Michigan study sees an industrial opening for small reactors, contingent on tax credits

A University of Michigan assessment says small modular reactors could find an early US market supplying hydrogen to heavy industry, while struggling to earn a return from selling electricity alone. The university's September 22 account puts ammonia production and oil refining among the most promising uses. Its large deployment and emissions figures are projections from a cost model, not results from operating US commercial reactors.

The study, published in Nature Communications on August 4, examined potential reactor projects against the energy needs of 34 ammonia plants, nine steel plants, 47 refineries and 955 facilities that use industrial process heat. The researchers also modeled reactor operations in wholesale electricity markets. That comparison matters because a reactor serving an industrial site could sell hydrogen or heat as well as produce electricity; a power-only project must compete with the prices available in the grid market.

The tax credit changes the result

Under the study's first-of-a-kind cost assumptions, the researchers estimated that 91 gigawatts electric of small modular reactors could be deployed profitably when projects receive the federal clean-hydrogen production credit. Without that credit, the modeled figure fell to about four gigawatts. Neither number describes reactors built or committed for construction. They indicate how strongly the modeled opportunity depends on federal support.

The US Department of Energy says the Section 45V credit can pay eligible projects up to $3 for each kilogram of clean hydrogen produced for 10 years. The amount depends on emissions and labor requirements. The Michigan researchers compared cases with and without that production credit, while retaining a separate modeled 30% reduction in reactor and electrolyzer capital costs associated with an investment tax credit. Their first-of-a-kind reactor cost estimates came from manufacturers.

The university's account reports projected profit margins of about 300% for hydrogen production serving ammonia plants, more than 50% at most refining sites and about 20% at steel sites. Only a handful of facilities needing process heat had positive margins, generally below 10%. Those are modeled margins for selected applications, not measured returns across existing projects. They also help explain why the researchers identify industrial hydrogen, rather than process heat generally, as the stronger initial opportunity.

Why power sales look less attractive

For dedicated power production, the study found that its modeled reactor electricity costs exceeded even the highest average state wholesale price in its comparison by at least 36%. That calculation used 2024 electricity prices. It does not establish what a future project could earn under a different market arrangement, but it does show why the paper's industrial case cannot be applied to grid-only reactors without changing the assumptions.

The university account describes a further scenario in which later construction costs fall by about 20%. With the hydrogen credit, modeled profitable capacity then rises to 171.9 gigawatts electric; without it, the figure is 7.8 gigawatts. The projected increase depends on achieving that cost reduction as well as the other model inputs. Commercial operating experience from the US projects being assessed has yet to establish their actual construction costs and returns.

A possible route to industrial emissions cuts

The researchers also estimated that deployment across the sectors they examined could reduce US industrial carbon emissions by 8% in an initial construction wave and 14% in a later one, measured against a cited 2020 industrial emissions baseline. These reductions have not occurred. They depend on reactors being built, their output displacing higher-emitting industrial energy, and the cost and policy conditions in the scenarios holding in practice.

There is an established reason to study industrial customers. In an earlier assessment, the International Energy Agency identified placing nuclear hydrogen production near its users as a way to avoid transport costs. It also warned that substantial reductions in nuclear investment costs would be needed to compete with the hydrogen alternatives in its own analysis. That independent assessment gives context for the Michigan approach; it does not verify the newer model's profit estimates.

The IEA's 2026 hydrogen review says global hydrogen demand exceeded 100 million tonnes in 2025, almost all of it in industry and refining, while production of low-emissions hydrogen remained below one million tonnes. The agency identifies high costs, uncertain demand, regulation and infrastructure as obstacles to expanding low-emissions supply. Those conditions make industrial hydrogen a consequential market to examine, while leaving the scale and timing of any reactor-driven shift unresolved.

Study co-author Brendan Kochunas said the work could guide companies deciding where to deploy small reactors and inform government policy. For now, the model's most favorable applications are tied to a production incentive and capital costs that future US commercial projects have not demonstrated. Whether developers can secure those incentives, control construction costs and deliver hydrogen at the projected returns remains the test of the proposed industrial path.