Nvidia commissioned a DGX GB300 supercomputer at the Naval Postgraduate School in Monterey, California on July 22, 2026, putting one of the company's most powerful AI platforms directly on the campus of the U.S. military's flagship technical graduate university. Chief executive Jensen Huang switched on the system in person, giving 1,500 in-resident students and 600 faculty on-premises access to model training and inference for defense-relevant workloads. The system runs on Nvidia Mission Control software and is anchored inside an Nvidia AI Technology Center already established on the Monterey campus.
The initial workload list spans weather prediction, cybersecurity, and disaster resilience and response planning — all domains where the Naval Postgraduate School already has faculty depth and where sovereign, on-premises compute matters more than cloud economics. The commissioning took place during the school's three-day Converge @ NPS event, and it extends an existing Nvidia collaboration from curriculum and tooling into hardware.
“Our nation depends on our men and women who fight on the front lines. Nothing is more valuable to you than information and insight, and information and insight in a timely way, and to understand its impact and consequence.”— Jensen Huang, Nvidia founder and CEO
The strategic logic is straightforward. NPS educates active-duty officers and international partners across space operations, ocean science, and other applied disciplines, and until now its ability to train foundation models or run high-fidelity simulations at scale was gated by external compute. The DGX GB300 collapses that constraint into a single on-campus system.
Key facts
- 01Nvidia commissioned a DGX GB300 system at the Naval Postgraduate School in Monterey on July 22, 2026.
- 02The system serves 1,500 in-resident students and 600 faculty with on-premises AI training and inference.
- 03Workloads include weather prediction, cybersecurity, disaster resilience, and Omniverse-based digital twins built with MITRE.
- 04Partners DDN, VAST Data, and Vertiv supplied storage, data platform, and power and cooling infrastructure.
- 05The commissioning took place during the three-day Converge @ NPS event.
Huang used the commissioning to argue that access to modern compute is now a leadership prerequisite, not a technical specialty. He told attendees the technology is a tool, not the mission, and pushed officers who aren't computer scientists to engage with AI systems directly. That framing is consistent with Nvidia's broader institutional pitch: put the hardware where the decision-makers train, and the software adoption follows.
Admiral Samuel Paparo, commander of U.S. Pacific Command, spoke alongside Huang and framed the deployment as part of a wider modernization of how the military educates its officer corps rather than a discrete procurement win.
The academic use cases already lined up include ocean modeling, atmospheric prediction, and digital twins of complex operational environments. Through a partnership with MITRE, NPS has built a framework on Nvidia Omniverse libraries that produces high-fidelity digital twin environments for simulating navigation and decision-making under uncertainty. Those workloads had been running at limited scale and will now sit on the DGX GB300 directly.
Nvidia also expanded its Deep Learning Institute footprint at the school, distributing instructor toolkits to NPS faculty so AI coursework threads across departments rather than sitting inside a single computer science track. The pattern mirrors what Nvidia has done at civilian research universities, adapted for a student body of commissioned officers.
The build itself pulled in three infrastructure partners. DDN supplied high-performance data infrastructure for AI workloads, VAST Data provided a unified data platform spanning edge, core, and cloud, and Vertiv delivered the racks, cooling, power, and commissioning support. The partner list is a useful signal of what a defense-grade GB300 deployment now looks like end-to-end.
“Many of you will command in AI-enabled environments. Information will move at lightning speeds, compressing response times.”— Samuel Paparo, Commander, U.S. Pacific Command
There are open questions the commissioning event did not address. Nvidia did not disclose the size of the DGX GB300 deployment in GPUs or the cost, and the school did not detail how classified or export-controlled workloads will be partitioned from open academic research on the same hardware. Those are governance details that typically get worked out after the ribbon-cutting rather than during it.
The broader signal for Nvidia is that its enterprise-grade AI systems are being placed inside the institutions that train U.S. and allied military leaders, not just inside the operational commands that will eventually deploy AI in the field. That is a slower but stickier form of platform entrenchment than a single procurement contract, and it lines up with the ecosystem Nvidia continues to build at events like the upcoming GTC Berlin on October 20-22. If the officers who learn to build with Omniverse and DGX at NPS end up specifying AI systems a decade from now, the standard they specify will already be Nvidia's.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
