With the launch of its artificial intelligence tool CentGPT, U.S. Central Command has initiated a pilot project as it works to establish its own high-performance computing infrastructure to address classified compute-related challenges, Federal News Network reported Wednesday.
Table of Contents
Assessing Compute Load Needs
As part of the pilot, CENTCOM procured and installed four NVIDIA H100 graphic processing units, or GPUs, on its classified network to support CentGPT and other AI initiatives.
The command has also teamed with the Air Force Research Laboratory and Lambda under a cooperative research and development agreement, which will enable CENTCOM to purchase an additional 24 NVIDIA H100 GPUs.
“It will give us a really huge, significant amount of compute capability that no one else — at least that I’m tracking — has in the Defense Department for classified networks,” Dan Leventry, chief data officer at CENTCOM, told FNN.
According to Leventry, the command expects the pilot to help establish baseline costs and serve as a testing ground to determine other combatant commands’ future requirements for computing power as they scale AI adoption.
“We think we’re going to be good for several years, but it allows us to benchmark what the requirements are for other combatant commands as we move into the cloud and the commercial cloud providers begin to provide that compute power in the cloud,” he said.
What Is CentGPT?
The recently introduced CentGPT is an AI-powered chatbot and document analysis platform built on open source models like Meta’s Llama. The tool integrates an AI framework called retrieval-augmented generation that enables users to upload, analyze and interpret documents.
CENTCOM launched CentGPT on its classified and unclassified networks and saw the number of users of the AI tool grow to 900 in six months.
Attend the Potomac Officers Club’s 2025 Digital Transformation Summit on April 24, and hear experts as they discuss how emerging technologies and latest innovations modernize government operations. Register now!
