The governance structure at OpenAI was a novel one, as I reported earlier this year—meant to limit the control one corporate entity could have over the development of the important, but potentially dangerous technology of artificial general intelligence. Now that unusual structure has backfired.
Late on Friday, the board of OpenAI’s nonprofit suddenly ousted CEO Sam Altman and pushed former OpenAI president, Greg Brockman, off the board (he and key OpenAI researchers resigned in protest of the board’s actions shortly thereafter). The subsequent revolt within OpenAI’s ranks led to talks of bringing Altman back, though by late Sunday evening, those talks had dissolved, with the board naming Emmett Shear, who cofounded Twitch, as interim CEO.
OpenAI’s shareholders have been involved in weekend efforts to course correct and bring back Altman, using whatever influence they can: Tiger Global and Sequoia Capital had been negotiating behind the scenes, according to a report in The Information.
But OpenAI’s venture investors—which include Khosla Ventures, Tiger, a16z, and Sequoia—don’t hold much leverage at the company and hold no actual control at the organization, since they purchased shares in the company through tender offers, and none of them have board seats. Investors have a cut in a waterfall profit-share model, but that doesn’t translate into influence. And it evidently doesn’t translate to visibility either, as OpenAI’s investors were caught off guard by the sudden change. An OpenAI investor, who spoke with me on condition of anonymity, said that investors were not apprised of the developments. (Reid Hoffman and investors or representatives at Khosla, Tiger, a16z, and Sequoia declined to comment or didn’t respond to my request for comment before press time.)
OpenAI’s unusual structure gives its nonprofit full governance rights over a holding company that is owned by OpenAI employees and minority investors including Microsoft and venture capital firms, according to a corporate structure map from late last year that was reviewed by Fortune. OpenAI’s nonprofit corporate board, which made the decision to fire Altman, has six seats—and none of the people on the board may hold a financial interest in OpenAI.
All of this gives shareholders limited protection from a crisis quite like this. And there doesn’t appear to be anything equivalent to a “Key Man” clause in the investor agreements, according to investment materials from earlier this year that were reviewed by Fortune.
Microsoft apparently also has limited visibility into the situation, though it does have more leverage and weight to throw around—more so due to its operating relationship with OpenAI than through anything else. As my colleague Jeremy Kahn reported on Sunday, OpenAI is “entirely dependent on Microsoft’s cloud computing data centers to both train and run its models.”
As he wrote:
By turning to a single corporate entity, Microsoft, for the majority of the cash and computing power OpenAI needed to achieve its mission, it was essentially handing control to Microsoft, even if that control wasn’t codified in any formal governance mechanism. Ironically, if OpenAI’s nonprofit board chooses to ask Altman to return and resigns, as they are contemplating, it will prove that Altman’s structure failed—OpenAI was not able to both raise billions of dollars from a big tech corporation while somehow remaining free from that corporation’s control.
All of the events from this chaotic weekend suggest that OpenAI’s unusual governance structure—organized by Altman himself—backfired on him. You can read the rundown of what happened this weekend here. If you have more information or thoughts on the situation to share, reach out to me below.
See you tomorrow,
Jessica Mathews
Twitter: @jessicakmathews
Email: jessica.mathews@fortune.com
Submit a deal for the Term Sheet newsletter here.
Joe Abrams curated the deals section of today’s newsletter.
