Cellular Intelligence has signed four scientists to its advisory board as it works toward a foundation model of cell behavior, with a Parkinson’s therapy already in development as the clearest test of the approach.
Robert Langer, who co-founded Moderna, and Yann LeCun, Meta’s chief AI scientist, are among the new advisers. The same panel takes in Fabian Theis, who directs the Computational Health Center at Helmholtz Munich, and Jens Nielsen, chief executive of the Novo Nordisk Foundation’s BioInnovation Institute. Langer also picks up an observer seat on the board of directors.
The Boston company answered to the name Somite AI until it rebranded in January and published a white paper on what it calls cell fate engineering. Its own framing leans on AlphaFold: protein structure was the first prediction problem to fall, and the company argues the next one is how a cell answers a sequence of biochemical signals.
A capsule-based platform runs many experiments at once. According to the company, the platform has grown from 27,000 experimental conditions to over a million unique sequential signal combinations, and that growth is what feeds its transformer models, which are built for dynamic cellular responses rather than static snapshots.
Funding stands above $62 million, and a Parkinson’s therapy is headed for Phase 2 trials, which gives the platform something to prove outside the lab.
Why it matters. Most AI biology works from fixed measurements taken at a single moment. If a cell’s response can be modeled as a sequence, the ceiling for virtual experiments rises sharply, and so does the value of the data generated to train them.
