Gefion: Supercomputer Elevates Denmark's AI Hub
A new wave of AI computers is coming. They have processors designed specifically for AI programs. This is a huge change in computation as AI offers new possibilities for individual users, governments, and scientific organizations.
AI computers are able to process massive datasets and learn new skills. No wonder countries compete to lead the AI computing sector to expand their capabilities in private and public sectors. One of the most essential sectors that can benefit from AI is healthcare. AI in healthcare is set to optimize diagnostics, treatment, and research.
Danish Novo Nordisk Partners With NVIDIA for AI Innovation Center
AI use in healthcare has taken off with significant developments. Recently, Novo Nordisk, a Denmark-based leading healthcare company, announced it’s building an AI Innovation Center in the country. For this purpose, Novo Nordisk partners with the Export and Investment Fund of Denmark (EIFO) and tech giant NVIDIA.
The heart of Denmark’s AI Innovation Center will be the Gefion supercomputer. Gefion is set to empower research and development in healthcare, life science, and quantum computing. To bring the idea into reality, the Novo Nordisk Foundation is initially investing €80 million, and the EIFO around €8 million.
The Gefion computer is expected to be fully operational in early 2025. It’ll be one of the most powerful computers in the world. The intelligent machine will use NVIDIA’s processor model H100 Tensor Core GPU, designed for purpose-built, large-scale computing tasks. There will be 15,128 GPUs in the Denmark Innovation Center to perform computation tasks like protein structure prediction.
Gefion will connect to the NVIDIA Quantum-2 InfiniBand networking platform to ensure fast communication between data centers. The computer’s installation and delivery will be provided by the fintech company Eviden. Gefion will be hosted by the Digital Realty data center using 100% renewable energy.
According to Charlie Boyle, vice president and general manager of DGX Systems at NVIDIA:
AI is helping researchers find solutions to many of humanity’s biggest challenges in healthcare, sustainability, energy and more. Built by Eviden with NVIDIA DGX H100 AI supercomputing, the Novo Nordisk Foundation Gefion system can advance research and development in Denmark with transformative, full-stack AI infrastructure.
Modern AI Data Center. Source: nvidia.com
The AI Impact on Computation
The growth of AI and its integration into data centers changes the research and development sector. It enables efficient and intelligent use of data. Machine learning algorithms help recognize patterns and manage workloads.
What can’t be achieved through regular computer chips can be achieved through purpose-built AI chips. These chips are the engines of AI computers that can do a variety of tasks: process massive datasets, learn new skills, and enhance cyber security. AI capabilities reach new heights. During the GTC 2024 Conference, NVIDIA announced the world’s most powerful AI chip, Blackwell.
On the flip side, modern AI techniques need enormous resources and computational power. Organizations are investing heavily in AI computing and application development to offer innovative solutions first. It’s interesting to follow the growing use cases and see who will take the lead in the AI competition.