Denmark has launched its first AI supercomputer, Gefion, built on NVIDIA’s DGX SuperPOD architecture. The system is powered by 1,528 NVIDIA H100 Tensor Core GPUs and interconnected using NVIDIA Quantum-2 InfiniBand networking. This marks a significant milestone in the country’s efforts to accelerate research and innovation in quantum computing, drug discovery, and green energy transition.
The Gefion supercomputer resulted from a public-private partnership between the Novo Nordisk Foundation and the Export and Investment Fund of Denmark (EIFO), which funded the creation of the Danish Centre for AI Innovation A/S (DCAI). Key individuals involved in this project include Morten Bødskov, Minister for Industry, Business, and Financial Affairs; Jensen Huang, founder and CEO of NVIDIA; Nadia Carlsten, CEO of DCAI; Mads Krogsgaard Thomsen, CEO of the Novo Nordisk Foundation; and Peder Lundquist, CEO of EIFO.
Gefion is expected to drive innovation in various fields, including life sciences, climate research, and quantum computing. The system will be used by select customers during a pilot phase, with six winning proposals already announced from researchers at institutions such as the University of Copenhagen and Technical University of Denmark (DTU), as well as startups like Go Autonomous and Teton.
Denmark’s First AI Supercomputer: Gefion
Denmark has taken a significant leap in the field of artificial intelligence (AI) with the launch of its first AI supercomputer, named “Gefion”. This powerful machine is designed to accelerate research and development in various fields where AI plays a crucial role. The supercomputer is hosted by Digital Realty, a global data center provider, in one of its AI-ready facilities in Denmark.
A Private-Public Partnership
The establishment of Gefion is the result of a groundbreaking private-public partnership between the Novo Nordisk Foundation and the Export and Investment Fund of Denmark (EIFO). The Novo Nordisk Foundation has committed approximately DKK 600 million towards the initial costs of the center, while EIFO has contributed DKK 100 million and owns a minority stake of 15% in the company. This partnership represents a strategic focus on AI, quantum computing, the green transition, life sciences, and the commercialization of research.
The Power of GPUs
Gefion’s massively parallel processing capabilities are made possible by its use of graphics processing units (GPUs). Over the past decade, GPUs have seen explosive growth following the emergence of deep learning and generative AI models that run much faster and more efficiently on these processors than on general-purpose central processing units (CPUs).
Sustainable Data Center
The data center hosting Gefion is designed and built to be sustainable and runs on 100% renewable energy. This aligns with the Novo Nordisk Foundation’s vision of improving people’s health and the sustainability of society and the planet.
First Use Cases: Flagship Projects
The supercomputer and collaboration with NVIDIA will enable Denmark to pursue large-scale projects in various 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 from NVIDIA to co-develop solutions to complex problems.
Some of the first use cases include:
- Research in pharmaceuticals and biotechnology using the NVIDIA BioNeMo platform, including protein design
- The acceleration of the green transition
- The development of fault-tolerant (i.e., error-free) quantum computing using NVIDIA CUDA-Q, an open-source hybrid quantum computing platform
Pilot Projects
A competition to identify innovative AI use cases that can fully harness the potential of Gefion saw over 50 proposals submitted in just three weeks. Six winning proposals were selected to run as part of Gefion’s pilot phase. These projects come from researchers and startups, including:
- “Large-scale distributed simulation of quantum algorithms for quantifying molecular recognition processes” (University of Copenhagen)
- “Unravelling CO2 reduction in Non-Metal Formate Dehydrogenase (FDH) using Machine-Learned Force Fields” (Technical University of Denmark)
- “Multimodal genomic foundation model” (University of Copenhagen)
- “Multi-Modal Document Understanding: Transforming Data Entry with Multi-Modal Precision” (Go Autonomous)
- “Building an AI Care Companion with Large Video Pretraining” (Teton and the University of Copenhagen)
- “SAPIEN – Skilful Atmospheric Prediction with Intelligent Environmental Networks” (Danish Meteorological Institute)
These projects will utilize Gefion to process very large data sets to find new solutions to problems that are extremely hard to solve without access to GPU-accelerated computing.
External Link: Click Here For More
