Biotech
2026-10-10
GSK said Friday it's expanding its work with Chai Discovery, an AI startup that designs proteins for drugmakers. Chai's models generated candidate protein designs for GSK's own drug targets without any training specific to those targets. GSK then tested the designs in its own labs and found working binders for every target it tried.
An AI that had never seen these specific drug targets still produced working proteins. It got them right on the first real try. That didn't happen in Chai's own tests — it happened in GSK's. Chai isn't new to this. Endpoints reports GSK is its sixth pharma partner in nine months. Eli Lilly, Pfizer, Novartis, Argenx, and Bristol Myers Squibb came before it. Endpoints also reports that string of deals has pushed Chai's valuation to $3.8 billion.
Does a working lab binder actually turn into a real drug? That's still the open question. Designing a protein that binds correctly in a dish is an early, encouraging step, not a finished medicine. Every one of these candidates still has to survive animal studies. Then it faces years of human trials. Only after that would any patient see it. The thing to watch isn't whether AI can design proteins that work in a dish anymore — it increasingly can. It's whether any of Chai's six pharma partners turns one of these designs into an approved drug.
This story is written by AI from the sources above, checked against them before publishing. If something here still reads wrong, tell us and we'll correct it.