Chris Butler: Isomorphic's in-house AI models keep drug work on track

Isomorphic Labs was founded in 2021 as a DeepMind spinoff, and its origins under DeepMind founder Demis Hassabis place it inside the same AI lineage that OpenAI and Anthropic — the companies whose calls to pace frontier work Butler was asked about — also come from. Butler's remarks at the Semafor Future of Health Forum on Tuesday draw a line between those public slowdown debates and Isomorphic's own operations: because its models are locked down in-house, he argues, decisions made elsewhere do not determine its pace. That framing matters because it separates the policy conversation around frontier AI from the day-to-day work of a company whose output is measured in drug candidates, not model releases.

The $2.1 billion round in May, led by Thrive Capital at an undisclosed valuation, gives Isomorphic the means to act on that independence. Butler says the money goes toward building out the pipeline and evolving drug design, and the team is growing with particular interest in oncology and immunology. Those are two of the largest and most competitive therapeutic areas, so the hiring and pipeline language points to where the company expects its in-house models to prove themselves first.

The stakes in Butler's own words are progress toward the clinic. That is the test he names, and it is the point at which an AI-designed approach stops being a capability claim and starts being a candidate that can be tested in people. Whether in-house models shorten that path is not something Butler asserts as an outcome — it is the bet the funding and the hiring are meant to support.