Basecamp Research, an AI-for-biology startup backed by Anthropic and Nvidia, has raised $140 million to advance a bold idea: using a biological “foundation model” to design medicines that reprogram a patient’s own cells inside their body.
The concept: a foundation model for biology
Just as large language models learned the patterns of human language from vast text, Basecamp is building a model — called EDEN (environmentally-derived evolutionary network) — that learns the patterns of biology to design new biological parts. The aim is in vivo cell therapies: treatments that reprogram cells within the body, rather than the current approach of extracting cells, engineering them in a lab, and reinfusing them. If it works, that could make cell therapy far simpler and cheaper.
The data advantage
AI models are only as good as their training data, and this is Basecamp’s central asset. EDEN was trained on the Trillion Gene Atlas, which the company describes as the world’s largest proprietary biological dataset, assembled through partnerships across more than 30 countries. In an industry where public datasets are widely shared, a large proprietary trove of biological sequences is a meaningful moat. The company reports the model achieved a 63.2% functional hit rate in testing and can design complex DNA sequences, enzymes and peptides.
The financing
The $140 million Series C was oversubscribed and led by Palo Alto firm S32, with a striking roster of backers that signals how AI and biology are converging: Anthropic, NVentures (Nvidia’s venture arm), Catalio, the Rockefeller Foundation, the NATO Innovation Fund and others. The money will fund training next-generation EDEN models and developing in vivo cell therapies for diseases including cancer and autoimmune conditions.
Why it matters
The involvement of an AI lab (Anthropic) and a chipmaker (Nvidia) alongside biotech investors underscores a broader shift: AI companies are moving deeper into drug discovery, and the tools of machine learning are being pointed at designing biology itself. “We believe the future of medicine lies in reprogramming the body to repair itself,” said co-founder Glen Gowers — an ambitious framing for where cell therapy could go.
The caveats
Enthusiasm should be tempered. A high “functional hit rate” in the lab is not the same as a working therapy in a patient, and in vivo cell reprogramming is a hard, largely unproven frontier. AI-designed biological parts still must clear the same brutal gauntlet every drug faces — safety, efficacy, manufacturing, regulation. The funding buys a strong platform and runway; whether it yields approved medicines remains the open question. Business news, not investment advice.