Data centers are coming to the US military, and the race for AI infrastructure enters military bases.
A short while ago we received a truly dramatic update!
Until now, the data center race revolved around tech giants.
Now the Pentagon is also entering the game, not as a client, but as one providing military land for the establishment of massive AI infrastructure.
The US Department of Defense is advancing a plan to establish AI-based data centers at more than 12 military bases across the United States.
This is a strategic move designed to significantly increase the computing power available to the military, without the government funding the construction of the facilities.
The first projects have already been selected: 1. Fort Bliss, Texas – Carlyle was chosen to establish a campus on approximately 1,384 acres.
2. Dugway Proving Ground, Utah – CyrusOne, owned by KKR and BlackRock funds, will establish a data center on approximately 1,201 acres.
The investment volume for each of the projects is estimated at approximately $2 billion.
The economic model is brilliant.
Instead of the government investing billions in building server farms, it grants private companies the right to use military land.
In return? 1. The companies finance the entire project – planning, construction, equipment, operation, maintenance, and future decommissioning. 2. The Department of Defense gains access to the centers' computing power without upfront government capital investment (but will, of course, pay rental fees for the use of the computing power).
In practice, this is a "land for computing power" deal.
This is further proof that computing power has become a strategic resource, just like military bases, runways, or aircraft carriers.
The US military is no longer content with purchasing cloud services.
It wants the next generation of AI infrastructure to be physically established within military facilities, with direct access to security systems, a high level of security, and dedicated infrastructure for compute-intensive tasks.
The Secretary of Defense even defined artificial intelligence as a "strategic asset" and a "force multiplier" for the armed forces.
The move also strengthens the thesis that we have been seeing in the last two years:
The demand for computing power has long ceased to come only from cloud giants.
Now the defense sector also joins as a significant consumer of AI infrastructure, a move that could increase over the years the demand for data centers, servers, AI chips, cooling systems, electrical equipment, and communication networks.
It seems we have reached a stage where the market no longer asks if more computing power is needed, it asks where more land can be found to build it.