TACC Triples AI Inference System With SambaNova
The National Science Foundation awarded $5 million for the expansion, including $3.5 million for AI accelerator hardware.
Updated on Nov. 16, 2026 in Artificial Intelligence

The Texas Advanced Computing Center is tripling the size of its SambaStack AI inference system with SambaNova. The expansion adds SambaRacks to Sagebrush, the system TACC uses to provide AI inference services to researchers.
Why it matters
TACC's supercomputers are busy running scientific simulations and other demanding jobs, so moving AI inference onto dedicated hardware frees that computing power for other work. The expansion gives Texas researchers more capacity to run trained AI models without competing for supercomputer time.
The National Science Foundation awarded $5 million for the project, with $3.5 million going to AI accelerator hardware, and SambaNova claims its SN50 chip runs five times faster than rival chips. Independent verification of that speed claim and the deployment timeline have not been reported.
The players
Texas Advanced Computing Center
TACC is a research computing center at the University of Texas at Austin that operates some of the most powerful supercomputers available to academic researchers.
SambaNova
SambaNova is an artificial intelligence hardware and software company that designs its own AI chips and the SambaStack platform used to run AI models.
National Science Foundation
The National Science Foundation is a federal agency that funds basic research and education across science and engineering in the United States.
The details
A model is trained on a TACC supercomputer, and SambaNova hardware then handles inference tasks afterward. TACC first chose SambaNova technology in 2024, and it is keeping its existing supercomputing systems in place while adding the new hardware.
Timeline
2024: TACC first chose SambaNova technology to run models trained on traditional supercomputers.
Earlier in 2026: SambaNova introduced its SN50 AI chip.
November 16, 2026: TACC announced the expansion with SambaNova.
Big Picture
TACC is involved in the National Artificial Intelligence Research Resource Super Intelligence Infrastructure for Science initiative, and the San Diego Supercomputer Center leads the NSF-funded NAIRR Operations Center. The Texas expansion places a major AI inference resource within that federal effort to give researchers shared access to AI computing.
Researchers at Texas universities and institutions that use TACC resources can run trained AI models through the expanded inference system instead of waiting on time for the main supercomputers. The added capacity should mean faster access to AI tools for academic projects, though the extraction does not specify when that access will begin.
The takeaway
Dedicated AI inference hardware lets research supercomputers focus on core simulations while AI workloads run separately. Researchers planning AI-driven projects in Texas may find more shared computing capacity becoming available through federal programs.
Further reading
For related coverage, explore more Artificial Intelligence stories on Wisevoter.
Source note: This article includes information reported by HPCwire.









