Artificial intelligence is slashing the time and expense of bringing new drugs to market, according to a survey of biopharma executives that reveals how deeply the technology is reshaping early-stage research. Preclinical costs and timelines are compressing by as much as 70%, creating a surge in demand for AI-powered software, sequencing tools, and computational models that promise to flood pipelines with experimental treatments over the next five years.
The TD Cowen survey of 80 industry leaders found that the largest opportunity lies in advanced software that can simulate biological processes and predict drug interactions, toxicity profiles, and dosage adjustments for vulnerable populations like newborns and pregnant women. These so-called "in silico" platforms allow scientists to run thousands of virtual experiments in seconds, replacing months of traditional lab work.
The economic stakes are substantial. Industry executives project new drug development programs could expand by more than 10% within three to five years, with companies expected to funnel at least an additional $1 billion into software, sequencing, and lab infrastructure to support the computational shift.
"The goal is to create more shots on goal," explains Brendan Smith, director of life sciences equity research at TD Cowen, "and to generate enough data to train AI models that increase the odds of clinical success." This continuous feedback loop between computation and laboratory validation is fundamentally changing how drug designers approach their work. Computer screens are becoming as central to the process as the traditional wet labs where scientists still must ultimately test compounds in living systems.
The transformation carries particular significance as the Trump administration signals its intent to reduce animal testing in biomedical research. That policy shift is expected to accelerate adoption of computational alternatives, 3D human tissue models, and other tools that can predict how compounds behave without relying on animal trials.
Yet the technology comes with real caveats. No AI system has yet discovered a drug that won FDA approval, and some investors remain skeptical about whether algorithmic optimization can adequately account for the complexity of human biology. A core concern is that all the engineering and tweaking happening on computers may not sufficiently capture how different people respond to drugs before candidates enter human trials. Without that deeper insight, skeptics warn, the roughly 90% failure rate for new drugs in clinical development could persist.
Competition is intensifying on the global stage. China's biotech sector continues building capacity at lower costs and faster speeds, already attracting billions in fresh investment and threatening U.S. research dominance. Meanwhile, the administration is navigating conflicting pressures to maintain a light regulatory touch on AI while simultaneously addressing concerns about safety and privacy, creating uncertainty about the long-term policy environment.
Author James Rodriguez: "AI is turbocharged drug development, but turning code into cures requires the skepticism to match the hype."
Comments