AI-designed drug push

Drugmakers are pushing artificial intelligence deeper into the lab, with Isomorphic Labs now targeting its first fully AI-designed drug trial by 2026. (moneycontrol.com) Demis Hassabis, the Google DeepMind chief who also founded Isomorphic Labs, said the company wants to move an AI-designed candidate into human testing on that timeline after earlier comments in January 2025 pointed to trials by the end of 2025. (moneycontrol.com, soci.org) The company is not working alone. Isomorphic Labs said in January 2024 that it signed multi-target small-molecule research deals with Eli Lilly and Novartis worth nearly $3 billion in potential milestone payments, excluding royalties, and it expanded the Novartis collaboration on February 18, 2025. (isomorphiclabs.com, storage.googleapis.com) In drug discovery, the basic problem is finding a molecule that fits a disease target the way a key fits a lock. Google DeepMind and Isomorphic Labs say AlphaFold 3 helps by predicting how proteins, DNA, RNA and drug-like molecules interact before chemists make and test them. (deepmind.google, isomorphiclabs.com) That does not remove the slow part. A drug still has to clear preclinical work, manufacturing checks and human trials, which is why the milestone investors watch is not a model demo but the first patient dosed in a regulated study. (moneycontrol.com, soci.org) Money is following the bet. Isomorphic Labs announced a $600 million funding round on March 31, 2025, led by Thrive Capital with participation from GV and follow-on capital from Alphabet, saying the cash would help advance therapeutic programs into the clinic. (isomorphiclabs.com) A parallel strategy is emerging outside Alphabet: use artificial intelligence to rescue old compounds instead of inventing brand-new ones. Toronto-based Biossil said it has bought or licensed 10 molecules in the past three years, with two already in advanced clinical trials. (europesays.com, biossil.co) Biossil says its current and planned programs span diseases including sickle cell disease, idiopathic pulmonary fibrosis, glioblastoma, breast cancer and Alzheimer’s disease, with 11 Phase 2 and 3 programs on its pipeline page. (biossil.ai, biossil.co) The pitch from both camps is speed: either design better molecules earlier or find new uses for drugs that already failed somewhere else. The test over the next two years is whether those software-guided bets can survive the same clinical and regulatory hurdles that have stalled drug programs for decades. (moneycontrol.com, europesays.com)