Drug discovery is entering a new phase where artificial intelligence (AI) is no longer a separate tool but an integrated partner working alongside scientists at every stage of research. Rather than simply predicting molecular properties, AI agents now collaborate with researchers to design experiments, synthesize data, and move from scientific questions to validated discoveries in a fraction of the traditional timeline (Source 1, 2, 3).

What's Driving the Shift From AI Tools to AI Agents?

The distinction matters. Earlier AI applications in drug discovery focused on prediction: identifying promising compounds or forecasting how a molecule might behave. Today's agentic AI systems go further, actively participating in the scientific workflow itself. These AI agents can reason over experimental history, design new experiments, execute computational biology tasks, and collaborate with human scientists in real time.

This shift reflects a recognition that the bottleneck in drug discovery isn't just computational power or better models. It's the speed at which scientists can move from asking a question to gathering evidence to making a decision. Ono Pharmaceutical, a Japanese company with over 300 years of scientific heritage, recently partnered with Phylo, an AI platform company, to embed agentic AI into the hands of every discovery scientist. The goal is straightforward: help researchers move faster without sacrificing scientific rigor.

"We believe AI will become a core capability for drug discovery. With patients waiting for new medicines, there is an urgent need to help scientists move faster without compromising scientific rigor," said Seishi Katsumata, Corporate Officer and Executive Vice President of Discovery and Research at Ono.

Seishi Katsumata, Corporate Officer and Executive Vice President of Discovery and Research, Ono Pharmaceutical

How Are Leading Pharma Companies Building AI Infrastructure?

The largest pharmaceutical companies are making substantial investments to embed AI into the fabric of their R&D operations. Bristol Myers Squibb announced plans to build what it describes as the most powerful AI factory in life sciences, partnering with NVIDIA, while Eli Lilly has invested in dedicated AI infrastructure and deployment. These are not simply larger computing clusters. They represent a strategic recognition that secure access to proprietary scientific data, advanced AI models, and the ability to deploy those models at enterprise scale are now core competitive capabilities.

At the same time, research institutions are strengthening their leadership to bridge the gap between discovery and translation. The Quantitative Biosciences Institute (QBI) at the UCSF School of Pharmacy appointed two nationally recognized leaders to newly created executive roles aimed at accelerating the translation of discoveries into therapies. John A. Young, a former Roche pharmaceutical executive with over a decade of experience in infectious disease research, became chief strategy officer. Andy Kilianski, who previously oversaw more than one billion dollars in biomedical research investments at the Advanced Research Projects Agency for Health (ARPA-H), became chief science and technology officer.

"John and Andy each bring an exceptional combination of scientific expertise, leadership experience, and translational vision. Together, they will help expand QBI's impact by building the partnerships and technological capabilities needed to move groundbreaking discoveries from the laboratory toward improving human health," said Nevan Krogan, director of QBI.

Nevan Krogan, Director, Quantitative Biosciences Institute at UCSF

Steps to Integrate AI Into Drug Discovery Workflows

• Secure Infrastructure First: Organizations must establish secure, scalable computing environments that allow scientists to develop and deploy AI capabilities while protecting proprietary data and maintaining scientific integrity across active discovery programs.
• Embed AI Into the Learning Loop: Rather than treating AI as a separate analysis step, integrate it across the entire discovery process, from initial ideation and molecular design through prediction, experimentation, and decision-making to compress the cycle from question to evidence.
• Focus on Validated Learning Over Speed Alone: AI can accelerate discovery, but the real advantage comes from moving teams more quickly from question to validated evidence to better scientific decisions, not simply generating results faster without judgment.
• Build Cross-Functional Leadership: Recruit leaders who understand both deep scientific expertise and translational vision, capable of bridging academic discovery, industry partnerships, and government policy to move discoveries toward patient impact.

Why Speed of Learning Matters More Than Raw Speed?

A critical insight emerging from industry leaders is that the advantage of AI in drug discovery is not simply speed. Abbas Kazimi, CEO of Nimbus Therapeutics, a drug discovery company integrating AI into its operations, explained that the real value lies in the speed of validated learning. Every output still must survive contact with the laboratory bench and the clinic. Without judgment, speed can simply generate more noise.

At Nimbus, AI is already delivering concrete results. The team built and deployed a custom molecular property model in nine days to resolve a design liability in one active program. In another program, coding agents helped rewrite and extend a large codebase in about a week, compressing work that might previously have taken months. Across the portfolio, predictive models increasingly guide which compounds scientists synthesize, reducing wasted effort on molecules unlikely to succeed.

"AI can compress that cycle by moving teams more quickly from question to evidence to decision. However, the advantage isn't simply speed. It's the speed of validated learning and every output still has to survive contact with the bench and the clinic," noted Abbas Kazimi, CEO of Nimbus Therapeutics.

Abbas Kazimi, CEO, Nimbus Therapeutics

The broader context matters too. Drug discovery typically represents only the first four years of a decade or more of total R&D, while late-stage clinical development accounts for most of the timeline and cost. Accelerating early discovery is valuable, but improving the quality of decisions and advancing better molecules into development may have an even greater impact by reducing the risks that dominate later stages of R&D.

What Separates Hype From Real Progress?

The pace of progress at the intersection of AI and medicine has been remarkable, with new scientific workbenches, evolving protein-design systems, generative chemistry platforms, and increasingly capable coding agents fueling large financing rounds and high-profile appointments. Yet honest questions persist among experienced drug hunters: How much is hype? Do we understand why drug candidates fail well enough to know where AI can help? And most importantly, where are the medicines?

The answer, according to industry leaders, is that the question is no longer whether AI can augment drug discovery. The question is what prevents organizations from effectively accessing these tools, learning to use them well, and applying them to scientific questions that matter. When powerful AI models become broadly available, what will differentiate one drug discovery organization from another is not the tools themselves, but how quickly an organization turns those tools into validated learning and better scientific decisions.

QBI's research platforms span proteomics, structural biology, CRISPR engineering, and AI-driven target discovery, positioning the institute to develop collaborations with biopharmaceutical companies and public health organizations. The institute's collaborative approach has already helped launch biotechnology companies, translating discoveries made at UCSF into new opportunities for drug development. With strengthened leadership focused on translational vision and industry partnerships, QBI is positioned to accelerate the path from laboratory discovery to patient impact.

The convergence of institutional investment, leadership talent, and agentic AI platforms suggests that drug discovery is entering a new era. The medicines that result from these accelerated discovery processes will ultimately determine whether the current wave of AI investment in pharma represents genuine progress or inflated expectations. For now, the infrastructure is being built, the partnerships are forming, and the first concrete results are emerging from organizations embedding AI into their everyday scientific work.