This week’s $100 billion investment promise from OpenAI and Nvidia, which includes supply chain data, highlights the US’s significant reliance on foreign vendors for project components.
Building conventional power plants, whether gas-fired or nuclear, requires a variety of specialized parts that the United States doesn’t make in significant quantities, Brandon Daniels, CEO of AI-powered risk and supply chain management solutions business Exiger, told.
There are four main types of equipment that are very costly and mostly imported, and this reliance occurs at a time when power building is already facing significant shortages and backlogs,” Daniels said.
First up are gas turbines
According to Exiger supply chain data research, three Original Equipment Manufacturers (OEMs) lead the global market for heavy-duty, utility-scale turbines: Mitsubishi of Japan, Siemens of Germany, and GE Vernova of the United States.
According to Daniels, the three oversee the supply of new turbines, and about half of them come from overseas. The infrastructure of [artificial intelligence] on this scale will be powered by these enterprises, he added.
Nuclear components
Ultra-large nuclear plant components and forgings fall into the second group.
The enormous, single-piece reactor pressure vessels and related hardware are no longer produced in the United States, according to Daniels. For instance, Doosan, a company located in South Korea, manufactured the steam generators and pressure vessels for the newly constructed reactors at the Vogtle facility in Georgia, which is the first new nuclear plant to open in the United States in a generation.
A third link in the supply chain includes large transformers that are necessary for transmitting power from producing facilities into the electrical grid. These suppliers are located abroad.
Daniels stated that vendors from nations ranging from South Korea to Germany to Canada produce over 80% of these high-voltage transformers.
In this supply chain, steel is the fourth component that is most likely to raise prices.
“Project developers often depend on imports to meet both cost and capacity needs, even though the U.S. and allies like the U.K. remain significant producers,” Daniels said.
Impact of Tariffs
Increased import duties only make these initiatives more complicated.
Because the planning and buildout would take years rather than months, and because U.S. domestic capacity may increase throughout that time, the tariff impact isn’t consistent, Daniels said. Lower tariff regimes also apply to certain associated suppliers; the United Kingdom, for instance, is now free from the 50% U.S. steel tariff. However, the amount of material and equipment needed will increase even with these carve-outs.
A 3%–6% tariff-driven budget rise on multibillion-dollar energy projects adds up to hundreds of millions of dollars, Daniels added.
“This is only in aluminum and steel,” Daniels remarked. Turbines, reactor vessels, and transformers are examples of components that are acquired from overseas, where risk is increased by backlogs and import dependency. In addition to being a technical difficulty, the combined effects of tariffs and sourcing realities present a substantial supply chain and trade policy barrier for the development of AI infrastructure.
Include the amount of labor assistance required to complete the plans.
According to Daniels, this calls for a massive expansion of the industrial base that is comparable to that of a war. Simply said, we lack the trained personnel necessary to expand this quickly. There is already a lack of electricians, machinists, and welders in the U.S. [construction] trades. The hardware itself might not be as much of a pain point as this. Over the next five years, we will need to face that truth in order to make this even somewhat feasible.
