Microsoft's Would-Have-Been $30B "Hackquisition" of OpenAI
When looking back at the origins of the "hackquisition" model for acquiring startups, Microsoft was clearly the godfather of such deals. Their March 2024 deal to not buy Inflection, but instead to bring on much of their team was groundbreaking in this regard, and it has led basically all of their rivals to copy (or expand upon) the same general structure.1 But actually, there was almost a "hackquisition" that Microsoft did before the Inflection deal... OpenAI.
I've long thought that the Inflection deal structure, and thus, the "hackquisition", was born out of a deal Microsoft nearly had to execute under duress. In November 2023, OpenAI's board ousted CEO Sam Altman and the ensuing weekend saw Microsoft, led by Satya Nadella on down, scramble to save the situation. The company had invested around $13B into OpenAI at the time – a lot of money in those days2 – and they were worried it was going to go to zero overnight.3 So they implemented a plan. Either OpenAI would reinstate Altman, or Microsoft would just hire the entire company.4
Sound familiar?
But wait, don't "hackquisitions" also involve some level of IP rights? You know, "non-exclusive licensing agreements"? For show, if nothing else? Yeah, Microsoft didn't need those here because they already had far more rights to OpenAI's IP, thanks to their deal terms following their massive investments. So yes, this would have been the first "hackquisition". And now we know how much it would have cost.
Thanks to text messages from OpenAI COO Brad Lightcap, uncovered as part of Elon Musk's lawsuit against the company, to Nadella and Altman, we have the number. $25B. Or $29B if they wanted to include Ilya Sutskever. Which of course Microsoft would have, but given his role in said coup d'état of Altman, it's not clear that would have happened...
As relayed by Rocket Drew of The Information:
Lightcap disclosed Sutskever’s financial stake after the OpenAI board fired CEO Sam Altman and Microsoft CEO Satya Nadella prepared to hire Altman and OpenAI colleagues. Lightcap wrote to Altman and Nadella that paying OpenAI employees for their equity would require $25 billion excluding Sutskever’s equity, or $29 billion including Sutskever’s vested units. (It’s not clear how much Sutskever had in unvested equity.)
The day following the exchange, Microsoft chief technology officer Kevin Scott said publicly that Microsoft would hire and match the compensation of every OpenAI employee who chose to resign from the startup, in the event that Altman was not reinstated at OpenAI. Those hires never happened because the OpenAI board rehired Altman later that night.
So yeah, Microsoft would have paid at least $25B, and perhaps $29B to do that would-have-been "hackquisition". That's not too surprising since OpenAI was valued at... $29B during secondary sales that were taking place at the time. But the math wouldn't have been quite that straightforward.
The key bit to remember here is that no one actually owned equity in OpenAI at the time. This was long before the conversion to the public benefit corporation, and everyone from Microsoft on down simply had percentages of the rights to future profits. Microsoft was clearly willing to "true up" those employees and match the on-paper value, and seemingly assume that those employee shares equated to 100% ownership of the company. Of course, there were other investors in OpenAI, even beyond Microsoft. So the bigger question remains if they would have given those other investors anything.5
That is, of course, another key component to the "hackquisition", paying out investors to ensure they go along with the deal. Given how freaked out those investors were during "The Blip", it seems reasonable to think they believed they'd be getting nothing for their investments. But it's also probably reasonable to think that Microsoft would have given them something to avoid future litigation, if nothing else. Again, this is what Microsoft did with Inflection investors several months later.
And if that's the case, that probably would have pushed the deal well past $30B – especially if you consider compensation for Altman, who famously had no direct equity – or even "Profit Participation Units" – in OpenAI at the time (and apparently still doesn't). But if Sutskever was going to get $4B... Microsoft probably would have wanted to keep Altman happy, in order to keep him.
Anyway, there was one massive shareholder Microsoft didn't need to pay out: Microsoft. Everyone now knows they own 27% of OpenAI, but again, at the time no one had any equity. Microsoft had the right to 49% of those profits (though actually a much higher percentage until their initial investment was paid back) and undoubtedly could have and would have argued that meant they "owned" 49% of the company. Unless it was regulators who were asking.
Had anyone else tried to "hackquire" – or even actually acquire – OpenAI, they probably would have wanted at least $15B for their troubles, if not more. And it highlights just how killer of a deal Microsoft would have gotten here.
Call it $30B - $35B to effectively buy OpenAI in late 2023. Given that the company is currently raising around an $850B valuation just over two years later...
Alas, by the end of the weekend, Altman was back in place atop OpenAI and "The Blip" was just that, a blip. This is clearly the outcome Nadella and Microsoft wanted, undoubtedly due to regulatory concerns if nothing else. But the mad scramble also pretty clearly planted the seed that would grow into the "hackquisition" model.
One more thing: if Sutskever's stake was worth $4B back at that $29B valuation, it would be worth $117.2B at that $850B valuation. Obviously, there has been a lot of dilution, but he also stayed and vested another six months before he was off to start Safe Superintelligence... So let's guess that Sutskever's OpenAI's stake is likely worth around $100B now – assuming he didn't sell in any of these tenders...
Update January 8, 2026: Further testimony and messages reveal even more about Satya Nadella's motivations after Sam Altman was ousted: he was teriffied that Altman and Greg Brockman might spin-up a competitive lab – epsecially since they were getting calls from the leadership at Google and Meta. So he quickly had offers drawn up for the pair...
1 Notably bringing on board OpenAI co-founder and Inflection CEO Mustafa Suleyman to head up Microsoft's AI efforts, presaging Google doing a similar move with Character.AI's Noam Shazeer and Meta doing that with Scale AI's Alexandr Wang... As an aside, the $650M price point seems quaint today! I mean, NVIDIA just did a "hackquisition" for $20B! ↩
2 Also sort of wild that Microsoft apparently hasn't really committed more money since that point. Sure, they've "participated" in some subsequent fundraises, but presumably only token amounts. That shows just how clearly they've outsourced the support of OpenAI to SoftBank and others post "Blip". They clearly decided they needed to try to own AI themselves... ↩
3 To be clear and fair, a lot of that money was in compute credits, and future capital commits. So Microsoft undoubtedly wouldn't have actually lost $13B, but still many billions. And arguably a lot more if you take into account where OpenAI is valued today... ↩
4 Well, anyone who wanted to come aboard. Which, judging from the immediate reaction from the employee base, seemed to be pretty much everyone. ↩
5 And I won't even go into the "grants" Elon Musk made to OpenAI at its inception. He clearly – clearly – would be suing Microsoft right now over those, either to block/unwind any deal, or get paid. But it may have been a harder argument without any legal equity conversion... ↩