Powering AI-Driven Drug Discovery with President of Isomorphic Labs Max Jaderberg

Max Jaderberg argues that drug discovery is still dominated by empirical trial and error, which is why costs have risen even as scientific knowledge has exploded. He says the industry’s core bottleneck is that failures only become visible in patients, when toxicity or lack of efficacy can no longer be hidden by preclinical proxies. Isomorphic Labs is trying to move that work onto computers by building predictive models that reach experimental accuracy before a molecule is synthesized or tested.

Jaderberg says the company’s bet is that AI can create compounding learning across programs, instead of treating each drug as a one-off frontier science project. He describes a hiring philosophy built around mixing core AI researchers with biology and chemistry experts, then training them to work together until cross-disciplinary scientists emerge. Internally, the company is focused on oncology, immunology, and inflammation, but the drug design engine is intended to be disease agnostic, target class agnostic, and modality agnostic.

He also argues that better molecular engineering could unlock safer drug combinations, including multi-drug cocktails that hit complex pathways at once. On validation, he says the real-world still matters for proving safety and efficacy, but the goal is to reduce how much learning has to happen there. He expects major progress in predictive biology and precise, less toxic molecule design within five years.