Isomorphic Labs just raised $2.1B, but the real tell sits in the February IsoDDE report. The engine claims more than 2x AlphaFold 3 accuracy on the hardest protein ligand generalization benchmark, 2.3x on antibody antigen interfaces, better binding affinity than FEP on curated sets, and novel pocket detection beyond ordinary structure cosplay.

Biotech finance usually worships platform language because platforms let investors defer contact with the corpse field of failed molecules. Here the platform claim has a colder shape. It is trying to turn medicinal chemistry from artisanal hypothesis traffic into an instrumented search problem where geometry, affinity, pockets, and modality selection move under one capitalized machine.

The clinic will still judge it with blood, toxicity, and time. Good. That is where the myth should be sent.