Microreactor Economics: Where Remote Power, Heat and Military Projects May Fit

60-second answer: Microreactors are not competing with grid power; they are competing with whatever their customer pays now, and that number varies by an order of magnitude across the three real markets. A remote mine burning trucked-in diesel can pay several times grid rates for electricity; an industrial site wants continuous heat; a military base prices energy security in mission terms. DOE's working definition spans 1–20 MW thermal (electrical output is less, after conversion losses), with long refueling intervals as design goals (DOE microreactor explainer). The 2026 test criticalities prove cores work; none of them establishes a commercial price. So this guide gives the break-even framework with the customer's own numbers as inputs, because as of 2026 there is no disclosed commercial microreactor price to quote.

The microreactor pitch is factory-built nuclear at the scale of a diesel generator hall. Whether the economics work is not one question but three, one per market, and each has a different competitor.

First, define the unit

Microreactor marketing mixes three quantities that differ by large factors:

• Thermal power (MWth): what the core produces. DOE's 1–20 MW range is thermal.
• Electrical power (MWe): thermal times conversion efficiency, often a third or less at small scale. Westinghouse's eVinci illustrates the ratio: about 5 MWe from roughly 13 MWth (Power).
• Delivered energy (MWh over time): power times availability, which is what a customer buys and the only basis on which a diesel comparison is honest.

A "10 MW microreactor" claim means little until the letters after MW are pinned down; the same discipline applies across reactor classes.

The three markets and their real competitors

• Remote power. Mines, Arctic communities and islands run on delivered diesel, whose all-in cost includes fuel, trucking or barging, storage, maintenance and backup. This is the one market where today's incumbent can cost several times grid power, which is what makes it the lead microreactor use case. The bar: beat delivered diesel on total cost over the contract, including the reactor's own logistics.
• Industrial heat. Continuous process heat at useful temperatures, competing with gas boilers and electric heating. High-temperature designs sell heat directly, skipping conversion losses; the competitor's price is local fuel cost, and the sale needs a customer comfortable hosting a reactor.
• Defense and government. Project Pele, the DoD's 1.5 MWth transportable demonstration, took delivery of its BWXT-built TRISO core in late 2025 for testing at INL from 2026 (BWXT). Government programs buy assured energy for missions, at program budgets that do not translate into commercial unit prices; a military demonstration is procurement evidence, not a price point.

A break-even worksheet, with the customer's numbers

The honest comparison for the lead market runs on inputs the operator already knows. For a remote site: annual electrical load (MWh), the delivered diesel price at the site ($/liter, including logistics), generator efficiency (roughly 3–4 kWh per liter), plus diesel plant maintenance and backup requirements. Multiply out and a site drawing 30 GWh a year at a delivered price of $1.50/liter spends on the order of $12–15 million a year on fuel alone before maintenance. That annual figure, over a contract term, is the budget a microreactor bid has to beat, covering its capital, fuel cores, staffing, security, decommissioning and its own backup arrangements. Every input above is the customer's own or explicitly assumed; the one number nobody can yet supply from evidence is the microreactor's side of the ledger, because no commercial unit has been delivered at a disclosed price. When bids become public, this worksheet is how to read them.

Fuel-side inputs for the nuclear column exist already: most announced microreactors specify HALEU, often as TRISO, with whole-core replacement on multi-year cycles, so fuel arrives as a periodic capital item rather than a continuous cost, and financing dynamics resemble small construction projects more than fuel contracts.

Factory manufacture versus the scale penalty

Microreactors push the SMR trade to its limit. Per kilowatt, small is expensive: containment, licensing, security and staffing shrink slower than output, and a 5 MWe unit cannot amortize them the way a gigawatt plant does. The offsetting bets are true factory production (the whole reactor ships in containers, Pele's system fits in four), site works measured in weeks, and fleets sharing licensing and operations overhead. Those bets are testable claims about manufacturing learning, and none has yet been tested at commercial volume. What can be said from evidence: transportability is demonstrated at demonstration scale, and refueling intervals of eight-plus years are design targets some vendors state (eVinci), not fleet operating history.

What the 2026 demonstrations prove

Five reactors reached criticality under DOE's pilot program between June and August 2026, several of them microreactor-class, and the NCERC test series added zero-power validation for eVinci's core (the licensing guide maps what those authorizations cover). The demonstrations establish that the cores go critical, the teams can execute, and the federal test pathway is fast. They do not establish commercial cost, NRC commercial licensing, fuel-supply economics at fleet scale, or a customer's willingness to host. The gap between a criticality and a commercial contract is where all three markets' economics will be decided, and the evidence to watch is contracts with disclosed terms, tracked as they appear on our reactor tracker and in company filings.

What to watch next

Commercial evidence, in rising order of strength: a customer contract with a disclosed price per MWh or per year; an NRC commercial license for a specific site; a delivered unit operating for a paying customer; and a second unit for the same customer, the first real datapoint on factory economics. On the uranium side, microreactor fleets would add demand in a distinctive shape (periodic whole cores of HALEU rather than annual reloads), but at single-digit MWe per unit the pounds stay small until fleet counts are large; the per-reactor math scales down accordingly.

Frequently asked questions

How much does a microreactor cost? As of 2026, no commercial microreactor has been delivered at a publicly disclosed price; every circulating figure is a target, a program budget or an estimate. The economics are therefore evaluated as a break-even against the customer's current energy cost, which for remote diesel sites can run several times grid rates.

What are microreactors used for? The three markets with real commercial logic: remote sites now running on delivered diesel (mines, isolated communities), continuous industrial heat, and defense installations buying assured energy. Each has a different incumbent to beat and a different willingness to pay.

Are microreactors the same as SMRs? Microreactors are the smallest end of the small-reactor spectrum, roughly 1–20 MW thermal per DOE's working definition, against tens to hundreds of megawatts for SMRs. The economics differ too: microreactors lean entirely on factory production and transportability to offset the per-kilowatt penalty of being small.

Did microreactors already reach criticality? Several test units did in 2026 under DOE authorizations, and the DoD's Project Pele core was delivered for testing. Those milestones prove working cores and a fast federal test path; commercial licensing, pricing and fleet operation remain undemonstrated.

How much uranium would microreactors use? Per unit, little: single-digit megawatts electric, fueled by periodic whole-core HALEU replacements on multi-year cycles rather than annual reloads. Meaningful uranium demand from this class requires fleet-scale deployment, which is exactly what current evidence does not yet show.

This article is for informational purposes only, not investment advice.