Moderna's Bob Langer and AI pioneer Yann LeCun join board of Cellular Intelligence, bringing AI into medicine via a Parkinson's drug
Fortune Catherina Gioino
Moderna co-founder Bob Langer and AI star Yann LeCun just joined Cellular Intelligence’s advisory board. The bet: AI may help turn living cells into faster, cheaper drug makers.
Based on reporting by Fortune, Catherina Gioino — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Cellular Intelligence is making a blunt, ambitious claim: if AI can learn language and images, it can learn cells. The Boston startup wants a model that predicts how living cells will behave and, eventually, how to steer them toward useful medical outcomes. That’s the pitch behind a company built around biology, data and machine learning rather than the old lab ritual of testing one possibility after another.
On Monday, it gave that pitch more credibility. Robert “Bob” Langer, the Moderna co-founder and longtime MIT Institute Professor, joined the company’s scientific advisory board, along with Yann LeCun, the Turing Award-winning AI researcher who helped build Meta’s Fundamental AI Research lab. Jens Nielsen, who runs the Novo Nordisk Foundation’s BioInnovation Institute, and Fabian Theis, a computational biologist at Helmholtz Munich, also joined the board. Langer is additionally taking an observer seat on the board of directors.
The mix is the point. LeCun brings AI muscle, Langer brings decades in medicine and biotech, and Cellular Intelligence is trying to sit in the middle. CEO Micha Breakstone described the company as “a marriage of three disciplines: genomics, developmental biology and AI.” He also framed the core idea in familiar AI terms: a “world model” that learns how a system behaves and then predicts what happens when someone acts on it. In this case, the system is a living cell.
There’s a commercial hook under the science. The company says it has a proprietary engine for generating biological data over time, a foundation model trained on that data, and its own clinical program to test whether the predictions matter in the real world. Langer said internal benchmarks already look better than existing models on healthy development and cancer, even though the system was never trained on a single cancer cell. That’s the kind of line startups love to put on slides. The harder part is turning it into medicine.
The most concrete proof point right now is STEM-PD, an experimental Parkinson’s cell therapy Cellular Intelligence picked up in May from Novo Nordisk, which had shelved it after closing its cell-therapy division last October. The program is Phase 2-ready, has FDA Fast Track designation and IND clearance, and remains the company’s only owned clinical asset. Novo took an equity stake in the deal and can still collect future milestones and royalties. Cellular Intelligence, founded in 2023 and formerly called Somite.ai, says it has raised more than $70 million from backers including Khosla Ventures, the Chan Zuckerberg Initiative and Novo Nordisk.
My take — AI-written commentary, not fact-checked reporting
This is the biotech version of hiring both the chef and the food critic to fix dinner. The AI crowd keeps promising “world models” for everything, and medicine is the easiest place to sound profound because patients are real and timelines are long. If Cellular Intelligence can make the biology legible instead of just trendy, fine; if not, it’s another expensive slide deck with a lab coat on it.
Read more about this at: Fortune