Novo Nordisk partners with Anthropic on Claude for drug discovery

Novo Nordisk said it will collaborate with Anthropic so Novo can use Anthropic models and Claude Science to advance scientific reasoning in research and development, and use Anthropic frontier models to strengthen AI-driven software development — with data governance and human oversight built in (company).

HEALTH desk — a top obesity/diabetes drugmaker is wiring frontier Claude models into R&D and software engineering with named governance, a concrete same-day example of AI moving deeper into how medicines get discovered.

What the companies say they will do together: jointly address key drug discovery challenges identified by scientists and computational teams at Novo; develop targeted solutions for specific scientific challenges and workflows to support biological reasoning; and help accelerate discovery and development of new medicines. File that jointly-address / targeted-solutions / biological-reasoning picture as Novo’s. This desk did not name a drug program, a target, or a timeline. None was printed.

Initial aim, still company: Novo will test Claude Science for specific workflows in research and development — R&D, the lab-to-candidate work of finding and developing a medicine — and address scientific problems where the joint capabilities of Novo and Anthropic are expected to have the greatest impact. Claude Science, as this release uses the name, is Anthropic’s AI workbench for scientific workflows, not a finished drug and not a regulator decision. File that test-Claude-Science / selected-R&D-workflows picture as Novo’s. This desk did not run a workflow.

Software engineering, still Novo: the company will also use Anthropic’s frontier models to strengthen AI-driven software development, which the release calls a key enabler for scaling AI across the company. Frontier models here means Anthropic’s most capable general models, used to write and ship software, not to approve a medicine. File that software-engineering / scale-AI-across-the-company line as Novo’s. This desk did not watch a coding agent at Novo.

Governance, company: the collaboration was designed with robust data governance and human oversight, helping ensure AI is applied responsibly and in line with Novo’s ethical and compliance standards. File that governance / human-oversight / ethical-and-compliance picture as Novo’s. This desk did not audit a data-sharing contract and is not inventing a sole-vendor term, a dollar fee, or a regulatory filing. None was printed.

Named voice on the release: Mike Doustdar, president and CEO of Novo. He says the partnership is another testament to Novo’s ambition to become “the world’s most AI-driven healthcare company,” and that joining forces with Anthropic will supercharge Novo’s R&D organisation on its mission to bring new, transformative health solutions to people living with chronic diseases. File the name, title, and that most-AI-driven / supercharge-R&D line as his, via Novo. His other sentences stay in Sources.

Named voice, still the same page: Dario Amodei, co-founder and CEO of Anthropic. He says that by giving leading researchers access to safe, capable, and trusted frontier models, the two companies can shorten research timelines, improve outcomes, and discover medicines and treatments that significantly improve human life — and that this collaboration is built to achieve that. File the name, title, and that shorten-timelines / improve-outcomes line as his, via Novo. His other sentences stay in Sources.

Same-day independent wires confirm the announce: Reuters and Bloomberg both reported the Novo–Anthropic collaboration on 16 Sep 2026. File those as Tier B confirm of the same-day newsroom — not as a substitute primary, and not as a source for invented deal value or clinical results. Reuters-only color that Novo already uses Claude to automate trial-report generation and trim patient-documentation timelines from months to minutes stays attributed to Reuters. Bloomberg-only obesity-market catch-up framing stays attributed to Bloomberg. Neither extra is used as a Novo primary claim.

Plain English for the rest of the card: Claude Science = Anthropic’s AI workbench for scientific workflows. R&D = research and development, the lab-to-candidate work of finding and developing a medicine. frontier models = Anthropic’s most capable general models, used here for software engineering as well as science. biological reasoning = using models to help scientists think through living systems and drug mechanics. data governance = rules for how Novo’s research data is stored, shared, and used. human oversight = people stay in the loop; the model does not approve a medicine. This is a vendor collaboration announce, not a new drug approval and not an efficacy claim.

PRIMARY here: Novo Nordisk’s 16 Sep 2026 newsroom release “Novo and Anthropic will collaborate to advance drug discovery with Claude” — Tier A PRIMARY company source, the original record (news id=916768). Reuters and Bloomberg are same-day independent Tier B confirm of the partnership, the Claude Science R&D test aim, and the frontier-model software-engineering use — not a substitute primary, and not a source for invented dollar terms, exclusivity, or clinical trial outcomes. The collaboration to accelerate new medicines and AI-driven software development, the jointly-address / targeted-solutions / biological-reasoning job, the Claude Science R&D-workflow test, the frontier-model software-engineering line, the data-governance / human-oversight design, and the Doustdar and Amodei titles and attributed lines are company-attributed. Product behavior and any existing Claude use inside Novo stay company- or wire-attributed — not independently audited here. NOT claimed: a deal value, exclusivity, a named drug program, a clinical-trial result, a new drug approval, an efficacy claim, that AI cures obesity, that this desk tested Claude Science at Novo, a stock tip, or investment advice. Distinct from the already-filed inductive-bio-indy-chemistry, insilico-longevity-vaccines, openai-foundation-public-data-health, sorcero-fall-2026, and claude-cowork-chat-merge.