Novo Nordisk partners with OpenAI
- Novo Nordisk said on April 14, 2026 it formed a strategic partnership with OpenAI to apply AI across drug discovery, manufacturing and commercial operations. - Bristol Myers Squibb said on May 20 it would deploy Anthropic's Claude across research, clinical development, manufacturing and commercial functions enterprise-wide. - Eli Lilly expanded TuneLab access through a May 20 integration with CDD Vault, adding another named platform in pharma AI.
Novo Nordisk’s partnership with OpenAI is less notable for the branding than for where the company says it will use the technology. The Danish drugmaker said on April 14 it would apply OpenAI’s models across global operations, from early drug discovery to manufacturing and commercial work, with a stated focus on analyzing complex data, improving target identification and accelerating development of therapies for chronic diseases. That scope matters because large pharmaceutical companies have spent years testing AI in narrow pilots. The latest announcements show several of them moving toward broader, operational deployments tied to research systems, internal knowledge and development workflows. Bristol Myers Squibb said on May 20 it would deploy Anthropic’s Claude across research, clinical development, manufacturing, commercial and corporate functions. Eli Lilly, through a separate May 20 agreement, expanded access to its TuneLab AI/ML models by integrating them into CDD Vault. (biopharminternational.com) ### Where does Novo say OpenAI will be used? Novo Nordisk said the OpenAI partnership will run across “global operations,” not only discovery labs. BioPharm International reported that the company plans to use advanced AI to analyze complex data, improve target identification and speed development of chronic-disease therapies. The company’s framing suggests a workflow play as much as a science play. (news.bms.com) BioPharm said Novo described the tie-up as spanning early drug discovery through manufacturing and commercial execution, which places the models inside multiple stages of the pipeline rather than at a single research step. ### Why does that stand out in pharma now? Bristol Myers Squibb used unusually broad language in its Anthropic announcement on May 20. (biopharminternational.com) The company said Claude Enterprise would become a “shared intelligence platform” across its global operations and that it would use agentic capabilities to connect people, systems and institutional knowledge at enterprise scale. Lilly’s TuneLab expansion points in the same direction. CDD said on May 20 that Lilly agreed to integrate TuneLab into CDD Vault, giving participating companies access to Lilly’s AI/ML drug discovery models through a platform already used for research data management. Lilly’s TuneLab site describes the service as a collaborative platform offering access to Lilly’s own models. (news.bms.com) ### Is this still mostly hype, or are AI-built drugs reaching the clinic? BioPharm International pointed to an earlier benchmark in AI drug development: DSP-1181, developed by Exscientia and Sumitomo Dainippon Pharma for obsessive compulsive disorder, was reported as expected to enter clinical trials after machine learning shortened parts of the design process. (finance.yahoo.com) That example is old, but it remains one of the clearest cited cases of an AI-designed candidate advancing into human testing. The current significance is that large drugmakers are now pairing model access with enterprise systems and internal data environments, rather than describing AI only as a discovery-screening tool. ### What does this change inside drug R&D? (biopharminternational.com) Drug discovery runs on fragmented data, long timelines and repeated go/no-go decisions. Novo said its OpenAI deal is aimed at candidate identification and data-driven decision-making, which puts the emphasis on narrowing targets faster and using large data sets more systematically. The comparison with Bristol Myers and Lilly shows the same operational pattern. (biopharminternational.com) One company is positioning a frontier model as a cross-company intelligence layer; another is packaging internal models for broader external use through an established research platform. Novo’s move fits that same buildout of AI as R&D infrastructure. ### What comes next? The next test is not another partnership announcement but whether these systems show up in named programs, trial candidates or measurable development timelines. For now, the concrete milestones are already on the calendar: Novo disclosed its OpenAI partnership on April 14, Bristol Myers announced its Anthropic deployment on May 20, and Lilly’s latest TuneLab expansion was announced the same day. (biopharminternational.com) (news.bms.com)