CH Health Tech Advisory

Christian Hein · 17 November 2026 · 2 min read

Isomorphic Labs at $50 Billion: A Bet on Human Data That Mostly Does Not Exist Yet

Isomorphic Labs raised $5 billion at a reported valuation of about $50 billion, with no drug in the clinic yet. I think that price holds only if Isomorphic solves the data problem in human biology.

TL;DR

Isomorphic Labs raised $5 billion today at a reported valuation of about $50 billion, with no drug in the clinic yet. I think that price holds only if Isomorphic solves the data problem in human biology. Rat studies predict human clinical toxicity only about 30% of the time, and little human toxicity data exists. I don't see anything comparable to the Protein Data Bank yet for how drugs behave in people.

Isomorphic Labs raised $5 billion today. Bloomberg reports a valuation of about $50 billion, roughly a fifth of Novartis's market value, for a company with no drug in the clinic yet. I think that price holds only if Isomorphic solves the data problem in human biology.

I told Handelsblatt for today's piece that the valuation is justified if Isomorphic can move most of drug design and optimization onto a computer. The central question is whether they can generate enough data on efficacy in humans for the models to predict reliably.

Where the scaling laws come from and where data runs short

Sequoia, which co-led the round with Greenoaks, says Isomorphic has found scaling laws. President Max Jaderberg explained where those come from on the No Priors podcast last week. Isomorphic generates its own data, in house and with CROs, because the repositories in pharma and the public domain are a biased by-product of decades of making medicines. That works for design. The harder part comes later. Jaderberg says rat studies predict human clinical toxicity only about 30% of the time, and that little data exists on human toxicity because only so many trials have ever been run.

AlphaFold, which won the 2024 Nobel Prize in Chemistry, was trained on protein structures that researchers had deposited in the Protein Data Bank since 1971. I don't see anything comparable yet for how drugs behave in people.

Isomorphic doesn't license its engine. It uses it for its own pipeline and for partnered programs with Lilly, my former colleagues at Novartis, and J&J. So the platform's value will show up in molecules that work in patients, and Isomorphic hasn't given a date for its first trial.

I told Handelsblatt this could change pharma completely, and I believe that. Whether it does comes down to how fast Isomorphic can build data on how drugs behave in people. This could be the first platform shift since the recombinant biotech revolution of the 1980s, and potentially even bigger.

Artificial Intelligence · Machine Learning · Tech Bio · Pre-clinical research · Clinical development · Venture Investment · Digital Transformation · Biotech