HyperAIHyperAI

Command Palette

Search for a command to run...

Alphabet's AI Drug Firm Isomorphic Labs Seeks $5B Before First Clinical Trials

Alphabet-backed Isomorphic Labs is in talks to raise about $5 billion at a potential $50 billion valuation, according to Bloomberg. The scale of the proposed round stands out because the company has not yet advanced any drug into human clinical trials, while the global AI drug discovery market was estimated at about $2.3 billion in 2025. The raise would deepen the company’s capital base as it tries to prove that artificial intelligence can move beyond molecular modeling and improve the odds of developing real medicines. The valuation bet rests on both AI capability and pharmaceutical experience. Isomorphic Labs is led by CEO Demis Hassabis, also chief executive of Google DeepMind and a 2024 Nobel laureate for AlphaFold-related work. President Max Jaderberg, a former DeepMind researcher, manages day-to-day operations. The company has assembled drug-development specialists including chief scientific officer Miles Congreve, a co-inventor of breast-cancer drug Kisqali; chief research officer Sarah Skerratt; chief medical officer Ben Wolf; and chief technology officer Sergei Yakneen. Its scientific advisory board includes Nobel laureates Jennifer Doudna and David MacMillan. Isomorphic Labs has said it does not sell API access or token services. Its core AI engine, IsoDDE, is intended to design drug candidates rather than function as a software product. The company is pursuing two tracks: internal programs in oncology, immune disease and inflammation, and partnerships with major drugmakers such as Eli Lilly, Novartis and Johnson & Johnson. Under those deals, partners contribute disease biology and development targets while Isomorphic Labs designs candidate molecules and earns upfront payments, milestone fees and possible future royalties. The key question is whether AI can reduce the high failure rate of drug development. In traditional pipelines, roughly 90 percent of compounds that reach human testing fail, often because of toxicity or lack of efficacy. Jaderberg argues that conventional methods rely too heavily on trial and error and that AI models combining reasoning and physical simulation can test more molecular designs computationally before experiments begin. The company says IsoDDE can predict protein-ligand interactions, antibody-antigen interfaces and binding affinities. In its Runs N’Poses benchmark, Isomorphic Labs claims the system achieved structure-prediction accuracy more than twice that of AlphaFold 3 on difficult generalization cases, and outperformed AlphaFold 3 and Boltz-2 in certain high-precision antibody-antigen tests. Columbia University computational biologist Mohammed AlQuraishi praised the reported results but noted the lack of technical details needed for independent replication. Isomorphic Labs plans to start its first human clinical trials by late 2026. That timeline will mark the transition from a heavily capitalized AI research organization into a clinical-stage drug developer. The market is betting that advanced AI can shorten years of failed experiments, but the company’s $50 billion valuation remains contingent on experimental data, regulatory progress and the ability to turn predictive models into effective, safe medicines.

Related Links