Novo Nordisk’s partnership with Anthropic is best read as a strategic effort to apply frontier artificial intelligence to the slower, more complex parts of pharmaceutical research, rather than limiting the technology to administrative work. The Danish drugmaker announced on 16 September that it would use Anthropic’s Claude models to help speed the discovery and development of new medicines.
Novo Nordisk, known for its diabetes and weight-loss medicines Ozempic and Wegovy, will test the technology in selected research and development workflows. The focus will include scientific reasoning, biological understanding and drug mechanics, areas that sit close to the core work of finding and developing medicines.
Under the agreement, Novo Nordisk scientists and computational teams will give Anthropic specific drug discovery challenges to work on. The companies said they would concentrate on problems where their combined capabilities could have the greatest effect, including working through complex biological questions and understanding how drugs function.
The arrangement builds on Novo Nordisk’s existing use of Claude for trial report generation. The company said that use has reduced patient documentation timelines from months to minutes. The new partnership moves the testing of the technology into research and development work that could affect the speed of drug discovery itself.
The collaboration will also cover artificial intelligence-driven software development across Novo Nordisk. The companies said data governance rules and human oversight would apply, with the stated aim of keeping artificial intelligence use within Novo Nordisk’s compliance and ethics standards.
Anthropic will provide Claude Science, its research workbench, for the work. Claude Science is built on Anthropic’s existing Claude models, including Opus 4.8, rather than being a separate new model. It connects to more than 60 scientific databases and includes prebuilt tools for genomics, proteomics, structural biology and cheminformatics.
The workbench is designed to divide scientific tasks across artificial intelligence systems. Its primary agent can create specialised sub-agents to manage different parts of a task. A separate reviewer agent checks citations and calculations before work is finalised. Those features are intended to support work involving molecular and biological research, where the company is testing Claude’s ability to reason through complicated questions.
Novo Nordisk is adding Anthropic to existing artificial intelligence arrangements with OpenAI and Amazon. That approach gives the company access to competing platforms as it assesses where frontier artificial intelligence can help its research and development operations, while retaining governance and oversight measures around its use.
The financial terms of the Anthropic agreement were not disclosed. The companies have also not set out a timetable for moving from testing to broader implementation, identified particular drug targets or therapeutic areas for the first phase, or given projected efficiency gains from the collaboration. The effect on Novo Nordisk’s existing development pipeline and project timelines is also unknown.
Anthropic has previously provided Bristol-Myers Squibb employees with access to Claude for research and drug development. Novo Nordisk’s agreement adds another pharmaceutical collaboration for Anthropic as drugmakers test whether artificial intelligence tools can shorten development work while operating under their internal compliance and ethics requirements.
Comments (0)