Drugmaker CEOs say biology limits what AI can do in drug discovery

Venture capital and private equity invested $8.4 billion globally in AI drug discovery in 2025

No drug designed by AI has won clearance from the Food and Drug Administration, and executives across the industry say human biology is too complicated for the technology to work alone. Drugmakers are using AI to pick targets, fine-tune experimental medicines and identify patients likely to benefit from treatments.

Vertex Pharmaceuticals Executive Chairman Jeff Leiden said the company used roughly 50 AI companies to comb through its data set of 10,000 drug compounds. Not one generated a working drug. “It just doesn’t work by itself,” Leiden said, “and any company that says they can do that—at least in our experience, we’ve looked at most of them—it’s not true.”

Leiden said the technology is useful as a scientist’s co-pilot. Drugmaker executives said their concern is whether AI can independently discover a drug that works, remains safe enough to advance through clinical testing and receives FDA clearance.

McKinsey Senior Partner Alex Devereson, who focuses on AI in drug development, said the technology can design experimental medicines more quickly but does not necessarily produce workable drugs. Around a quarter of failures in mid-stage to late-stage drug development can be traced to safety problems caused by off-target effects, meaning drugs do unintended things inside the body that can lead to unwanted or harmful side effects. Current AI models struggle to predict those effects.

Anthropic Chief Executive Dario Amodei posted on X in September: “I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound.” An unnamed technology executive joked that AI researchers had started smoking cigarettes because they were so confident AI would cure lung cancer. Leiden, recalling a recent conversation with a technology leader who claimed AI would cure 90% of human disease in the next seven years, called that premise “That’s complete BS.”

Executives leading the companies that discover new drugs said AI was unlikely to cure cancer anytime soon and that human biology remained too complicated even for superintelligence. Many of the unknowns in biology, they said, cannot be fixed through data analysis.

McKinsey data show that venture capital and private-equity firms invested $8.4 billion globally in AI drug discovery in 2025, up from roughly $4 billion in 2023. Only a few AI-native biotechs, including Insilico Medicine and Generate Biomedicines, have advanced drugs into late-stage patient testing, and the FDA has not cleared any of them. “Investment is flowing to the problems that are easiest to solve, not the problems that are most important,” Devereson said.

Technology companies have continued to expand their drug-development operations despite the caution from drugmakers. Silicon Valley’s largest players see drug development as a lucrative opportunity prime for disruption. Anthropic bought a stealth-mode biotech earlier this year and recently bought a one-person biotech operating almost entirely with AI. OpenAI announced last month that it had bought an AI drug-discovery biotech.

Most pharmaceutical companies also have partnerships with AI developers. Eli Lilly and Nvidia said in January that they would invest up to $1 billion over five years to research AI applications including drug discovery and robotics. Bristol Myers Squibb, Novo Nordisk and Sanofi have all announced partnerships with OpenAI or Anthropic.

AstraZeneca Chief Executive Pascal Soriot said, “If we are in the industry long enough, we quickly learn to be humble.” He added, “Disease is complicated, biology is complicated. To say AI is going to cure cancer is maybe a little bit of a jump.” AstraZeneca is still using AI. The British drugmaker recently signed a deal with AI biotech Insitro worth $40 million up front to discover new heart and weight-loss drugs. AstraZeneca also signed an agreement with Oracle to streamline the recruitment of patients for clinical trials, and a recently opened research and development site in Cambridge, Mass., uses AI to speed up experiments.

Regeneron Chief Scientific Officer George Yancopoulos, the holder of more than 1,000 patents, sees potential for AI to assist in laboratories but does not expect it to transform drug development on its own. “AI is not going to just miraculously come in and tell us how to do things and cut things from 15 years to five years,” Yancopoulos said.

A team at Germany-based Bayer’s gene therapy unit recently used AI to comb through 100 million patient records in a matter of days. The system examined genetic markers and patient histories to identify the five diseases best suited to Bayer’s gene-editing technology. Sebastian Guth, Bayer’s chief operating officer, said that application should not be confused with AI curing a disease. “Is that AI as such curing a disease? It isn’t,” Guth said. “The magic happens when you marry the brilliance of human scientists with AI.”

Regulators have recently proposed new frameworks governing AI use in regulatory submissions and trial-data analysis. Those proposals do not resolve more difficult parts of clinical development, such as recruiting enough of the right patients and monitoring them throughout studies.

As the technology evolves, AI is likely to become a more useful and potentially more powerful tool in drug development. The current record, however, is one of assistance to researchers rather than autonomous disease cures.