iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Novo Nordisk says AI now works alongside its laboratories in three places: screening very large numbers of virtual molecules before any compound is made, refining candidate molecules through design cycles, and analysing clinical-trial data to understand disease and select patients. The most specific figures come from the company’s own 2024 Annual Report. Its partnerships with Valo Health, Microsoft, NVIDIA, OpenAI and Amazon Web Services (AWS) support that work, but they sit at different stages, from reported early results to announced pilots and future targets.
What the evidence does not show is that an AI-assisted molecule has been approved as a medicine, or that AI has measurably shortened the path from discovery to an approved drug. The rest of this article separates what Novo Nordisk has reported, what its partners have announced, and what is still planned.
How Novo Nordisk describes AI in its research workflow
In its 2024 Annual Report, Novo Nordisk describes AI as one part of a research workflow that still depends on laboratory experiments. The company’s framing is that computational screening and predictive design narrow down which molecules are worth testing, and that data analysis then informs decisions about patients and trials. The report does not present AI as a replacement for laboratory science, and the examples it gives are discovery-stage work rather than finished products.
Screening virtual molecules before the laboratory
The clearest description in the report is of combining computer modelling with high-throughput lab work. In the company’s words: “By combining AI with high-throughput experimentation, we have assessed one billion virtual molecules via computer modelling and screened approximately 2,500 compounds in the lab” (Novo Nordisk, Annual Report 2024, AI research and development section). The sequence this implies is a wide computational net first, then a small set of compounds taken to physical testing. The two numbers describe different stages and should not be read as a ratio of success.
#1 Best Overall
Designing molecules in fewer rounds
The report links this workflow to a specific discovery result: a highly selective amylin compound that closely mimics the natural hormone. Novo Nordisk says the work required 50–75% fewer design rounds to reach it (Annual Report 2024). A design round is one cycle of proposing, making and testing candidate molecules, so fewer rounds means fewer cycles of that loop. It is a measure of discovery effort, not of time to market, and the report does not state how many rounds the conventional approach would have needed.
Using clinical-trial data for patient selection and trial design
On the clinical side, Novo Nordisk says it harmonised data from around 1,600 clinical trials, including its SELECT and STEP programmes, to support three uses: disease insights, patient stratification (grouping patients by how likely they are to respond or to develop a condition), and drug-target identification. The report also describes data analysis informing trial design, site selection and outcome forecasting. These are described as uses of the approach; the report does not give measured results for them.
Novo Nordisk’s Karin Conde-Knape, Senior Vice President of Global Drug Discovery, framed the direction in a Microsoft customer story published on 4 October 2024: “Using data with AI to define patients’ risk of developing diseases will help us provide better treatments. This is precision medicine, and it’s going to be important for us to operate in the future.”
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
The figures and what each one measures
Most of the quantitative claims about Novo Nordisk’s AI work come from one document, and each one measures something narrower than it may first appear. The table below gives each figure with the context needed to read it correctly.
| Figure | What it measures | Source | Status as stated |
|---|---|---|---|
| One billion virtual molecules assessed | Molecules evaluated by computer modelling | Novo Nordisk, Annual Report 2024 | Company-reported research figure |
| Approximately 2,500 compounds screened | Compounds tested in the laboratory with high-throughput experimentation | Novo Nordisk, Annual Report 2024 | Company-reported research figure |
| 50–75% fewer design rounds | Design cycles for one highly selective amylin compound | Novo Nordisk, Annual Report 2024 | Discovery-stage result; not evidence of regulatory approval or patient benefit |
| Around 1,600 clinical trials harmonised | Trial data (including SELECT and STEP) used for disease insight, patient stratification and target identification | Novo Nordisk, Annual Report 2024 | Company-reported; no outcome measure stated |
| Up to 20 cardiometabolic programmes | Programmes the expanded Valo collaboration aims to accelerate | Annual Report 2024; collaboration announcement of 8 January 2025 | Development ambition, not a count of results or medicines |
Two points follow directly from the table. None of these figures has been independently audited, so each should be read as Novo Nordisk’s account of its own work. And none of them measures how long a medicine takes to reach patients, which is the question most readers actually care about.
Valo Health: human data and cardiometabolic targets
The Valo Health collaboration is the most concrete example of AI applied to human data in the company’s pipeline. Novo Nordisk announced the expansion on 8 January 2025. It covers obesity, type 2 diabetes and cardiovascular disease, and draws on Valo’s human dataset together with AI-powered computation.
