

Nvidia’s Chips Power the Supercomputer That Could Change AI
12 snips Jan 24, 2025
Mads Krogsgaard Thomsen, CEO of the Novo Nordisk Foundation, leads a vital discussion on Gefion, an AI supercomputer powered by Nvidia. He shares insights on its groundbreaking role in revolutionizing drug discovery and protein design. The conversation touches on how machine learning enhances research efficiency and accelerates therapy timelines. Mads also explores the supercomputer's diverse applications across biological research and agriculture while emphasizing the importance of equitable access to AI technologies for societal advancement.
AI Snips
Chapters
Transcript
Episode notes
AI Supercomputer vs. Supercomputer
- Gefion's AI optimization makes it special, using NVIDIA's H100 chip.
- It's optimized for AI workloads, especially AI training.
Machine Learning in Drug Discovery
- Mads Krogsgaard Thomsen used machine learning to design a hemophilia drug.
- This process, without a GenAI computer, cut development time by two years.
Orders of Magnitude Improvement
- AI allows parallel processing, significantly speeding up drug discovery.
- Chemists can now test more molecules faster, with AI guiding the process.