The Trump administration is trying to solve the AI energy crunch by ripping up the old nuclear playbook. Instead of waiting a decade for a conventional plant to clear regulators, the White House is inviting tech giants to finance their own reactors and plug them directly into data centers, with approvals that could arrive in a matter of weeks. The result is a sweeping experiment in whether the United States can fuel Artificial Intelligence with nuclear power while sidestepping the layers of red tape that have defined the industry for half a century.
At the center of this push is a cluster of executive orders, agency pilots, and regulatory rewrites that tie nuclear energy directly to national security and AI competitiveness. The administration is betting that if it can compress timelines, shift costs onto Big Tech, and weaken state and local veto points, the United States will not only keep pace in AI but also quadruple its nuclear fleet by midcentury. Whether that gamble pays off will depend on how safely and quickly these new reactors move from policy paper to concrete and steel.
The three‑week reactor: Trump’s AI bargain with Big Tech
The most eye catching piece of the plan is the promise that companies like the largest cloud providers could get nuclear projects cleared in roughly three weeks instead of slogging through the traditional Nuclear Regulatory Commission process. Reporting on Nuclear Regulatory Commission reforms describes a proposal that would let tech firms build dedicated reactors for their own data centers, bypassing years of case by case review that the NRC historically required. The idea is to treat AI infrastructure as a national security asset, then tailor nuclear oversight around that priority rather than around the slow, litigation heavy model that grew out of the 1970s.
In parallel, the administration has leaned on the same companies to pay for the hardware. Coverage of the White House’s outreach to industry describes how The Trump team wants large platforms to finance new power plants as AI boosts electricity usage, a push summarized in one report’s Topline that The Trump administration will urge tech giants to underwrite capacity rather than pass higher rates to consumers after calls from Trump. Another account of the Davos pitch notes that Trump has talked up a three week approval window for nuclear projects tied to AI data centers, framing slow permitting as a central economic risk of 2026 and promising to fast track reactors that serve Big Tech’s energy hungry clusters in what one report dubbed Trump fast tracks for Big Tech.
Rewriting a 50‑year‑old regulator
To make any of this real, the White House is trying to overhaul the Nuclear Regulatory Commission itself. In a presidential action titled ordering the reform of the Nuclear Regulatory Commission, Trump directed the agency to align its structure and staffing with Congress’s original mandate to support both safety and deployment. A separate section of that order, labeled Sec 4, explicitly targets Reforming the NRC’s Structure, arguing that the current Structure of the NRC is misaligned with Congress and that the agency should speed the adoption of innovative technology instead of acting as a brake.
Senior aides have described this as a “total and complete reform” of a 50-year-old regulator that has long been cautious about new reactor designs. The same set of directives instructs the NRC to coordinate more closely with the Department of Energy’s national laboratories, effectively pulling research, licensing, and deployment into a single pipeline. Legal analysts have noted that these New Executive Orders Aim to Accelerate and Expand, with The US government using executive authority to compress environmental reviews, standardize approvals, and prioritize projects that serve national security or AI related objectives.
From military bases to data centers: the new nuclear buildout
The administration is not starting from scratch. Earlier orders already tied advanced reactors to defense infrastructure, and those precedents now serve as a template for AI. In a White House document titled Fact Sheet, President Donald, Trump Deploys Advanced Nuclear Reactor Technologies for National Security, officials explained that Trump signed an Executive Order to rapidly deploy advanced nuclear reactors on military installations, treating them as hardened power sources for critical missions. The same Fact Sheet notes that The Order directs agencies to support these projects as part of a broader push to modernize the grid and reduce reliance on vulnerable fuel supply chains, embedding nuclear into the national security toolkit rather than treating it as just another generation option.
Another section of that document, labeled DEPLOYING NUCLEAR REACTORS FOR NATIONAL SECURITY, spells out a timeline in which new systems are expected to meet security objectives within 30 months. That same logic now underpins the AI push: if the Pentagon can host small modular reactors on bases, the argument goes, then hyperscale data centers in Energy Silicon Valley can do the same. A separate Energy Department Fact Sheet underscores that Building out the next generation of American nuclear is framed as an investment in growing energy needs today that will become a backbone of U.S. small modular reactors, a description that neatly fits the power profile of AI campuses.
AI policy meets energy policy
The nuclear sprint is tightly coupled to a broader effort to strip away what the White House sees as obstacles to AI development. In a presidential action titled Ensuring a National Policy Framework for Artificial Intelligence, the administration’s Section 1, labeled Purpose, states that United States leadership in Artificial Intelligence will promote United States national and economic security, particularly for start ups that might otherwise be constrained by a patchwork of state rules. The order is explicit that federal policy should eliminate state law obstruction of national AI policy, clearing the way for a unified approach to both data and infrastructure.
That philosophy carries over into the administration’s AI strategy documents. One analysis of the All American AI blueprint notes that the report released in Jul calls to “reject radical climate dogma and bureaucratic red tape” that it believes will interfere with expansion of AI, a line that appears in coverage of the plan’s environmental stance and is echoed in Jul reporting on how it would reshape agency obligations and authorities. A separate review of the America AI Action Plan explains that the new plan favors removing regulatory barriers to AI development and largely counters previous initiatives, with Scarlett Evans and Con highlighting that the document explicitly celebrates cutting red tape for AI, a stance captured in coverage that quotes the line that the new plan favors removing regulatory barriers to AI.
Emergency auctions, direct hookups, and a 400 g ambition
Even with nuclear in the spotlight, the administration is scrambling to keep today’s grid from buckling under AI demand. In early Jan, President Trump used Truth Social to announce that his administration is working with major U.S. tech companies on an Emergency Energy Auction to prevent data center driven rate increases, a plan described in detail by analysts who note that the auction is designed to steer companies toward firm power as opposed to renewable energy. That same analysis of the Jan proposal underscores that the White House wants AI operators to compete for dedicated supply rather than simply outbidding households on the open market.
Regulators are also changing how data centers connect to the grid. Federal regulators will allow tech companies to effectively plug massive data centers directly into power plants, issuing approvals that let them bypass some traditional transmission constraints in order to support growth in artificial intelligence and revive domestic manufacturing, according to an account that highlights the role of Federal regulators in reshaping grid rules. Inside government, Jan briefings on 4 things to watch at DOE in 2026 note that DOE is also moving to deploy data centers at four federal sites, a plan expected to involve nuclear power and to treat those centers as a national security issue, a framing that appears in coverage of DOE priorities for the year.
More from Morning Overview