What’s the role of generative AI in drug discovery?

Investing.com -- Generative AI is emerging as a transformative long-term force in drug discovery, a field historically characterized by lengthy timelines, high costs, and low success rates.

According to brokerage Jefferies, the average drug development cycle spans 8 to 10 years, with success rates below 10% and costs exceeding $1 billion per drug. AI has the potential to cut risks by over 50%, accelerating timelines, improving success probabilities, and lowering development costs.

"Generative AI is poised to accelerate the very slow and risky drug discovery process, reducing the time from bench to clinic while increasing the success rate," Jefferies analysts said in a Wednesday note.

The trend is evident across the industry, with major pharmaceutical companies, contract research organizations, and emerging biotechs integrating AI into various stages of development.

From early target identification to compound screening and toxicity prediction, AI platforms are reshaping the pipeline. For instance, Schrodinger (NASDAQ:SDGR) uses a hybrid physics and machine learning approach that enables large-scale virtual screening and significantly reduces preclinical timelines.

Meanwhile, Recursion Pharmaceuticals Inc (NASDAQ:RXRX) runs more than 2 million experiments weekly through its AI-driven platform and supercomputing infrastructure, digitizing biology to streamline drug design and testing.

“With their extensive datasets and computational power, RXRX has introduced several cutting-edge multimodal models. These models provide deeper insights into how cells may respond to new drug candidates, pushing the boundaries of drug discovery and development,” analysts explained.

These platforms not only enable faster drug candidate selection but also support toxicity prediction.

Schrodinger, with support from the Bill & Melinda Gates Foundation and Nvidia (NASDAQ:NVDA), is developing an AI-based predictive toxicology initiative aimed at identifying off-target effects early in development.

Jefferies notes that such tools can enable a "fail early, fast, and cheap" approach that prioritizes safer compounds and improves the odds of clinical success.

Schrodinger expects to launch its predictive toxicology capability in the second half of 2025.

Jefferies also highlights the economic rationale behind AI adoption. “For a hypothetical blockbuster generating $1B in peak sales, pulling the launch forward by 1 year increases the Net Present Value (NPV) by ~20-40% under a standard set of assumptions.”

Earlier launches not only drive additional years of exclusivity but also enhance investment returns.

As regulatory agencies like the FDA show signs of warming to AI-supported applications, Jefferies expects adoption to broaden. The broker estimates AI-related R&D spend at $3–5 billion globally, with the market projected to grow to $8–10 billion in five years and reach $30–40 billion by 2040.

AI is also making strides in personalized medicine, with companies such as Acrivon Therapeutics (NASDAQ:ACRV) and AnaptysBio (NASDAQ:ANAB) using AI to match therapies to patient populations.

As AI matures, Jefferies sees it becoming "a critical tool in drug discovery and precision medicine."