AI drug discovery has attracted billions in investment over the last decade.
The thesis was straightforward: if AI can find better drug candidates faster, the economics of pharmaceutical R&D change permanently. That thesis has proven correct. AI-led discovery pipelines are now generating more candidates than the industry can develop.
That last part is worth paying attention to.
Having a candidate is not the same as having a drug. Every candidate still needs to clear preclinical studies, clinical trials, regulatory review, and real-world deployment. That process takes 12 to 15 years and costs between $1 billion and $2.6 billion per drug, regardless of how the candidate was found.
Discovery and validation move at very different speeds. AI has accelerated one. The other has not changed much.
Validation has always been the more expensive, more time-consuming part of drug development. And the coordination infrastructure that determines whether any of those candidates ever reach a patient, including the data rails, compliance frameworks, and clinical networks, has seen far less investment than the discovery side.
The validation gap is real. Whether it is a science problem, an infrastructure problem, or something in between is a question the industry is only beginning to ask seriously.
17
2
2
1July 1, 2026 973 2
