushing for. Behind the second was a faster listing, possibly this autumn, at a price the market would actually bear. Mr Altman, OpenAI's chief executive, declined to walk through either. Any cut to the trillion-dollar figure, he told one person involved, was "a nonstarter" — which would leave, by elimination, the wait.

The news that OpenAI is leaning toward postponing its IPO until 2027 has been read, reasonably enough, as a flinch. Markets have been choppy; tech stocks have dragged the indices; SpaceX, whose June listing was the largest in history, has spent its first fortnight as a public company wiping out most of its post-debut surge. Retail enthusiasm, OpenAI's advisers now warn, may not be there at the price. All of that is true, and none of it is quite the story. The choice to wait, if it comes, is not the market's to make. It is Mr Altman's, and he would make it to protect a number.

Holding pattern

OpenAI's last private valuation, set in February's $110bn round, was $730bn. The trillion is not a mark anyone has paid; it is an aspiration, a round figure Mr Altman instructed his bankers to find a way to reach. At $1trn, OpenAI would trade near 40 times its current revenue run-rate of roughly $25bn, and near 77 times the $13bn it actually booked last year. Microsoft, a profitable near-monopoly, trades around twelve times sales; Walmart, nearer one. The company is not believed to have turned a profit, and it is pouring money into data centers at a rate that shows no sign of slowing. The trillion is a number OpenAI has chosen, not one its financials produce.