> ## Content Index
> Fetch the complete content index at: https://top-margin.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Army Is Building Its Own Power Plants
- URL: https://top-margin.ghost.io/the-army-is-building-its-own-power-plants/
- Published: 2026-09-13T11:40:33.000Z
- Updated: 2026-09-13T11:40:33.000Z
- Author: Marcus Cole

At the end of August the Army named five installations that will get nuclear microreactors and the five companies that will build them, with awards worth up to $2.2 billion over five years. Fort Bragg, Fort Campbell, Fort Hood, Fort Benning and Fort Drum are the sites. The reactors run somewhere between one and twenty megawatts each depending on the design, the Army expects to build more than twenty of them in total, and there is a stated deadline to have an advanced reactor operating on a domestic military installation by the end of September 2028.  
  
That deadline is the interesting part. Nuclear projects in the United States have historically been measured against decades. This one is being run against a date roughly two years out, which tells you the requirement is being treated as urgent rather than exploratory.

### From the Battlefield to the Balance Sheet

The military rationale is grid independence. A base that draws power from a commercial utility inherits that utility's vulnerabilities, and the same reasoning that put backup generators everywhere now extends to something that can carry a full critical load indefinitely without a fuel convoy. That is a real requirement, and it is worth noting that the Army is buying resilience rather than cheap electricity, because at these sizes the power will not be cheap.  
  
The economic significance is what the order does for the supply chain. Small reactor developers have spent a decade with credible designs and no customer willing to be first, which is the classic problem for capital-intensive technology facing a licensing process built for something else. A committed buyer with five sites, a deadline and a budget solves the first-of-a-kind problem that private financing could not. It also produces the thing the sector has never had, which is repetition. The second reactor of a given design costs less than the first, the fifth less than the second, and no developer has previously been allowed to find out where that curve goes.  
  
For allocators the read is that this is a manufacturing learning-curve bet dressed as a defense program, and it should be judged on delivery of the first units rather than on the announcement.

### **The Dual-Use Reality Check**

The civilian demand behind this is not subtle. Data center construction has outrun the ability of utilities to deliver interconnection in several regions, and the queue for large loads now runs years in many markets. A three to twenty megawatt reactor that can be sited next to a load and does not require a transmission upgrade addresses exactly that problem, and the companies on these Army awards are the same ones talking to hyperscale operators.  
  
That makes the defense program a subsidy for the commercial market in the most direct way available. The government is paying for first-of-a-kind engineering, regulatory precedent and fabrication experience on sites where security and land access are simple, and the resulting designs, supply chains and trained crews are then available for private buyers who could never have absorbed the first-unit cost. If these units get built on anything like the stated schedule, the beneficiary is every industrial and computing load currently stuck behind an interconnection queue. If they slip, the lesson will be about licensing and fabrication capacity rather than about the technology.

### **The Capital Signal**

The capital signal is a government moving from research funding to purchase orders in a sector where that transition has never happened at scale. Five sites, five vendors and a fixed date is a different instrument than a demonstration grant, and it converts a set of engineering companies into manufacturers with a delivery obligation.  
  
The structural read is that the constraint will be fabrication and fuel rather than design. Pressure vessels, reactor internals and the specialist welding and forging behind them come from a very small supplier base, and the fuel for several of these designs requires enrichment capacity that is only now being rebuilt domestically. Those chokepoints will decide the schedule, and they sit well upstream of the companies whose names are on the awards. Watch the fuel supply and the heavy forging base, because that is where a 2028 deadline is either met or missed.  
  
*Signal: The Army has become the first-of-a-kind customer the small reactor industry could not find commercially, and the schedule now depends on forgings and fuel.*

![](https://storage.ghost.io/c/e3/0e/e30efc82-f19f-427f-b131-8005b5e7416a/content/images/2026/09/army-is-building-its-own-power-plants-photorealistic-1.jpg)

Marcus Cole, Top Margin