Together with the Export and Investment Fund of Denmark (EIFO) and world-leading AI computing company NVIDIA, the Novo Nordisk Foundation is establishing the Danish Centre for AI Innovation.
Once operational, researchers from Denmark’s public and private sectors will be able to gain access to a state-of-the-art NVIDIA AI supercomputer optimised for large-scale projects using AI, as well as world-leading NVIDIA software platforms, training, and expertise. The goal is to accelerate research and innovation in fields from healthcare and life sciences to the green transition, supporting the development of innovative solutions to the world’s biggest problems.
Denmark’s new supercomputer – a large-scale NVIDIA DGX SuperPOD – is powered by NVIDIA H100 Tensor Core GPUs and interconnected using NVIDIA Quantum-2 InfiniBand networking.
The supercomputer and collaboration with NVIDIA will help enable Denmark to pursue large-scale projects in countless fields where AI is a valuable tool, as well as within AI research itself. In selected flagship project areas, researchers will be able to engage with expert teams at NVIDIA to co-develop solutions to complex problems.
These include research in pharmaceuticals and biotechnology, including protein design; the acceleration of the green transition; and the development of fault-tolerant (i.e. error-free) quantum computing.
Eviden, an Atos Group company and European leader in advanced computing, will deliver, install, and configure the supercomputer, named Gefion, and provide support with the start-up phase. The centre is expected to be ready for pilot projects before the end of 2024.
The Foundation has established a public limited company – the Danish Centre for AI Innovation A/S – that will own and operate the supercomputer, with all revenue reinvested in the initiative. EIFO is a minority owner, holding 15% of the shares.
Global data centre provider Digital Realty will host the supercomputer in one of its AI-ready facilities in Denmark. The data centre is designed and built to be sustainable and runs on 100% renewable energy.
The supercomputer will have the highest level of security and support Danish data sovereignty. Once fully operational, it will therefore be able to support projects involving sensitive data. No data will be stored permanently in the centre, and the centre’s users will be in full control of their data at all times.
Differential payment models are being developed that will balance commercial and academic use with the aim of meeting the centre’s current and future costs.
Frequently asked questions
When will the centre be up and running?
A pilot phase will start in October 2024, during which DCAI will set up the necessary software and processes. Selected projects that are sufficiently mature will be able to start during this phase, with users providing valuable feedback to DCAI.
The goal is then to go live for most project types in the first half of 2025. A solution for projects involving sensitive data will come later – we will share more information as we have it.
Can I work at DCAI?
We will be hiring later in the year and advertising these positions on the Foundation’s LinkedIn.
Can you say more about the flagship project areas?
Several broad areas have been developed and agreed upon by the Novo Nordisk Foundation together with EIFO and NVIDIA. They are still evolving, but are currently as follows:
– Foundational AI research
– Development of democratic and human centric AI healthcare solutions
– Development of Danish large language models, especially within health and biomedicine
– Research in pharmaceuticals and biotechnology, including protein design
– Acceleration of the green transition, including research in energy materials
– Data collection and AI for Earth-2 like projects, using multi-modal sources and remote sensing for understanding and predicting the climate, weather, forestry, agriculture, oceans etc.
– Acceleration of the development of fault tolerant quantum computing, and simulation of quantum states
In projects within these areas, researchers will be able to engage with expert teams at NVIDIA to co-develop solutions to complex problems and push technological boundaries.
What if I want to use Gefion for a project outside these areas?
Researchers and innovators from any field are welcome to use the supercomputer, provided they fulfil the access criteria regarding ethical use and agree to the terms of use (both will be laid out in the Access Charter).
NVIDIA will be able to provide general support to all projects, primarily through upgrading skill levels and via train-the-trainer engagements. However, resource allocation by NVIDIA’s teams for co-developing new solutions will be reserved initially for the flagship project areas.
Can the supercomputer be used to develop large language models (LLMs) that meet Danish needs?
Yes, this is one possible use case. The development of Danish large language models, especially within health and biomedicine, has been identified as a possible point of collaboration between Danish researchers and NVIDIA experts. Gefion’s architecture is optimised for training and deployment of large AI model computational workloads, and training LLMs is a good example of this.
Who will prioritise what large language models (LLMs) Gefion will be used for?
As long as sufficient capacity can be allocated, users are welcome to utilise Gefion to train or finetune LLMs. For such projects, as with many others, DCAI and/or NVIDIA would be able to provide technical expertise to help accelerate the training process. For the development of health-related or biomedical LLMs – one of the flagship project areas – researchers will be able to engage with NVIDIA’s expert teams at NVIDIA to co-develop solutions.
How can I get access?
The details are being worked out. We are aiming for a process that will enable researchers and innovators in the private and public sectors in Denmark to access the facility’s services relatively quickly and smoothly, while ensuring that all users fulfil some basic criteria, including for ethical use.
I may want to use the facility in 2025. How can I prepare?
Gefion is a powerful compute platform designed for large, at-scale AI and computational workflows using GPUs. Familiarity with the NVIDIA software stack and with scaling up your computational workload on GPUs will therefore be a good place to start.
There will also be some training events later in 2024. We will share more details when we have them.
How much will it cost to access?
The aim is to ensure that all researchers, innovators and companies in Denmark who wish to use the facility will be able to do so, while ensuring that DCAI can meet the centre’s operational costs and the cost of future upgrades.
There will therefore be a differential fee model, with higher costs for some types of users than others. We will share more information as it becomes available.
What if the data is sensitive?
The environment will be designed to – eventually – be able to facilitate a secure, multi-tenancy setup, where each user will have their data and computational tasks isolated and secured from others.
We aim to have this solution as soon as possible, but it’s a complex process to create a platform that provides this level of security. We will share more information on an expected timeline when we have it.
How can I find out more?
This page will be updated as new information becomes available. We will also share major announcements across the Novo Nordisk Foundation channels. Later in 2024, there will be an official website for the company and the supercomputer.
