Novo Nordisk Just Hired the Brainiest AI in the Room
Novo Nordisk partnered with Anthropic to deploy Claude AI across its drug development pipeline, adding to deals with OpenAI and NVIDIA. The world's most valuable pharma company wants to become 'the most AI-driven healthcare company,' and the broader industry is sprinting to keep up.
A Major Pharma Company Made a Very Expensive New Friend
Imagine you're running a restaurant kitchen with 100 chefs. Each one is brilliant, but they all spend half their time flipping through recipe books instead of actually cooking. Now imagine handing each chef a sous-chef who's already memorized every cookbook ever written, can prep ingredients in seconds, and never takes a bathroom break.
That's roughly what Novo Nordisk is trying to do with its drug development pipeline.
On September 16, the Danish pharma giant announced a collaboration with Anthropic, the AI company behind the Claude family of models, to deploy Claude's AI tools across drug discovery and development. The stated goal is almost absurdly ambitious: Novo wants to become "the world's most AI-driven healthcare company."
Bold words from a company that already ranks among the pharma industry's biggest players, largely on the back of its blockbuster obesity and diabetes drugs. But this isn't just corporate puffery. It's the latest (and arguably biggest) signal that Big Pharma is done treating AI like a science fair project and is ready to make it core infrastructure.
What Exactly Did They Agree To?
The short answer: a lot of ambition, not a lot of disclosed numbers.
Novo and Anthropic said they'll work together to tackle key drug discovery challenges identified by Novo's scientists and computational teams. The plan is to develop targeted solutions for specific scientific workflows, starting in R&D areas where Claude Science (Anthropic's research-focused AI workbench) can have the biggest impact.
Beyond the lab, Novo also plans to use Anthropic's frontier models to improve its software engineering capabilities, which matters more than it sounds. Scaling AI across a global pharma company requires serious internal tooling, and better software development accelerates everything else.
What we don't know: the financial terms. No contract value, no upfront payment, no milestones, no exclusivity details, no duration. The announcement focused on functional goals and governance (robust data oversight, human-in-the-loop safeguards, ethical compliance), but the business side stayed behind the curtain.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
That's not unusual for partnerships like this, but it does mean we're grading this on strategy rather than dollars for now.
Claude Science: The Swiss Army Knife for Lab Coats
So what is Novo actually getting? Claude Science isn't a single AI model. It's a research workbench built on top of Anthropic's Claude models, designed specifically for scientists.
Think of it as a turbocharged research assistant that can:
Search and summarize biomedical literature, then generate testable hypotheses from it
Run bioinformatics and genomics workflows, like sequence analysis and related pipelines
Render 3D protein structures for structural biology work
Handle chemical structures for drug design and optimization
Connect to over 60 scientific databases, so researchers don't have to toggle between a dozen different tools
The workbench also keeps a record of the code, environment, and conversation history behind every output. In a regulated industry like pharma, that reproducibility and audit trail is the difference between a useful experiment and a compliance nightmare. Anthropic's flagship models now support context windows of up to 1 million tokens, which means Claude can digest entire protocol documents, large literature sets, and multi-document analyses in one sitting.
Put simply: if drug development is a treasure hunt, Claude Science is offering Novo a metal detector, a map, and a very fast pair of legs.
The AI Arms Race in Pharma Is Getting Ridiculous (in a Good Way)
Novo didn't wake up one morning and decide AI was cool. The company has been stacking partnerships like a kid collecting Pokémon cards.
In January 2025, it expanded a collaboration with Valo Health for AI-driven obesity and diabetes research. That same year, it partnered with NVIDIA to build a GPU-powered supercomputer for drug discovery simulations. In April 2026, Novo inked a broad deal with OpenAI covering discovery, clinical development, manufacturing, supply chain, and corporate functions. And it's already rolled out Microsoft Copilot to more than 20,000 employees.
The Anthropic partnership adds yet another layer, and the pattern is clear: Novo isn't betting on one AI horse. It's building a stable.
