Breakingviews: OpenAI’s value comes full circle

Sam Altman lost his job as OpenAI boss last month in the boardroom coup heard around the world. Four frenzied days later, with support from backer Microsoft and its Chief Executive Satya Nadella, he landed back in the top seat at the artificial intelligence startup. The incident unveiled profound tensions between OpenAI’s safety-driven, not-for-profit parent and the corporate entity it controls. Yet despite the whiplash, there’s a strong argument that the roughly $86 billion valuation OpenAI was seeking in a share sale prior to the ouster remains largely unchanged.

OpenAI explicitly tasked its board with putting safety over profit. So it’s not entirely surprising that just after directors flexed their muscles in firing Altman, some investors told Bloomberg they were considering writing down their stakes in the company to zero. The majority of OpenAI employees threatened to jump ship alongside their leader, highlighting the ability of one key executive to determine whether the company was perceived to be worth tens of billions of dollars, or nothing at all.

Altman’s eagerly awaited return brought some changes to the company renowned for its humanlike chatbot service, ChatGPT. Three of the four directors responsible for his exit are no longer on the board, according to a company blog post. New members, such as ex-Salesforce boss Bret Taylor, bring more business experience than the original cohort. That could be a boon to its valuation.

Assessing an AI company’s worth, especially for such a nascent business, is in some ways an exercise in guesswork. Yet investors have good reasons not to dismiss OpenAI’s mooted valuation. The company is set to make $1.3 billion in revenue this year, roughly 10 times more than its close competitor Anthropic, according to technology outfit The Information. Yet Anthropic as of October was in talks to raise capital at a $20 to $30 billion valuation.

To be sure, there are risks. OpenAI’s revenue may struggle to cover the costs it incurs to run and train its models. Plus, competition is steep. But these challenges existed well before November’s boardroom drama.

Indeed, OpenAI’s new board members seem less likely to make drastic moves. And Microsoft now has a seat allowing it to observe board meetings, even if it can’t vote. If anything, Nadella’s handling of the fiasco gives the software giant greater power over OpenAI without formal control. From a purely commercial perspective, the coup that didn’t kill OpenAI might have even left it stronger.