The announcement reports that several novel targets have been identified and that multiple small-molecule preclinical programmes are underway. “Preclinical” means work before human trials begin, so these programmes are early. The annual report describes the goal as accelerating up to 20 cardiometabolic programmes. That is a planning figure. It tells readers how big the effort is meant to be, not how many candidates have succeeded.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsThe technology partners and what each one contributes
The other partnerships supply different parts of the infrastructure, and they should not be compared as interchangeable products. Microsoft and AWS provide cloud and AI platforms. NVIDIA supplies computing and model-building tools. OpenAI is described in terms of general AI use across the business, not drug discovery alone.
Microsoft: cloud, data science and cardiovascular risk models
Novo Nordisk and Microsoft announced a collaboration on 12 September 2022 covering cloud, AI and data science. The stated scope included analysis of scientific literature and cardiovascular risk modelling. Microsoft’s own customer story from 4 October 2024 describes an Azure-based AI platform and says initial cardiovascular risk models were published. Because the source is Microsoft, these are Microsoft’s descriptions of a vendor-supported project, not an independent evaluation of the models.
NVIDIA: supercomputing and model tooling
NVIDIA’s newsroom announcement of 11 June 2025 names the Gefion supercomputer and several software tools: BioNeMo, NIM, NeMo and Omniverse. It outlines proposed research in predicting cellular responses, designing molecules and building biomedical language models. The word that matters is “proposed”: the announcement describes planned use cases, not reported outputs.
OpenAI: pilots across the business
Novo Nordisk announced a strategic partnership with OpenAI on 14 April 2026. The company describes pilots across research and development, manufacturing and commercial operations, and says governance and human oversight are requirements for that work. It also targets full integration by the end of 2026. Announcing the partnership, Mike Doustdar, President and CEO of Novo Nordisk, said: “Integrating AI in our everyday work gives us the ability to analyse datasets at a scale that was previously impossible, identify patterns we could not see, and test hypotheses faster than ever.”
AWS: a London hub and cloud services
Novo Nordisk announced a strategic partnership with AWS on 10 August 2026. The announcement describes a London co-innovation hub, cloud and AI services, and drug-discovery use cases. It is the newest of the partnerships, and the announcement describes use cases rather than results.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Reported results, announcements and plans
The partnerships differ most in how far along they are. The table below sorts each one by its stated status.
| Partner | Date of announcement | Role in the workflow | Status as stated |
|---|---|---|---|
| Valo Health | 8 January 2025 (expansion) | Human dataset and AI computation for obesity, type 2 diabetes and cardiovascular disease | Several novel targets identified; multiple small-molecule preclinical programmes underway |
| Microsoft | 12 September 2022; customer story 4 October 2024 | Cloud, AI, data science, literature analysis, cardiovascular risk modelling | Initial cardiovascular risk models published (per Microsoft) |
| NVIDIA | 11 June 2025 | Gefion supercomputing; BioNeMo, NIM, NeMo, Omniverse; cellular-response prediction, molecule design, biomedical language models | Proposed research use cases |
| OpenAI | 14 April 2026 | Pilots across R&D, manufacturing and commercial operations | Pilots under way; full integration targeted by end of 2026 |
| AWS | 10 August 2026 | London co-innovation hub; cloud and AI services; drug-discovery use cases | Newly announced partnership; use cases described, no results stated |
Read this way, only the Valo programmes and the Microsoft models are described as work already completed or underway. The NVIDIA and AWS items are plans or newly launched collaborations, and the OpenAI target is a commitment with a date that has not yet arrived.
What the evidence does not establish
- Approval. None of the sources links an AI-assisted molecule to a regulatory approval. The amylin compound is described as a discovery-stage result.
- Speed to market. Fewer design rounds is not the same as a shorter development timeline, and no published figure measures how much AI has shortened the total path from discovery to an approved drug.
- Patient outcomes. The trial-data work is described as supporting insight and selection. The sources do not report improved outcomes for patients.
- Independent validation. The key metrics come from Novo Nordisk, and the partner claims come from the companies involved. Peer-reviewed or regulator-reviewed evaluation of these methods is not cited in the sources.
What to watch next
Several developments would show whether these early results carry through. Readers can track the following:
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →- Whether the amylin compound or any other candidate from the screening workflow enters human trials, and under what name or identifier the company reports it.
- Whether any Valo-derived target moves from preclinical work into clinical development.
- Whether Novo Nordisk reports the OpenAI integration as complete by the end of 2026, and what it says about governance and human oversight.
- Whether the company publishes measured results from the AWS hub or the NVIDIA-supported projects, which the current announcements do not yet contain.
Novo Nordisk’s R&D and commercial partnering page describes interest in organisations with complementary expertise in novel targets, modalities, research technologies, clinical development and digital health. That is a category-level statement of strategic interest, not an open invitation tied to this work.
The Bottom Line
Novo Nordisk is using AI as a working part of its discovery and trial analysis, and it has reported specific early figures and a discovery-stage compound. Whether the approach shortens development or produces approved medicines is not yet shown by the available evidence.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
