Ian Buck Kicks Off NVIDIA’s Week at SC26 as Science Enters the Age of Agents

What happened
At the SC26 conference in Chicago, Ian Buck of NVIDIA introduced new AI agent tools and high-performance computing products. He announced the release of Nemotron for Science, the Rubin HPC GPU, and an expanded Vera CPU platform. Buck also revealed collaborations with IBM and Fermi National Accelerator Laboratory focused on quantum computing. The announcements included new capabilities for researchers to run AI-driven scientific experiments and simulations on their own systems.
Key details
- Nemotron for Science uses the Nemotron 3.5 Super VL open model, an open chemistry dataset, and post-training recipes for life sciences and chemistry.
- The Rubin HPC GPU, part of the Vera Rubin platform, has over 50,000 CUDA cores and is expected in 2028.
- The Vera CPU platform has been selected for the HLRE-5 supercomputer at the German Climate Computing Center, with deployment planned for early 2028.
- NVIDIA, IBM, and Fermilab are launching a quantum computing co-design platform, pairing NVIDIA DGX B200 systems with IBM quantum computers.
- A new CUDA-Q decoding algorithm, TCBP-JOBS, reduces quantum error-correction research computation from 12 weeks to 45 minutes.
Why it matters
Researchers in science, engineering, and life sciences can now use new AI agent tools and hardware to automate and accelerate experiments and simulations on their own infrastructure. High-performance computing centers, such as those in Germany and Texas, will gain access to new CPUs and GPUs designed for demanding workloads, with deployments planned over the next several years. The collaboration with IBM and Fermilab provides quantum computing researchers with new platforms and faster error-correction capabilities, reducing research times and enabling more complex quantum experiments.
