The Incredible Valuation Heights of the OpenAI "Constellations"
Last March, I set out to map the OpenAI "Constellations" – that is, the startups not only in OpenAI's orbit, but those directly tied to it, founded by entrepreneurs that had previously been at OpenAI. No surprise, given both our moment in time with AI and the current funding environment around AI, there are a lot. So I set a threshold at the $1B valuation mark. By my math back then – again, a whole 10 months ago – those entities had a combined valuation of around $200B. And when you added in the valuation of OpenAI itself at the time, we were right at the $500B mark.
Just six months later, those numbers had changed drastically, so I checked back in on those valuations. At the end of August, that $200B aggregate valuation was past $400B. And with OpenAI's updated valuation, we were right around $1T in combined value.
Well, we're not quite six months since then, but given that Anthropic, xAI, and OpenAI itself – the three main drivers of those valuations, obviously – have all either just raised or are in the process of raising again, let's check back in...
Ursa Super Major
- Anthropic was in the process of raising the round that would value them at $183B back then. Just months later, they're working on a new round which will value them at $350B – before the investment. And while that process may have been kicked off by NVIDIA and Microsoft – yes, the same Microsoft that owns 27% of OpenAI – committing to invest, that money will apparently be on top of this new $10B that they're likely to raise here from other investors. If history is any indication, Anthropic will raise more than $10B, but let's just keep it there for now and add it to the $15B in commitments from Microsoft and NVIDIA, and that gives them a new post-money of $375B.
- xAI was said to be trying to raise at $170B to $200B last August, which I gave them the credit for because of course Elon Musk was going to be able to raise whatever he wanted, comps be damned! Well, it took a bit longer, but they have now announced the $20B Series E, which they really, really would like you to know exceeded their target. They didn't announce the valuation, but reports put it "above $230B" which is weirdly worded in there, but I take to mean $235B.
Given that these two companies formed by those formerly affiliated with OpenAI are valued above $600B by themselves, it's probably worth breaking them into their own category, and we know how much AI companies love to append "Super" to everything, so let's go with "Ursa Super Major".
Ursa Major
- Safe Superintelligence, speaking of "super", it seems like Ilya Sutskever's startup is still "stuck" at the $32B valuation from their "seed" round. I'll comically note that this seems almost prudent given the rate at which everyone else is raising. Sutskever may be giving his company some optionality, as a few of the big players would undoubtedly still be open to acquiring them – well, "hackquiring" at least – in this general price range, to get access to Sutskever if nothing else. And given that they're down a co-founder (and first CEO) thanks to Meta's mad scramble to catch up in AI (which included trying to acquire Safe Superintelligence, but instead they ended up investing – before they poached Daniel Gross...). Sutskever has opened up a bit more about what SSI is doing recently, and that included an explanation of how they can compete without needing to raise endless capital, like their rivals. Have I mentioned that Sutskever's own OpenAI shares may be worth $100B now, so he can probably fund SSI's "return to research" himself. Still, don't be shocked if he raises again sometime soon and resets their valuation...
- Thinking Machines Lab, would provide the reason for SSI to raise again if they're able to complete a round valuing them at $50B or $60B! That was rumored a few months ago already and doesn't seem to have happened yet. So I'll keep them behind SSI for now on the list, but give them credit for the reported valuation jump from the "mere" $12B valuation "seed" round in July. They have also lost a co-founder, also to Meta, since then. But hey, at least they have a product in market now. And promises of more to come this year...
- Perplexity would also have been vaulted if Thinking Machines gets that round done. Their most recent round reportedly pushed their valuation to $20B, up from the $18B just a couple months prior (when I last checked in). One of my predictions for 2026 is that someone will buy Perplexity, as they may need to step off the hamster wheel of funding. Samsung is the most obvious acquirer given their partnership (and great financial situation at the moment), but they've also been going deeper with Google... just as the other would-have-been buyer, Apple, has been...
These combined numbers add up to over $100B now, just for Ursa Major, without the 'Supers'!
Ursa Minor
- Harvey jumped up to an $8B valuation in December, up from $5B at the last check in.
- Periodic Labs may be a bit above the $1B mark during the last check-in as the round seems to have grown from a rumored $200M to a confirmed $300M. Some reports suggest $1.5B or even perhaps $2B.
- Cresta is one I actually missed in the previous reports, probably because it was co-founded way back in 2017 by early OpenAI researcher Tim Shi. It currently sits at a $1.6B valuation, which it hit back in 2022.
- Eureka Labs still hasn't raised any outside capital, it seems. Though Andrej Karpathy obviously could at any time, probably at any valuation.
So that's a solid haul for this group just over $10B.
Okay, so grouped all together, we're looking at an aggregate valuation of just over $730B – a number which has nearly quadrupled in less than a year. And nearly doubled just since August. Obviously it's top-heavy, but even taking out the 'Supers', you'd still have a group valued collectively at $123B. Would that alone be the highest valued group post-PayPal Mafia? Or even if you remove Elon from the full equation, you're looking at almost exactly a $500B aggregate valuation for the OpenAI diaspora.
And if you add OpenAI itself into the mix, with their newly targeted $830B valuation, well, you're truly in the stars now: just over $1.5T.
That's awfully close to Meta's market cap. And if/when OpenAI and Anthropic go public, both may shoot past that mark by themselves given likely investor demand (depending on timing). And if/when xAI merges with Tesla, making them an AI company as well...
M.G. Siegler
M.G. Siegler
M.G. Siegler