Silicon Valley Promises AI Will Cure Disease. Drugmaker CEOs Aren't Buying It

Silicon Valley leaders predict that AI could eliminate most disease within a decade, but pharmaceutical executives say those claims vastly overstate what the technology can do. No AI-designed drug has received FDA approval, and Vertex Pharmaceuticals’ experience is a warning: after testing tools from roughly 50 AI companies on data covering 10,000 compounds, it found none that produced a workable drug. Drugmakers do use AI to help identify targets, improve experimental medicines and select trial participants, but they see it as an aid to scientists, not a substitute for them. Executives argue that biology is too complex and incompletely understood for data analysis alone to deliver cures.

Investment in AI drug discovery reached $8.4 billion globally in 2025, roughly twice the 2023 total, while only a handful of AI-focused biotech firms have advanced treatments to late-stage trials. AI can design candidates faster, but speed does not solve major barriers: current systems struggle to predict unintended effects, which contribute to about a quarter of mid- to late-stage drug failures. Major pharmaceutical companies are nevertheless striking partnerships with AI firms, and companies such as AstraZeneca are using the technology in research and trial recruitment. Bayer used AI to analyze 100 million patient records in days and identify diseases suited to gene editing. Its executives say the value comes from combining AI’s analytical reach with human scientific judgment, not expecting the technology to discover cures on its own.