Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
A few months ago we marvelled at how the AI bacchanal had enticed venture capitalists into moving from backing pre-revenue companies to pre-product companies like Ilya Sutskever’s Safe Superintelligence.
Oh how sweetly innocent we were back then . . . The new hot thing seems to be pre-plan companies. From MainFT on Friday evening, with Alphaville’s emphasis below:
OpenAI’s former chief technology officer Mira Murati has raised $2bn for her new artificial intelligence start-up, in a deal which values the mysterious six-month-old company at $10bn.
The deal, which closed recently, according to multiple people familiar with the transaction, was one of the largest “seed” — or initial — funding rounds in Silicon Valley’s history.
San Francisco-based Thinking Machines Lab had not declared what it was working on, instead using Murati’s name and reputation to attract investors, said those familiar with the fundraise.
. . . Because of its highly clandestine nature, a number of funds that Murati pitched to passed on the deal, said multiple investors who were approached. One of these people added Murati’s pitch offered no information about a product or financial plans.
Another person said Thinking Machines was working on “artificial general intelligence”, a hypothetical point where computers have similar or superior levels of intelligence to humans. But they added that, at the moment, the group was still “strategising”.
Well, at least that $2bn should get Andreessen Horowitz, Conviction Partners and other VCs some oversight rights. Right?
The deal would give Murati a board vote that is structured to be equal to all other board votes combined — plus one. In other words, Murati would have a level of control beyond even the tight grip wielded by supervoting-share-owning founders like Mark Zuckerberg.
Oh.
We’re not even sure what stage of the cycle it is when investors are willing to fork over $2bn to a founder with no disclosed product, financials or business strategy, and bake in zero control.
But this is a world where Meta is willing to make a $14bn acqui-hire of an AI CEO with good connections but uncertain technical prowess, and a small Nvidia-dependent, highly leveraged cloud computing company is valued at $88bn (17x revenues), so maybe Murati’s Thinking Machines will prove a smart bet.
PitchBook’s latest venture capital monitor report showed how AI start-ups are now consuming over 70 per cent of ALL venture dollars being deployed in North America. PitchBook’s database indicates that 454 AI-related companies have been founded already this year. And who can blame them, when VCs are this deal drunk.
We should give Murati props for a good if odd company name though. “Thinking machines” is what the author Frank Herbert called AI-powered robots. In his Dune universe they rose up against humans and had to be defeated in the Butlerian Jihad, leading to a commandment across the universe that “thou shalt not make a machine in the likeness of a human mind”.
Perhaps Murati’s secret plan is actually to soak up as much VC money as possible to avoid it going into AGI development?
