NVIDIA AI Supercomputer Comes Online at Naval Postgraduate School

Jensen Huang flew into Monterey this week not to sell a product, but to flip the switch on one. NVIDIA's CEO personally commissioned a DGX GB300 system at the Naval Postgraduate School, the graduate university that trains active-duty officers and international military partners. It's a notable move: rather than shipping a black box, NVIDIA is embedding one of its most powerful AI platforms directly into a campus that already runs on national-security stakes.

The numbers give a sense of scale. NPS has more than 1,500 in-resident students and 600 faculty, all of whom now get on-premises access to model training and inference power that used to require cloud contracts or defense-budget-sized procurement cycles. NVIDIA's Mission Control software runs alongside the hardware, and the whole setup sits inside an NVIDIA AI Technology Center that's been quietly built out on the Monterey campus. This isn't a one-off donation — it's infrastructure meant to anchor coursework in weather prediction, cybersecurity, and disaster response for years.

Admiral Samuel Paparo, who leads U.S. Pacific Command, framed it in blunt terms: officers commanding in AI-enabled environments need to decide faster, and that means training with the same class of tools they'll eventually rely on in the field. Huang's own pitch to the room was less about hardware specs and more about permission — he told attendees, many of whom aren't computer scientists, that engaging with AI tools is easier than they assume, because the technology is designed to keep getting smarter without demanding expert-level fluency from every user.

Beyond the classroom, the GB300 will power some genuinely applied science. NPS runs regular hackathons that have already produced work in autonomy and ocean research, and the school's partnership with MITRE has produced an NVIDIA Omniverse-based framework for building digital twins of real-world environments — useful for simulating navigation and decision-making when conditions are murky. NVIDIA's Deep Learning Institute is also handing instructor toolkits to NPS faculty, so the goal isn't just raw compute but curriculum that actually uses it.

Getting a system like this running took more than NVIDIA showing up with silicon. DDN and VAST Data handled the data infrastructure side — the unglamorous but essential work of moving and securing data across edge, core and cloud. Vertiv built out the racks, cooling and power, and ran the installation and commissioning. It's a reminder that a supercomputer announcement is really a systems-integration story wearing a press release.