Isomorphic Labs talks a $40B valuation for AI drug discovery
A DeepMind spinout is hunting one of the biggest private marks in AI this year, Europe's sovereign-model bet just lost a marquee customer to Chinese open weights, and the math avalanche from OpenAI has the cryptography world digging in.
Isomorphic Labs is in early talks to raise funds at a valuation of at least $40 billion, with one person familiar putting the figure as high as $50 billion. Bloomberg reported the discussions on Wednesday, citing people familiar with the matter, and stressed that terms are not final and could change — Isomorphic did not respond to requests for comment. The timing is what makes the number interesting: the Google DeepMind spinout, best known for building on AlphaFold, closed a $2.1 billion Series B led by Thrive Capital only in May 2026, with Alphabet funding "a large part" of it, on top of a $600 million round in March 2025 — roughly $2.7 billion disclosed in total. A fresh mark at $40 billion and up, five months later, says investors are no longer pricing AI drug discovery as a research bet but as an industry with a winner-take-most shape — and that Isomorphic's platform story, not just its pedigree, is carrying the weight. If it lands, expect a wave of copycat rounds across AI-biology; if the talks stall, that tells you the $40 billion-plus tier of the market is already thinner than the headlines suggest.
Germany's Ecosia is dumping Mistral for open-weight models — including Chinese ones — citing disappointing quality. Founder and CEO Christian Kroll told Politico "we are disappointed with the quality of Mistral", said the French lab is "a year behind", and argued Ecosia was "simply too large a customer for Mistral" to get reliable service, with recurring server overloads. Ecosia now serves Qwen, GLM and Kimi through the Melious platform and says it has roughly halved its costs while improving quality — a sharp turn from May 2026, when it switched from OpenAI to Mistral. Mistral chief scientist Guillaume Lample pushed back publicly, inviting Ecosia to test its new model with early access; Mistral shipped Large 4 this week. What gives the switch weight beyond a vendor squabble: Ecosia is used by Germany's federal environment ministry and the UK's NHS, so Europe's home-grown sovereignty play is losing a government-facing customer on pure quality and price — the one comparison Mistral can't spin.
AI labs have "quietly started probing" whether their models can break important cryptographic protocols, Scott Aaronson reports — and crypto's own researchers are treating it as real. The theoretical computer scientist, citing unnamed sources, wrote that "the AI companies have now started, gingerly and discreetly, investigating whether their latest internal models can break important cryptographic protocols and primitives" — and noted cryptography was "conspicuous by its absence" from OpenAI's math dump. The claim is unverified by the labs themselves, but it lands in the same week as Ethereum researcher Justin Drake's call for the industry to prepare a "bunker mode", with Vitalik Buterin urging that "we should take the risks to cryptography from AI-accelerated math seriously" while telling users not to scramble their wallets. Investor Haseeb Qureshi framed the shift precisely: "The risk is not quantum, but just conventional mathematics overturning unproven cryptographic hardness assumptions." It follows directly on OpenAI's release of 722 manuscripts covering 372 problem families — we covered that drop in OpenAI drops 372 math results, nearly all from one prompt.
What to watch: whether Isomorphic's talks firm into signed terms — and whether any lab confirms or denies the cryptography probing.
If AI-accelerated math is going to break the cryptography under your wallet before quantum does, should labs be required to disclose what they find? Tell us in the comments.