But Novo isn't the only one galloping. The broader pharma industry has gone all-in on AI partnerships over the past two years, and the deals are getting bigger and more ambitious:
Pfizer expanded its AI collaboration with PostEra in January 2025 in a deal worth up to $610 million, targeting antibody-drug conjugate payload design
Roche's Genentech struck a deal with Orionis Biosciences worth over $2 billion in milestones to discover molecular glues using AI
AstraZeneca partnered with Modella AI for oncology, then liked the results so much it acquired the company in January 2026
Merck signed a multiyear collaboration with Google Cloud valued at up to $1 billion to deploy Gemini Enterprise across R&D, manufacturing, and commercial operations
The common thread? Pharma companies aren't licensing narrow AI point solutions anymore. They're trying to embed foundation models across entire organizations, from target identification all the way through regulatory submissions.
Why Anthropic, Though?
This is where it gets interesting. Novo already has a deal with OpenAI. So why add Anthropic to the roster?
Anthropic has been making aggressive moves in healthcare and life sciences throughout 2025 and 2026. It launched Claude for Life Sciences in late 2025 and Claude for Healthcare (with HIPAA-ready compliance tooling) in January 2026. The company's reported pharma client list reads like a who's who: AstraZeneca, Sanofi, Genmab, Bristol Myers Squibb, and Eli Lilly have all been linked to Claude deployments.
Anthropic also built an implementation ecosystem through partnerships with AWS, Google Cloud, Accenture, and others, making it easier for regulated industries to actually deploy the technology at scale. And it expanded into academic research through collaborations with the Allen Institute and Howard Hughes Medical Institute.
Compared to its competitors, Anthropic's edge seems to be industry-specific packaging. While OpenAI and Google compete through broad enterprise offerings, Anthropic has tailored its tools more explicitly for life sciences workflows: literature synthesis, genomics, structural biology, clinical operations, and regulatory compliance.
For Novo, running multiple AI partnerships isn't redundancy; it's hedging. Different models have different strengths, and the field is evolving so fast that locking into a single provider would be like choosing your fantasy football lineup in April. You want options.
The Bull Case and the Bear Case
Let's be honest about what's exciting and what's uncertain.
The bull case is straightforward. Drug development takes an average of 10 to 15 years and costs billions. If AI can compress even a fraction of that timeline (faster target identification, quicker hypothesis testing, more efficient clinical trial design), the return on investment is enormous. If Novo can move faster than competitors in expanding beyond obesity and diabetes into next-generation therapies, AI could be the engine that gets it there.
The bear case is equally straightforward. This collaboration is early-stage. It's focused on "targeted workflows," which is corporate-speak for "we're still figuring out where this works best." No financial terms were disclosed, which makes it hard to gauge how much skin either side has in the game. And the practical impact on clinical results or financial performance could take years to materialize.
The honest answer is probably somewhere in the middle. AI won't magically produce a blockbuster drug next quarter. But the companies that figure out how to integrate these tools into their R&D machinery now will have a meaningful advantage when the technology matures. It's less about today's output and more about building the muscle.
What This Means for the Rest of Us
Zoom out for a second. A pharma giant just added a third major AI partnership to its portfolio in less than two years. It's investing in supercomputers, foundation models, and enterprise-wide AI deployment simultaneously.
This isn't a pilot program anymore. This is infrastructure.
The signal for the broader industry is unmistakable: AI is no longer optional in pharma R&D. It's table stakes. Companies that treat it as a side project risk falling behind on everything from target discovery to regulatory submissions.
For patients, the promise is faster, cheaper drug development. For investors, it's a bet on R&D productivity improvements that could reshape pipeline economics. And for the AI companies themselves (Anthropic, OpenAI, Google), pharma is shaping up to be one of the most lucrative enterprise verticals on the planet.
Novo Nordisk wants to be the most AI-driven healthcare company in the world. Whether or not it earns that title, one thing is clear: the race to claim it just got a whole lot more crowded.
Funding & Financings4 min read
The Pill That Wants to Kill a $2 Billion Needle
Sling Therapeutics just raised $123 million to develop an oral pill that could replace IV infusions for thyroid eye disease. With Phase 2 data already in hand and a $2 billion market in its sights, this small company is picking a fight with Amgen and Viridian.
