Introduction
This article was flagged due to widespread concerns about the United States’ ability to domestically source crucial parts required for a $100 billion artificial intelligence infrastructure initiative by OpenAI and Nvidia. With public scrutiny focused on foreign dependency and potential slowdowns linked to supply chain bottlenecks, we examine whether these claims hold up against the available evidence.
Historical Context
For decades, the U.S. has claimed technological leadership in both artificial intelligence and large-scale infrastructure while gradually shifting significant portions of manufacturing abroad. From heavy machinery to highly specialized components required for power generation or tech infrastructure, a sizable share is sourced globally due to cost savings, capacity, and limited domestic output. The current push for massive AI data centers and the energy systems to run them has sharpened debate over America’s industrial readiness and economic security.
Fact-Check of Key Claims
Claim #1: “Heavy-duty turbines, nuclear plant components, and transformers are largely foreign-sourced and indispensable for powering AI infrastructure.”
The article states, “There are four major categories of equipment that are both extremely expensive and largely foreign-sourced,” highlighting turbines, nuclear components, and transformers. According to the U.S. Department of Energy and the International Energy Agency, the utility-scale turbine market is indeed dominated by GE Vernova (U.S.), Siemens (Germany), and Mitsubishi (Japan). However, GE Vernova is based in the U.S., though it still sources some parts internationally. For nuclear plant pressure vessels and ultra-large forgings, U.S. domestic manufacturing capacity has dramatically decreased, with the last domestic producer ceasing these parts in the early 2000s. Recent domestic nuclear projects, like Vogtle in Georgia, sourced main components from Doosan (South Korea), supporting the article’s assertion. When it comes to large transformers, the U.S. Department of Commerce recognizes that over 80% of high-voltage, utility-scale transformers are imported, highlining dependence on foreign suppliers for grid expansion. This claim is accurate and well documented by industry data and government records.
Claim #2: “Heavy reliance on imported steel and related tariffs will significantly drive up costs and add complexity to the AI infrastructure build-out.”
The article notes that “higher tariffs on imported goods only add to the complexity…” and specifically cites steel—where even with U.S. and U.K. production, imports are often needed for cost and capacity reasons. According to U.S. International Trade Administration data, the U.S. imports around 25-30% of its steel each year, relying heavily on Canada, Mexico, Brazil, and some European countries. Recent tariffs—including the 25% steel and 10% aluminum tariffs originally established under “Section 232” in 2018—remain in effect for most countries, with some exceptions for U.S. allies. The U.K. and Japan, for example, are currently exempted from increased tariffs due to negotiated agreements. Industry sources confirm that tariffs can add 3-6% to total project budgets, a point echoed by the Congressional Research Service and widely reported in major media outlets when discussing large-scale energy and infrastructure projects. The claim that tariffs and heavy import reliance drive up costs is accurate and aligns with industry analysis.
Claim #3: “A critical shortage of skilled U.S. labor could become as significant a challenge as material supply issues.”
The article quotes, “We simply don’t have the skilled workforce in place to scale this fast. U.S. [building] trades are already in decline, with a shortage of welders, machinists and electricians.” According to the Bureau of Labor Statistics and the Associated General Contractors of America, the U.S. construction and manufacturing sectors have faced chronic labor shortages for several years, particularly in skilled trades such as welding, machining, and electrical work. These reports note persistent unfilled job openings, a declining pipeline of new skilled workers, and increased retirements, all of which hamper large-scale projects. The need for a rapid expansion of the industrial labor base—likened in the article to “wartime”—accurately reflects both quantitative and qualitative industry assessments.
Conclusion
Assessing the article as a whole, its depiction of America’s heavy reliance on foreign-sourced parts for large-scale AI infrastructure is accurate according to available government, industry, and labor data. While some U.S. manufacturing capacity remains, critical components such as reactors, turbines, and transformers are either largely or substantially imported, which substantiates the supply chain concerns. The reporting also accurately frames tariffs and labor shortages as key add-on obstacles. The article avoids overt sensationalism and presents a fact-based, contextualized analysis. However, while it highlights risks, it could offer more detail on measures to boost domestic capacity or on ongoing government and industry responses. Overall, it substantially reflects current realities without major evidence of misinformation or bias.
Take Action Now
Stay empowered against misinformation. Download the DBUNK App to join a trusted community and submit your own fact-check requests for free.
Link to Original Article
Read the original coverage at this CNBC link.
