NVIDIA announced Rubin HPC at SC26 on November 16, 2026: a Vera Rubin platform GPU with more than 50,000 CUDA cores and seventh-generation NVLink, built to speed the simulations that science and engineering agents run. It is expected in 2028.
What changed with Rubin HPC
NVIDIA announced Rubin HPC at SC26 on November 16, 2026, as a new GPU in the Vera Rubin platform to accelerate simulation and data processing for science, engineering and analytics.
It has a record number of CUDA cores — more than 50,000 on a single GPU — plus a sparse data acceleration capability and seventh-generation NVLink. The design scales data movement and compute together, and it targets sparse matrix operations, stencil computations, sorting and reductions — the work behind climate models, engineering simulations, genomic analysis and financial risk calculations. With Rubin HPC, CUDA-X libraries give agents optimized implementations of the operations they use most. The GPU is expected to be available in 2028.
Rubin HPC by the numbers
- CUDA cores on a single Rubin HPC GPU: More than 50,000 (Nov 16, 2026)
Why Rubin HPC matters for business
Research-driven businesses in engineering, materials, genomics and finance can plan around a purpose-built GPU for the simulation step that agents depend on — faster simulation means agents can test more candidate designs and molecules before physical experiments — and the 2028 timing makes it a roadmap and procurement-planning item, not capacity to buy this year.
Top comments (0)