Isomorphic Labs unveils AI engine to design drugs end to end

What’s the deal? Isomorphic Labs, the London-based AI drug discovery startup spun out of DeepMind, has launched a new AI platform called the Isomorphic Labs Drug Design Engine. It says the system improves on AlphaFold 3 in predicting protein-ligand interactions and binding affinity, key steps in early drug design.

The company claims the platform can identify new binding pockets using only a protein sequence and do so faster and at lower cost than traditional lab methods. It plans to use the engine to advance its own drug candidates towards clinical trials.

Isomorphic Labs raised $600 million in 2025 in its first external funding round. The capital is being used to scale its technology and internal pipeline.

Why now? The launch builds on the momentum of AlphaFold 3, released in 2024, which expanded AI’s role from protein structure prediction to modelling molecular interactions. Drugmakers are under pressure to cut R&D costs and timelines, making AI-led discovery more attractive.

The company has also signed research partnerships with Novartis and Eli Lilly, giving it commercial validation as it moves from platform development to pipeline delivery.

What could go wrong? Drug development remains slow, expensive, and highly regulated. Strong computational predictions do not guarantee clinical success.

Isomorphic Labs has delayed the start of its first clinical trials to 2026, highlighting the gap between algorithmic design and human testing. Other AI drug startups have struggled to translate technical advances into approved therapies.

The signal: AI drug discovery is shifting from point tools to integrated platforms that aim to design and evaluate full drug candidates. Investors are backing companies that combine proprietary models, in-house pipelines, and pharma partnerships.

If platforms like this can produce viable clinical assets, they could change the economics of early-stage drug discovery. Until then, proof will come not from benchmarks, but from trial data.

Source:
Isomorphic Labs

A.M.