The Novo Nordisk and AWS partnership isn't about yet another "cloud migration"—it's about making AI agents a practical tool for scientists. The Danish pharma giant is expanding its use of AI in drug discovery: agents are meant to help identify biological targets, design therapies, and automate research processes.
Cloud, science, and agents: the essence of the partnership
Under the agreement, AWS becomes Novo Nordisk's preferred cloud provider and strategic AI partner. The companies are combining chronic disease research with AWS's cloud infrastructure and AI services for life sciences. This means computational analysis is no longer a separate step—it's being embedded into everyday scientific work.
Joint innovation center in London
For hands-on work, the companies have set up a joint innovation center at Novo Nordisk's London office. There, Novo Nordisk's R&D staff and AWS engineers—including AI specialists, applied scientists, and professional services teams—will work side by side.
The center's core idea is to reduce the handoff between computational analysis and lab research. Instead of spending a long time building models first and then testing them in the lab, researchers will be able to close the "hypothesis—computation—experiment" loop faster. The companies say the goal is to shorten the path from target identification to first dose in humans.
From chatbots to scientific scenarios
Novo Nordisk already has experience deploying generative AI, and it's fairly extensive. More than 25,000 employees use an internal platform built on Amazon Bedrock, which hosts chatbots for over 2,500 use cases. These are mostly non-regulated processes: information retrieval, document preparation, and reference functions.
One major deployment works with roughly 140,000 documents and processes over 26,000 requests per month. In a clinical documents use case, a system using Anthropic's Claude 3.5 model via Amazon Bedrock cut preparation time for certain materials by more than 90%. Previously, this work required 40–50 people and could take up to 15 weeks; now a team of three handles it in minutes. Importantly, medical specialists still review the final documents—so this is about accelerating expertise, not replacing it.
Amazon Bio Discovery: computational biology and lab feedback
The new agreement takes the collaboration far beyond internal assistants. The key innovation is the use of Amazon Bio Discovery, a service for computational biology and drug development powered by AI models and agents. The platform provides access to more than 40 biological models. Agents select and coordinate the right models for specific tasks on their own, and organizations can connect AWS models with their own proprietary models in multi-step processes.
Novo Nordisk plans to use the service to identify potential drug targets, design therapies, and analyze biological information. The platform can generate and rank potential drug candidates before they even reach the lab for synthesis and testing. At the same time, experimental results can feed back into the computational process—allowing models to be refined and linking computer-aided drug design with physical testing.
Additionally, under the agreement, Novo Nordisk will use Amazon Bedrock to build AI applications that work with clinical, genomic, and imaging data. This way, early research results can directly influence the design of future clinical trials. This is an important shift: previously, AI mainly helped employees; now it's being integrated into the scientific loop.
AgentCore: agents in research and operational processes
Another part of the partnership is the Amazon Bedrock AgentCore infrastructure. It's an environment for deploying and managing AI agents: they can work with corporate data, connect to internal systems, and execute multi-step workflows. Novo Nordisk hasn't detailed exactly where it will use AgentCore yet, but plans to deploy it in research and operational processes.
This is a notable expansion of the previous usage model: until now, generative AI deployments at the company were mostly employee-facing. Now the focus is shifting toward scientific and operational drug development scenarios. According to AWS representatives, the goal of the partnership is to apply agentic AI to drug discovery, shorten discovery timelines, handle complex research datasets, and translate results into specific therapeutic areas.
AWS engineers from the Forward Deployed Engineers group will work directly with Novo Nordisk teams on systems built under the agreement. The partnership isn't limited to drug development, though: it extends the broader relationship between Novo Nordisk and Amazon, including Amazon Pharmacy, Amazon Ads, and One Medical. The pharma giant also continues to run its own AI projects outside the AWS ecosystem.
