Nvidia donates GB300 Grace Blackwell supercomputer to U.S. Navy school nonprofit

Vertiv said it deployed the power, liquid cooling, racks and installation services for the Naval Postgraduate School’s NVIDIA DGX GB300 system, outlining a three-rack AI infrastructure build for the Monterey campus project.

Summary

Nvidia’s donated GB300 Grace Blackwell system at the Naval Postgraduate School has now been paired with a three-rack physical infrastructure deployment from Vertiv, which said it provided integrated power, liquid cooling, rack systems, engineering and installation services to bring the high-density AI environment online. The system was provided to NPS by Nvidia and supports education, research, digital engineering, modeling and simulation at the U.S. Navy-run graduate institution in Monterey, California. Vertiv said the project shows how an existing facility can be upgraded to handle the electrical, thermal and operational demands of next-generation accelerated computing without waiting for a new purpose-built site. The configuration was built with NVIDIA, NPS, the NPS Foundation and other partners using Vertiv’s SmartIT integrated AI infrastructure, including power protection and distribution, AI racks, liquid-cooling technologies, fluid management and monitoring, facility interfaces, installation sequencing, and testing and commissioning support. Reuters previously reported the donation marked the first direct use of Nvidia’s most advanced servers by the U.S. military, even as Blackwell chips are also planned for U.S. Department of Energy supercomputers. Nvidia CEO Jensen Huang has said “AI will be a backbone of America’s defense,” while Adm. Samuel J. Paparo said the system would help students and faculty understand both the opportunities and responsibilities tied to the technology.

Terms & Concepts
  • liquid cooling: A cooling method that removes heat from high-performance computing equipment using circulating liquid rather than air alone.
  • accelerated computing: Computing systems that use specialized processors and tightly integrated infrastructure to speed up demanding workloads such as AI.
  • commissioning: The testing and validation process used to confirm that installed infrastructure is operating as designed before full use.