AI shows promise in drug labs; commercial payoff years away

Only one in 10 drug candidates reaches market

Artificial intelligence has already accelerated parts of the drug discovery process that humans know how to do — predicting protein folding, identifying potential drug targets and screening millions of molecules virtually — but the technology has yet to demonstrate it can change the core economics of drug development, according to industry executives, researchers and investors.

At Roche’s Genentech, computational biologist Aviv Regev said she has built a “lab in the loop” in which AI models predict promising targets and molecules, researchers test those predictions experimentally, and the resulting data improve the next round of models. “AI is not smarter,” Regev said in an interview. “But what helps our scientists is that it encodes information very, very broadly.”

The approach has expanded the range of research programs scientists can realistically pursue, Regev said. But expanding discovery does not directly address the industry’s most persistent problem: turning ideas into medicines that work in people. Only about one in 10 drug candidates that enter human trials reach the market, and failures often occur after years of research and billions of dollars have been spent.

Jack Scannell, who co-wrote the 2012 paper that coined the term Eroom’s Law — Moore’s Law spelled backward — described the data challenge facing AI in drug development. Training an AI on today’s biological data, he said, is “like trying to train your Waymo for San Francisco by getting a frog to ride a bike around Albuquerque.” Whereas autonomous vehicles have millions of miles of real-world driving data, medicine lacks a clean map of the human body, he added.

The distinction between AI’s laboratory productivity and its financial impact creates a dilemma for investors, the Wall Street Journal reported. Pharma CEOs including Eli Lilly’s David Ricks and Novartis’s Vas Narasimhan are committing billions to computing power and new research platforms. Goldman Sachs estimates that the present value of AI’s benefits to drug development could reach $400 billion over the next decade by shortening timelines, lowering costs and improving the odds that medicines succeed, according to the report.

Eric Kauderer-Abrams, who leads life sciences at AI company Anthropic, said the industry is in the “second inning” of AI’s potential. He said AI could bend the curve by attacking multiple bottlenecks to boost a drug’s clinical probability of success, but acknowledged that proving its potential will take several more years.

Biotech investor Rod Wong, managing partner at RTW Investments, said competition from China could become a catalyst for change. China’s advantage, he said, is in the speed at which companies can move from research ideas to clinical evidence. That pressure, he said, could force the U.S. to rethink a clinical trial system that has become increasingly slow and expensive.

If those pieces come together — richer human data, better ways to monitor patients, and faster clinical trials — Eroom’s Law might finally begin to bend, the report concluded. But the winners may not be the companies that build AI models, but rather the largest established pharmaceutical companies that combine powerful tools with deep biological data and global development programs, according to Citi healthcare strategist Traver Davis. Biology runs on its own clock, not the semiconductor cycle. The revolution might yet be real, but we won’t know for certain for several more years.