Isomorphic Labs prepping AI drug trials

Drug discovery is the business of turning a molecular hunch into something safe enough to put into a human body. That sounds obvious, but it’s the whole bottleneck. Isomorphic Labs — the Alphabet-owned company spun out of DeepMind — says it is moving its AI-designed drug programs toward first-in-human trials. But the key update is not that trials have started. It’s that the company has spent the last year shifting from “the model can design molecules” to “the company can do the ugly real-world work needed before dosing a patient.” ### What is Isomorphic Labs actually building? Isomorphic Labs is not a chatbot for chemists. It is an AI-first drug design company built around DeepMind’s biology work, including AlphaFold, with the aim of designing small-molecule medicines across multiple disease areas and then developing some of those medicines itself. The company also works with big pharma partners, which matters because those partners bring clinical-development muscle that pure software companies usually lack. (clinicaltrialsarena.com) ### So what changed? The change is that the company’s own language has gotten more concrete. In July 2025, president Colin Murdoch said Isomorphic was preparing to dose its first patients and “staffing up” for trials, with oncology programs leading the way. Then in January 2026, Demis Hassabis gave a firmer outer bound — first clinical trials by the end of 2026. That is later than the end-2025 target he had discussed earlier. (isomorphiclabs.com) ### Why does the delay matter? Because it tells you where the hard part is. Designing a promising molecule on a computer is only the opening move. Before a human trial, a company still has to run preclinical toxicology, manufacturing, formulation, dosing work, and regulatory prep. Basically, AI can compress the search phase, but it does not erase the clinical-development gauntlet. The move from “very close” to “by the end of 2026” is a reminder that biology and regulation still set the pace. (clinicaltrialsarena.com) ### Why is $600 million such a big clue? Because money tells you what phase a company thinks it is entering. Isomorphic raised $600 million on March 31, 2025, in its first external funding round, led by Thrive Capital with GV and Alphabet participating. The company said the cash would help advance therapeutic programs into the clinic. That is not discovery-lab language. That is scale-up language — hiring, preclinical packages, manufacturing, and all the expensive steps between a design engine and an investigational drug filing. (marketscreener.com) ### What do the Lilly and Novartis deals add? Proof that pharma thinks the platform is worth paying for. In January 2024, Isomorphic announced collaborations with Eli Lilly and Novartis worth nearly $3 billion in potential value, excluding royalties. Those deals do not prove the drugs work. But they do show that established drugmakers were willing to put real money behind the company’s design engine well before any human efficacy data existed. (isomorphiclabs.com) ### Why is human testing the real threshold? Because this is where the story stops being about elegant predictions and starts being about patients. Drug discovery is a bit like solving a lock on a screen, then finding out the real lock is rusty, warped, and attached to a door in a storm. Plenty of molecules look great before they hit living systems. The first human trial will not prove AI has “solved” drug discovery. It will prove Isomorphic has crossed from computational promise into translational medicine. (isomorphiclabs.com) ### Is this an AI milestone or just normal biotech? It’s both. The AI angle is real — Isomorphic exists because DeepMind’s biology tools made structure-based design much more powerful. But once a candidate gets near the clinic, the company starts to look a lot more like a normal biotech. It needs capital, trial operations, regulatory discipline, and patience. Turns out the future still has to pass through the FDA-shaped present. (marketscreener.com) ### Bottom line? The news here is not that Isomorphic Labs has already begun human trials. It hasn’t. The real story is that its timeline is now clearer, its funding is in place, and the company is doing the unglamorous prep work required to test whether AI-designed drugs survive contact with human biology. If that first trial starts by late 2026, the field gets a real benchmark instead of another demo. (marketscreener.com) (isomorphiclabs.com)