NVIDIA (Intends to) Invest (Up to) $100B in OpenAI (Over Time)
I'm old enough to remember when $40B deals were insanely massive deals. So I'm at least eight months old. That's when word first started to trickle out that OpenAI would raise not just the biggest round in the history of venture capital, it was a round so large that it would be more money than any company had ever raised in any IPO. And it's an amount so large that it's still actually being raised.
Wait, what's that on my phone? A billion push notifications about NVIDIA investing $100B into OpenAI. But headlines can deceive....
To be fair, OpenAI doesn't actually use that in their headline announcing the deal. But everyone else sure seems to! Reuters' is particularly good: "Nvidia to invest $100 billion in OpenAI". Doesn't get more straightforward than that.
But again, it's not that straightforward. "To" is not just doing a lot of work there, it's doing all the work. Because it's not "NVIDIA Invests $100B in OpenAI" – that would tell us that it has happened. "To" means it's going to happen. And it may. Or it may not. But critically, it's not happening right now.
Let's look at what the press release actually says:
NVIDIA and OpenAI today announced a letter of intent for a landmark strategic partnership to deploy at least 10 gigawatts of NVIDIA systems for OpenAI’s next-generation AI infrastructure to train and run its next generation of models on the path to deploying superintelligence. To support this deployment including datacenter and power capacity, NVIDIA intends to invest up to $100 billion in OpenAI as the new NVIDIA systems are deployed. The first phase is targeted to come online in the second half of 2026 using NVIDIA’s Vera Rubin platform.
First and foremost, it's a letter of intent. What's up with OpenAI announcing those of late? (We'll get to that.) But the real key is in the second sentence: "NVIDIA intends to invest up to $100 billion in OpenAI...".
That's not one qualifier, it's two! We get "intends" and "up to". In other words, they may not invest $100B. Or they may not invest the full $100B. Or both. Or neither. And the sentence doesn't end there! "...as the new NVIDIA systems are deployed." In other words, this investment has stipulations.
Of course, so did the $40B round, which is why the money is still coming in months later. These stipulations do seem more straightforward – i.e. deploying systems versus say, converting the structure of the company – but they also seem to come with a time horizon that is far longer. "The first phase is targeted to come online in the second half of 2026 using NVIDIA’s Vera Rubin platform." Phase one starts about a year from now. How many phases will there be? Nobody knows.
This release doesn't even make it clear if NVIDIA's investment here will be for equity or if it's just more of a straight-up business deal. The "in OpenAI" part suggests that it's equity, but we probably need more details there? And it matters for more than the semantics of it all, obviously. We're talking about the most valuable public company in the world potentially owning a sizable stake in the most valuable private company in the world. And they're perhaps the two single most important companies when it comes to the future of AI.
So yeah, it's a big deal.
I'm just simply pointing out that it's not as straightforward of a deal as the headlines may have you believe. Which is perhaps obvious to some, but clearly not to all as messages I'm getting would have me believe. And there are clearly other layers to this – again, the most valuable company in AI is striking a $100B deal with the biggest startup in AI.
Sorry, "up to"!
One of those other layers may be the notion that OpenAI is said to be working on their own AI chips. Undoubtedly not full-on NVIDIA replacements – at least not any time soon – but still, if they can move some of their workloads off of NVIDIA's stack, that's a huge story. And at a high level, it's a move that OpenAI was thought to need to do at some point to try to control compute costs, and thus, their own destiny. And this was not a good narrative for NVIDIA, obviously.
This deal suggests OpenAI now might not be doing that. Or, at least, they may slow down their efforts. And if those efforts were mainly around inference, that also helps NVIDIA because it removes another story about how that's the one area – and perhaps the key area going forward – where they had some weakness and were vulnerable to attack. Now, if it's good enough for the biggest AI company at scale...
Another angle would be on the OpenAI side. Everyone now knows they need to convert to a PBC pronto, and not just for unlocking the aforementioned tranches of capital. Given the aforementioned cash burn, they also undoubtedly need to go public at some point in the near future to fully access all the capital required. These headlines seem to at least take some of that pressure off, optically, if not realistically – because, again, how long will it take this money to come in, and in what form is it coming in? Is it coming in the door just to go right back out the door in the form of NVIDIA chip purchases?
That's, um, an interesting layer in and of itself...
Anyway, that's not to say this is all misleading or nefarious in some way. There is a lot going on here, but at the highest level, there is also undoubtedly real intent for this to be a massive deal and partnership between the two sides too.
In fact, it could end up being one of the biggest business deals of all time! I'm simply pointing out that it's not the simple "investment" that you might think on first read of some headlines. NVIDIA investing $5B into Intel is a more straightforward investment. Hell, even the US Government's investment in Intel is a more typical investment!
This is two companies announcing a massive partnership – which seems tied to a partnership they already announced in the form of the Stargate Project. Sam Altman and Jensen Huang have learned a thing or two from Donald Trump, it seems. That includes the importance of a big headline number – even if it may or may not happen to that full extent. And if it does, even if it will be years from now...
One more thing: there's someone else who will undoubtedly not like these headlines – someone OpenAI would like to steal any spotlight from right now...
M.G. Siegler
M.G. Siegler
M.G. Siegler