Alphabet's Isomorphic Labs prepares for first human trials of AI-designed drugs

Alphabet’s drug discovery venture is about to put its silicon-generated molecules to the ultimate test.

Isomorphic Labs, the Google DeepMind spinoff founded in 2021, is preparing to launch its first-ever human clinical trials for AI-designed cancer drugs, according to company president Colin Murdoch.

From code to the clinic

Born directly out of DeepMind’s Nobel Prize-winning AlphaFold breakthroughs, Isomorphic isn't just modeling biology anymore — it's trying to manufacture cures.

While early versions of AlphaFold predicted static protein structures, AlphaFold 3 unlocked the ability to simulate how proteins bind with DNA and small-molecule drugs. Murdoch confirmed that teams of machine learning researchers and pharma veterans in London are already collaborating with AI models to map out novel oncology treatments.

To fund that ambition, Isomorphic pulled in $600M in April 2025 in a round led by Thrive Capital. That capital sits alongside massive multi-billion dollar research deals inked in 2024 with pharmaceutical giants Novartis and Eli Lilly.

Isomorphic isn't merely acting as a software vendor for big pharma. The company is actively pursuing its own internal pipeline focusing on oncology and immunology, with plans to advance candidates through early-stage human trials before licensing them out. Murdoch noted that pharma typically suffers from a brutal 10% success rate once candidates enter human trials; Isomorphic’s goal is to simulate away that failure rate before a single patient is dosed.

Murdoch even floated the sci-fi endgame: one day, scientists could theoretically pick an illness, push a button, and watch an AI generate the exact compound to treat it.

Why it matters

AI drug discovery has spent the last five years generating endless digital hype, but molecules in a simulation don't cure diseases — molecules in people do. Taking an entirely in-house AI-designed compound into human dosing marks the true litmus test for whether AlphaFold's computational brilliance translates to messy human biology.

For the biopharma landscape, the stakes are enormous. Developing a drug today routinely takes over a decade and billions of dollars, with failure baked into the business model. If Isomorphic can meaningfully move the needle on clinical trial success rates, it will force traditional drugmakers to overhaul their entire bench-science pipeline.

We think Isomorphic’s hybrid strategy — taking Big Pharma's partnership cash while keeping internal oncology candidates for eventual out-licensing — is a masterclass in risk management. However, "solving all diseases with a click of a button" sets an impossibly high bar. AI can accelerate candidate discovery in seconds, but human clinical recruitment, toxicology, and multi-year patient trials still obey the unyielding laws of biology.