Why Microsoft Will Likely Keep Its Iron Grip on OpenAI’s Future Profits; Newsom’s Strange AI Veto
As OpenAI nears a much-needed capital raise, some investors in the company have privately voiced hope that its largest shareholder and benefactor, Microsoft, will relinquish its rights to a large share of OpenAI’s future profits—an arrangement that effectively lowers the value of shares held by newer investors.
I wouldn’t bet on it.
Microsoft has the power to veto any proposed changes to those profit rights, and its executives haven’t wanted to change them as part of the funding round, according to someone briefed on the deal talks. Notably, Microsoft’s existing deal with OpenAI protects its rights to future profits even if the startup becomes a for-profit company, as it has been planning to do, according to this person.
To be sure, OpenAI is still losing billions of dollars a year, so those profit rights aren’t valuable right now. But its profit potential could improve next year if its $11.6 billion revenue projection turns out to be right.
The profit rights stem from Microsoft’s $10 billion investment in the company early last year. The way it works is that after OpenAI pays its earliest investors—Khosla Ventures, Reid Hoffman’s charitable foundation, the University of Michigan, Y Combinator partner Paul Buchheit and Y Combinator—with its first profits until their principal investment is paid back, Microsoft gets 75% of OpenAI’s profits until its principal investment is paid back ($13 billion in total) and 49% of profits after that until it hits a theoretical cap, which could be 120 times that amount.
The deal also grants Microsoft the right to use OpenAI’s technology in its products and to resell it to its own cloud customers up until OpenAI achieves artificial general intelligence, or AI that is on par with human intelligence, according to both companies. (Defining what exactly constitutes AGI is another question!)
The deal also makes Microsoft OpenAI’s exclusive cloud provider; the current contract runs through 2030 and can only be altered if both Microsoft and OpenAI agree to changes, though the two companies aren’t currently planning on making any such changes, according to the person briefed on the deal talks.
Under OpenAI’s existing structure, once profit distributions reach a cap, future profits would go to the nonprofit that currently governs the company. More recently, OpenAI has planned to convert to a for-profit corporation that’s no longer governed by the nonprofit.
Some investors expect the for-profit conversion could remove caps on profits shared with investors. But even if that happened, the special nature of its deal with OpenAI means Microsoft would retain its dibs on the profits, according to the person familiar with the companies’ agreement.
Some senior Microsoft executives have voiced support for the for-profit conversion, which could theoretically give Microsoft more influence in the form of shareholder voting rights. It’s not clear whether Microsoft would get traditional equity shares under such a plan. (As Semafor recently reported, Microsoft is concerned about antitrust scrutiny from officially owning a piece of OpenAI.)
It would be unusual for a for-profit company to guarantee a share of profits to specific shareholders. Companies typically distribute profits through dividends or share buybacks. But as CEO Sam Altman said last week in a memo announcing several tough personnel departures, OpenAI is “not a normal company.”
Profits or no, OpenAI is still a boon for Microsoft in more ways than one. As we laid out in detail here, the company extracts financial value from OpenAI through all kinds of revenue sharing that boosts its Azure cloud business and beyond.