U.S. Nuclear Buildout Faces Heavy-Forging Capacity Gap

On June 23, DOE’s Office of Energy Dominance Financing conditionally committed up to $17.5 billion in loan financing to help Westinghouse and as many as five utility or energy-company partners purchase long-lead equipment for ten AP1000 reactors, organized as five projects of two units each. DOE says the financing could accelerate deployment by as much as three years and supports a goal of having all ten under construction, with complete designs, by 2030. Westinghouse has signed letters of intent with seven potential partners, but the five that will ultimately receive financing have not yet been named, and each loan remains conditional on technical, legal, environmental, and financial requirements.

The financing is a genuine accelerant, not a substitute for manufacturing capacity that does not yet exist at scale. The single most consequential long-lead item in an AP1000 build is the reactor pressure vessel, the steel container holding the fuel, coolant, and reaction for the plant’s operating life, and the network of facilities capable of producing its largest components at commercial scale is small, concentrated overseas, and already booked years out.

Why Reactor Vendors Need Extremely Large Integral Forgings

A reactor pressure vessel is not forged as a single seamless object. It is assembled from several exceptionally large forged sections, including shell rings, heads, flanges, and nozzles, welded together into the finished vessel. Vendors prefer those individual sections as large and integral as possible, because every weld in a pressure vessel has to be inspected for decades of plant operation. Producing bigger single-piece shell rings reduces the number of welds and the long-term inspection burden that comes with them, which is why forging capacity, not welding capacity, sets the pace for how quickly components can be produced.

The World Nuclear Association reports that very large Generation III+ reactor components can require forging presses in the range of 14,000 to 15,000 tonnes and, for some designs, ingots weighing 500 to 600 tonnes. Westinghouse has separately identified a 15,000-ton press capable of taking 350-tonne ingots as its stated minimum for the largest AP1000 forgings specifically, a distinct and smaller figure that should not be conflated with the broader Generation III+ range.

A Small Global Supplier Base, Not One Overseas Press

The realistic supplier pool for AP1000-scale forgings is narrow, but not limited to a single machine. Japan Steel Works operates two 14,000-ton presses in Japan and claims the largest share of the world’s heavy nuclear-forging market, with plans announced this year to double nuclear-component production capacity by 2028. Doosan Enerbility, in South Korea, runs both a 13,000-ton and a 17,000-ton press, the latter handling ingots up to 540 tonnes, the largest single figure in the World Nuclear Association’s global table. Taewoong, also in South Korea, operates a 15,000-ton press. China First Heavy Industries and Shanghai Electric maintain heavy-forging capacity in China, and several Russian facilities round out the list. Commercial qualification, export controls, and existing vendor relationships narrow which facilities can realistically supply a given Western design, but the constraint is a small qualified ecosystem, not a single overseas press without alternative.

Throughput across that ecosystem is limited and uneven rather than fixed. The World Nuclear Association says roughly four pressure-vessel sets per year has been common for large presses scheduled alongside other orders, though potential capacity varies: Japan Steel Works has reported the ability to produce up to twelve sets annually, Doosan and China First Heavy Industries roughly five each, and Shanghai Electric around six. The binding constraint is qualified, scheduled capacity at a specific facility, not a uniform industry-wide rate.

Doosan’s Role in America’s Only AP1000 Projects to Date

Doosan Enerbility’s Changwon complex is one of the most important qualified suppliers available to U.S. AP1000 developers, and it has already supplied major components for the four original American AP1000 construction projects: the two completed units now operating at Vogtle in Georgia, and the two at V.C. Summer in South Carolina that were left unfinished when construction stopped in 2017. Santee Cooper and Brookfield are evaluating a potential completion of the V.C. Summer units, though a final construction decision has not been made. Some original equipment remains at the site, but its condition, licensing status, and economic case for reuse would still need to be verified before it goes into an operating reactor.

Fermi America provides an early example of how developers are responding to this constraint. In October 2025, it signed an agreement with Doosan to begin work associated with reactor pressure vessels and steam generators for its proposed Project Matador campus in Texas, well ahead of any construction start. Rolls-Royce SMR made a similar early-supplier move in May 2026, selecting Doosan and the Czech firm Škoda JS for pre-production and manufacturing-readiness work on nuclear-island components, including the pressure-vessel body, for its small modular reactor design. That reflects early supplier engagement rather than a confirmed forging slot, and an SMR’s vessel requirements are not necessarily comparable to the largest AP1000 forgings in scale.

The U.S. Has Press Force but Not an Integrated Heavy-Forging Supply Chain

Domestic capacity exists but does not match the requirement. Allegheny Technologies operates a 15,000-ton press, meeting the nominal force Westinghouse specifies for its largest AP1000 components, but the facility handles ingots of approximately 175 tonnes, well below the 350-tonne minimum Westinghouse requires. Lehigh Heavy Forge runs a 10,000-ton press capable of roughly 270-tonne ingots, and North American Forgemasters operates a comparable 10,000-ton press handling around 170 tonnes. The gap is not simply press force. It spans melt capacity, ingot size, component handling, precision machining, and nuclear qualification, an integrated production system the World Nuclear Association says the U.S. never rebuilt after domestic nuclear orders dried up in the 1970s. BWXT remains one of North America’s most important qualified nuclear-component manufacturers, but its principal location for large commercial nuclear components is Cambridge, Ontario, not the United States.

DOE’s $17.5 Billion Commitment Secures Components Before Construction

The financing does not pay for ten completed reactors. It helps Westinghouse and its eventual project partners order long-lead components early, standardize purchases across projects, and reserve manufacturing capacity before individual construction schedules are finalized, the same logic already driving how utilities and manufacturers are reserving scarce transformer and switchgear capacity years ahead of need. A forging slot booked early is difficult to replicate on short notice, however well-financed a later entrant might be. Edwin Lyman, director of nuclear power safety at the Union of Concerned Scientists, has noted that the loan package is a fraction of what ten reactors would likely cost in total, and none of the five potential projects yet has a construction contract in place.

Britain’s New Press Won’t Close the Gap Either

Sheffield Forgemasters, owned by the UK Ministry of Defence, is building a new forge for a 13,000-tonne press as part of a broader £1.3 billion recapitalization tied mainly to the AUKUS submarine program. Installation is scheduled to run from July 2027 through roughly July 2028, after which commissioning and qualification work would still be required before the press could produce reactor-grade components. That press falls below Westinghouse’s stated 15,000-ton minimum for the largest AP1000 forgings, so on its own it would not eliminate the UK’s reliance on overseas capacity for that specific design, though it would add qualified capacity for other reactor types and components.

The World Nuclear Association is explicit that this challenge is not confined to reactor pressure vessels. It extends to steam generators, turbines, and other large engineered components, each drawing on an overlapping, similarly constrained set of qualified manufacturers. The practical lesson for utilities and financiers underwriting this wave of projects is not that one machine determines the schedule. It is that even projects clearing every domestic regulatory milestone still depend on an integrated global manufacturing base the U.S. can finance its way into faster than it can rebuild.