2026 Is the Year AI Drug Discovery Meets Clinical Reality

Google DeepMind CEO and Isomorphic Labs Founder Demis Hassabis has described his company's long-term ambition as helping "solve all disease" through AI drug discovery company Isomorphic Labs and its AlphaFold platform. In May, Isomorphic Labs raised $2.1 billion in a Series B funding round , one of the largest investments in the sector. At the time of the company's latest public timeline, however, no patients had yet been dosed with an Isomorphic-designed drug after the start of its first clinical trial shifted from late 2025 to late 2026. The contrast reflects where AI drug discovery stands today. Roughly $60 billion has been invested in the sector since 2019, around 175 AI-originated drug programs have entered human trials, and none has yet received FDA approval. Industry estimates indicate that 15 to 20 of those programs could enter pivotal Phase III trials this year, making 2026 an important point for evaluating whether those investments translate into late-stage clinical success. The Proof Problem AI has delivered its clearest gains during the earliest stages of drug discovery. Researchers now use machine learning to identify disease targets, screen candidate molecules, and optimize compounds before laboratory testing begins. Clinical success rates, however, have changed little. A 2022 analysis found that roughly nine of 10 drug candidates entering human trials still fail. Medicinal chemist Derek Lowe has noted that most failures continue to result from efficacy or safety issues that emerge only during human testing. AI drug discovery company BenevolentAI illustrates both the promise and the limitations of the technology. Its platform identified baricitinib as a potential COVID-19 treatment in 2020, but the company's lead eczema program later failed a mid-stage clinical trial after the biological target did not demonstrate sufficient clinical benefit. AI drug discovery company Insilico Medicine has produced a leading example of an end-to-end AI-developed drug candidate. Its experimental therapy rentosertib used AI to identify a biological target and generate the molecule. Insilico Medicine Co-CEO Feng Ren described its Phase II dosing as a milestone for AI-driven drug discovery. Rentosertib remains in Phase IIa, with data collected from 71 patients. The therapy must still complete larger Phase III trials before regulators can evaluate it for approval. Despite that progress, some investors remain cautious. Bessemer Venture Partners Partner Andrew Hedin has cautioned against treating every AI-enabled drug as evidence that the broader field has matured. Researchers also continue debating whether fully autonomous AI drug discovery is technically achievable. A recent ACS Central Science paper from researchers affiliated with Eli Lilly and Insilico Medicine outlines a vision for autonomous "prompt-to-drug" pipelines. Separate academic research on agentic AI systems concludes that current systems remain limited to individual stages of drug development rather than the complete path from target identification through clinical testing. Competition Intensifies Competition has expanded beyond specialist drug discovery companies into the largest AI developers. Anthropic launched Claude Science on 6/30, introducing a research platform designed for life sciences. Shares of computational drug discovery company Schrödinger fell as much as 8.3% during trading following the announcement. The launch places Anthropic alongside Google DeepMind and OpenAI among the largest AI companies pursuing pharmaceutical research partnerships. OpenAI has established a collaboration with Amgen, while Anthropic also introduced a neglected-disease research initiative to support its life sciences strategy. Traditional pharmaceutical companies continue expanding their own investments. In March, Eli Lilly committed $2.75 billion to expand its partnership with Insilico Medicine. Other large drugmakers continue investing in internal AI capabilities and external collaborations. Jared Auclair, Vice Provost for Research Economic Development and Director of Bioinnovation at Northeastern University, has cautioned that general-purpose AI systems can hallucinate or miss important nuances in regulatory guidance or assay design. He says those errors can carry significant consequences in drug development. Why 2026 Matters Several important clinical milestones are expected this year. Recursion Pharmaceuticals plans updates for REC-394 and REC-1245. Schrödinger also continues advancing zasocitinib toward late-stage development. Together, these programs are expected to provide some of the first meaningful evidence of how AI-developed drug candidates perform in advanced clinical testing. The regulatory landscape is also evolving. The FDA is expected to finalize guidance covering AI in drug development during 2026. Some analysts estimate there is roughly a 60% chance that the first AI-designed drug could receive regulatory approval by 2027. Any approval

2026 Is the Year AI Drug Discovery Meets Clinical Reality

The Proof Problem

Keep reading. It's free.

Related stories