Bernstein Report: 70 AI-Driven Drug Projects in Clinical Trials, Real Outcomes Expected in 3–5 Years

According to Bernstein’s research report dated August 26, 2026, approximately 70 AI-driven drug development projects have entered clinical pipelines, with the vast majority still in early stages—the real test lies in the next three to five years. Recently, the CEOs of DeepMind and Anthropic claimed that “most diseases will be cured within 5 to 10 years,” but Bernstein views such statements more as defenses of AI’s societal value rather than grounded assertions supported by the realities of drug development. From target identification to clinical trials to healthcare system capacity, each layer presents a bottleneck; biology ultimately must be validated in humans, and this pace cannot be compressed by algorithms. Bernstein believes AI’s core value in pharmaceuticals lies in improving decision quality, not merely accelerating processes. A 20% increase in success rates generates far greater R&D returns than equivalent reductions in time or cost. If AI is applied to tackle difficult-to-treat diseases where biological understanding remains limited, even modest overall success rates can generate substantial value. Major pharmaceutical companies show significant variation in their AI strategies: Eli Lilly has the broadest external partnerships (25), Roche has made the largest computational investments, and Amgen benefits from the deepest data moat via deCODE Genetics. AI has not yet produced measurable impacts on FDA approval data—recent annual approvals averaging 48.5 new drugs reflect broader regulatory modernization and the rise of biologics. While AI is already contributing to molecular design, virtual screening, and target identification, “curing all diseases” requires overcoming multiple bottlenecks—this is not a problem solvable by algorithmic iteration alone.