The Pentagon is pushing ahead with plans to build artificial intelligence data centers on military bases, leasing federal land to private companies to house the computing infrastructure behind the U.S. military's accelerating AI ambitions, a move that is now drawing sharp scrutiny from Congress.
Lawmakers in both chambers are advancing measures that would force the Defense Department to slow down and be more transparent about where and how it builds data centers, according to a report by Politico. The push reflects growing unease on Capitol Hill that the military is moving faster than oversight can keep up, leasing sensitive federal installations to private firms with little public accountability. For more context on this story, see our ongoing AI industry coverage.
Leasing the Base for the Server
The Department of Defense is already leasing land to private companies to construct data centers proposed for military bases, Politico reported, citing the emerging congressional concern. The arrangement allows the Pentagon to rapidly expand the computing capacity it needs for AI, from battle management and threat detection to logistics, without building and operating the facilities itself.
That speed comes at a cost. Critics and lawmakers worry that leasing base land to commercial operators raises questions about who controls the infrastructure, how the electricity and water consumed by the facilities are accounted for, and what happens to the communities and military operations surrounding them. Data centers are notoriously power-hungry, and siting them on military installations adds a layer of national-security sensitivity that existing commercial data-center rules were never designed to handle.
Congress Pushes for Transparency
Most of the congressional measures under consideration would not outright block the Pentagon's build-out. Instead, according to Politico, they would require studies, evaluations, and additional planning that would force the Defense Department to take into account, and be transparent about, proposed data centers before breaking ground.
The provisions are working their way through the legislative process as part of broader defense policy debates, including around the National Defense Authorization Act, which has become a recurring vehicle for lawmakers seeking to impose guardrails on military AI. The Senate Armed Services Committee has already advanced its own military AI guardrails for the fiscal 2027 defense bill, reflecting a bipartisan appetite for tighter oversight even among lawmakers broadly supportive of the technology.
Behind the Push: A $30 Billion "AI Arsenal"
The congressional pushback comes as the Pentagon pursues one of the most ambitious AI infrastructure build-outs in the federal government. According to budget documents reported by DefenseScoop, the Defense Department requested close to $30 billion, specifically about $29.5 billion, for fiscal 2027 to purchase and enable next-generation AI supercomputers and modernize the military's computing infrastructure.
Much of that spending is organized under what the Pentagon calls its AI Arsenal initiative, an effort to build a portfolio of highly secure data centers and centralize supercomputing assets across the joint force. An official cited by DefenseScoop tied the effort to the White House's AI Action Plan, which emphasizes the importance of robust AI infrastructure. The scale of the request underscores how central AI computing has become to the military's modernization plans, and why Congress is reluctant to let it proceed unchecked.
A Familiar Tension
The clash over military data centers mirrors a broader national debate over the rapid spread of AI infrastructure. Across the country, local communities, utilities, and environmental groups are pushing back against hyperscale data centers, citing soaring electricity demand, water consumption, and the strain on local power grids. When those same facilities are sited on military bases, the concerns are amplified by questions of national security and democratic oversight.
Private industry is moving in parallel. Earlier reporting indicated that SpaceX has been in talks to supply the Pentagon with AI computing power worth billions, a sign that the Defense Department is pursuing multiple avenues, from leased base land to commercial partnerships, to feed its appetite for compute. Critics argue that the breadth of these arrangements makes them harder to monitor.
The Stakes for Military AI
For the Pentagon, the calculus is straightforward: AI is now central to warfighting, and the department argues it cannot afford to wait. U.S. defense planners have repeatedly framed artificial intelligence as essential to keeping pace with adversaries, particularly China, in a fast-moving technological competition. That urgency has translated into record spending requests and a willingness to experiment with new contracting models.
For Congress, the risk is that a sprawling, opaque network of military data centers, some operated by private companies on leased federal land, gets built before anyone has fully assessed the consequences. Lawmakers' concerns span cost, environmental impact, cybersecurity, and the basic question of whether private operators on military bases should be held to the same standards as the government.
How that tension is resolved will help shape not just the future of military AI, but the precedent for how the U.S. government manages the infrastructure boom now reshaping the entire economy. Lawmakers' demand is, at heart, a modest one: that before the servers go live on the nation's bases, the Pentagon should have to explain exactly what it is building, and why.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
