Danish pharmaceutical company Novo has partnered with AI developer Anthropic to bring Claude models into drug discovery and software engineering workflows. The deal, which follows Novo's earlier agreement with OpenAI, reflects the pharmaceutical industry's accelerating investment in AI tools that could shorten the path from research to approved medicine.
Claude enters the R&D pipeline
The collaboration will first target scientific challenges identified by Novo's researchers and computational teams. The companies plan to build AI solutions for specific research workflows, including tools that support biological reasoning. Novo will test Claude Science on problems where the two organisations believe AI can have the greatest impact, such as analysing complex scientific questions and understanding human biology and drug mechanisms.
Novo CEO Mike Doustdar said AI could increase productivity within R&D and shorten the path between research and a marketed medicine. Beyond efficiency, he sees opportunities for AI to support new forms of scientific reasoning. Anthropic CEO and co-founder Dario Amodei said increasingly capable AI models could shorten research timelines and contribute to the discovery of new medicines and treatments. How much these tools can improve drug development in practice will require further evaluation.
The partnership includes data governance measures and human oversight intended to keep the technology aligned with Novo's ethical and compliance standards. Novo also plans to use Anthropic's models for software development, which the company views as an important part of scaling AI across its organisation.
A broader push into AI infrastructure
The Anthropic agreement is part of Novo's larger strategy to embed AI throughout pharmaceutical research. Earlier this year, the company reached a deal with OpenAI to use GPT-Rosalind, a scientific AI model designed to analyse complex datasets and help identify potential drug candidates. In 2024, the Novo Nordisk Foundation partnered with Nvidia and the Export and Investment Fund of Denmark to establish the Danish Centre for AI Innovation, which operates Gefion, Denmark's first AI-ready supercomputer.
Anthropic is also expanding its life sciences footprint. In May 2026, the company announced a collaboration with Bristol Myers Squibb to deploy Claude across research, clinical development and manufacturing. The Novo collaboration underscores the industry's growing interest in AI throughout R&D, though the current focus remains on testing how the technology can complement researchers rather than proving it can independently deliver faster drug development. For research scientists and R&D engineers looking to build expertise in these tools, AI Scientific Research Courses and AI R&D Engineering Courses offer structured learning paths aligned with these developments.
AI's expanding role in preclinical research
Separate research published in Cell Systems last year demonstrated how AI can analyse changes in 3D cell shape to predict drug responses. The system combines 3D imaging with deep learning to examine cellular "fingerprints" that provide richer detail than conventional 2D imaging. After analysing nearly 100,000 3D images of melanoma cells, the AI identified which drug had been applied with 99.3 percent accuracy, including subtle differences between drugs with similar effects. The approach was also tested on red blood cells, brain vessels and stem cells, suggesting broader applications.
Researchers estimate the technology could substantially shorten drug development, particularly the preclinical phase, while helping identify promising treatments earlier. The Institute of Cancer Research is developing the technology further through its Center for Cancer Drug Discovery and spin-off company Sentinal4D, with the aim of applying it to cancer drug development and potentially other diseases.
Why this matters for healthcare, science and research professionals
These developments signal a shift in how pharmaceutical R&D teams will work. AI models are moving from experimental pilots into active research workflows - not as replacements for scientists, but as tools for analysing complex biological data and generating hypotheses faster. For professionals in drug discovery, preclinical research, or computational biology, familiarity with how large language models and deep learning systems integrate into lab and data environments is becoming a practical requirement, not a future consideration. The technology's real-world impact on timelines and success rates remains unproven, but the infrastructure and partnerships are being built now.
Your membership also unlocks:
