- Elon Musk proposed a $97 billion offer to buy part of OpenAI.
- Lots of folks, including OpenAI's Sam Altman, think Musk is more interested in hurting Altman than actually owning OpenAI.
- But when the world's richest man makes an offer — even if you're not sure it's real — you have to at least think about it.
Here is a question you don't normally ask when someone proposes a $97 billion dollar deal: Is that a serious offer, or a troll?
But it's 2025, and the person making the offer is Elon Musk. So that's a very appropriate question.
The deal in question: A $97.375 billion offer from a consortium, led by the world's richest man, to buy at least part of OpenAI.
The mechanics of this theoretical deal get complicated pretty quickly, since OpenAi is a… complicated organization. As is often the case, Matt Levine at Bloomberg has a very useful explainer, but it boils down to this:
- Right now, OpenAI is a nonprofit organization that controls a very valuable business. It's the group that created ChatGPT and other AI properties, and it has even grander plans.
- OpenAI leader Sam Altman wants to convert OpenAI into a for-profit business, and plans on giving the nonprofit part of OpenAI a stake in the for-profit business.
- Musk and his theoretical co-investors argue that this is a sweetheart deal that will short-change OpenAI, the nonprofit company. So they're offering to buy the nonprofit part of OpenAI for "no less than $97.375 billion," per a letter from Marc Toberoff, an attorney representing the Musk coalition. (I've asked both Toberoff and OpenAI for comment.)
When you have to summarize something with three bullet points, it's already pretty complicated. But it gets more complicated from here, because now we have to guess at whether Musk actually plans on doing this.
The conventional wisdom, which quickly coalesced after Musk made his offer on Monday, is no.
That argument mainly hinges on the idea that Musk is simply trying to jam up Altman, his former ally-turned-rival.
As The Wall Street Journal notes: "Musk's offer could force OpenAI's board of directors to reassess how it is valuing the nonprofit, which the board has said will be fairly compensated in the transaction and own a stake in the for-profit. The higher the valuation of the nonprofit, the bigger its stake would likely be in the for-profit OpenAI following a conversion."
That, in turn, will further complicate a giant funding deal that was supposed to value OpenAI at $300 billion; if Musk's offer raises the value of the non-profit part of OpenAI, it changes the calculations for investors in the for-profit part of OpenAI.
Or, in Altman's words: "I think he is probably just trying to slow us down. He obviously is a competitor."
And there are other reasons to raise at least one eyebrow at Musk's offer. The first is Musk's affinity for saying one thing and doing another. Like last March, when he wrote that he was "not donating money to either candidate for US President" — and then went on to donate more than a quarter of a billion dollars to Donald Trump's presidential campaign.
More specifically, Musk has previously floated deals that either aren't real or that he tried to back out of. Like in 2018, when he said he had "funding secured" to take Tesla private, but he didn't have funding secured, and he didn't take Tesla private. Or in 2022, when he signed a fast-tracked deal to buy Twitter and take it private — and then spent weeks trying to void the deal, saying that he'd been deceived by Twitter's management.
But the flip version of that argument is that betting against Elon Musk doing something is a very dangerous bet.
The standard version of that line is to point to the success of Tesla and SpaceX, two Musk companies that seemed like long-shots when he first engaged with them.
A more recent version is the one Liz Hoffman makes in Semafor: That even Musk's impulse purchase of Twitter, which led to a mass defection by advertisers and plummeting revenue, has actually worked out well. Even if the business itself is a shadow of what it was when Musk bought it, it is now an important political megaphone for Musk. And his investors may end up OK as well, because they're supposed to own a slice of xAI, Musk's own AI company.
I'm not convinced that just because Musk has won in the past doesn't mean you should take him seriously when he says something wild in the present. There's a reason that "past results don't guarantee future performance" is a standard business disclaimer.
But if I were Sam Altman, I'd at least have to game out a future where Musk really does mean what he says.
