US AI Data Centers Face Chinese Power Equipment Dependence

The United States’ growing reliance on China for critical components powering its burgeoning data center infrastructure is facing increased scrutiny. This dependence, amplified by the escalating competition for artificial intelligence supremacy between Washington and Beijing, could lead to escalating costs and exacerbated supply chain disruptions for the ongoing AI buildout.

US hyperscale operators are engaged in a multi-billion dollar race to construct data centers essential for AI development. However, a significant portion of the components required for these facilities are supplied by Chinese firms. Analysts indicate that these essential parts include transformers, switchgear, batteries, and advanced optical technologies.

In response to what is perceived as an “extraordinary foreign threat” to the U.S. bulk power system, President Trump recently signed an executive order declaring a national emergency. This directive empowers the Department of Energy to impose restrictions or conditions on certain transactions involving components utilized in the power grid and data centers. The move underscores growing concerns over national security implications stemming from foreign-sourced electrical equipment.

Tensions between the U.S. and China have intensified as Chinese AI models become more sophisticated and gain wider global adoption. While American companies lead in the development of cutting-edge AI chips and hardware, China’s influence over the power infrastructure for data centers is increasingly viewed as a strategic vulnerability for the United States.

“The spotlight on China’s role in the U.S. data center power ecosystem has intensified considerably over the past eight months or so. This represents a notable shift from the prior focus, which was almost exclusively on the compute stack,” Laveena Iyer, senior analyst at The Economist Group, commented to CNBC.

The power infrastructure, or “power stack,” of a data center is paramount to ensuring uninterrupted operations and minimizing downtime. This complex system encompasses transformers, which regulate high-voltage electricity from the grid to the lower levels required by servers and cooling systems. Switchgear, a centralized network of switches, fuses, and circuit breakers, is vital for managing power distribution and protection. Additionally, robust battery backup systems are indispensable for maintaining operational continuity during power outages.

“Several key areas within the AI infrastructure buildout exhibit a pronounced dependence on Chinese imports,” stated Ben Boucher, senior supply chain analyst at Wood Mackenzie. “Most prominent among these are substation transformers, which hyperscalers typically deploy on-site to reduce transmission voltage.”

“Our mid-term analysis highlights a significant concentration of exposure within grid connectors, specifically transformers, switchgear, and batteries,” noted Yury Dvorkin, an associate professor at Johns Hopkins University. “China’s market share in certain transformer and switchgear categories approaches 30%, and the nation accounts for over 40% of U.S. battery imports. Our investigations reveal an even deeper upstream reliance, encompassing essential raw materials such as copper, electrical steel, and battery cathode materials.”

Furthermore, optical technology, which leverages light transmitted through fiber-optic cables for high-speed data transfer, is another domain where Chinese dominance is pronounced. This technology is crucial for handling the immense data volumes associated with AI workloads.

Leading global suppliers in the data center optical transceiver market include companies such as Zhongji Innolight and Eoptolink. Collectively, Chinese firms are estimated to supply approximately two-thirds of the global unit volume for these critical components, according to research firm Counterpoint.

Earlier this year, in a separate Presidential Determination related to the Defense Production Act of 1950, President Trump identified transformers, transmission lines and conductors, substations, high-voltage circuit breakers, and power control electronics as being “essential to the national defense.” This designation signaled a strategic intent to secure domestic capabilities in these vital infrastructure areas.

Washington is actively pursuing strategies to mitigate its dependence on Chinese suppliers as it accelerates the development of the necessary infrastructure to power the AI revolution. The urgency is underscored by the rapid expansion of AI capabilities and the consequent demand for robust and secure data center operations.

“Since my first term, the threat to the United States regarding foreign supply of bulk-power system electric equipment has become even more acute,” President Trump stated in connection with the recent executive order. This sentiment highlights a sustained concern over the vulnerabilities introduced by international sourcing of critical energy infrastructure.

“The rapid growth of advanced manufacturing, data centers, artificial intelligence, and defense production has increased the Nation’s dependence on abundant, reliable electricity and magnified the consequences of a successful attack or supply disruption on the bulk-power system,” he further elaborated, emphasizing the critical link between energy security and national security in the context of emerging technologies.

The U.S. data center capacity is projected to experience substantial growth, soaring from an estimated 62 GW in March 2026 to 152 GW by 2030, primarily driven by the high-density computational demands of AI workloads, according to a June report by S&P Global. This projected expansion necessitates a secure and resilient supply chain for the underlying power infrastructure.

A White House spokesperson conveyed to CNBC that “Reshoring manufacturing that’s critical to our national and economic security has been a top priority for President Trump, and the Administration continues to deliver with a robust and nimble agenda of tax cuts, tariffs, and deregulation. Trillions in investments across key sectors – from steel to semiconductors to autos – prove that the Administration’s strategy is paying off.” This statement reflects a broader economic and industrial policy objective aimed at bolstering domestic manufacturing capabilities.

In a significant move to bolster domestic production of critical grid infrastructure, Hitachi Energy announced in September 2025 a $1 billion investment, including $457 million for a new large power transformer facility designed to meet the escalating demand from AI initiatives. Similarly, Siemens Energy stated in February its intention to invest $1 billion in U.S. manufacturing capabilities for grid and gas turbine equipment, specifically targeting the expansion of AI infrastructure and data centers.

Further underscoring the national security concerns, power inverters produced in foreign countries, essential for interconnecting energy sources for AI facilities, were added to the Federal Communications Commission’s (FCC) Covered List in July. This list comprises equipment and services deemed by the U.S. government to pose an unacceptable risk to national security.

Reports also indicate that the Trump administration is developing a ban on U.S. imports of new Chinese optical transceivers. These devices are fundamental for high-speed data transmission over fiber-optic cables between data centers. While the FCC did not provide comment on this matter, the White House has not yet officially addressed the reports.

Any restrictions placed on Chinese optical technology firms could potentially benefit U.S. companies such as Lumentum and Coherent. These companies have already attracted significant investment, with each receiving $2 billion from Nvidia in March. However, analysts caution that scaling manufacturing to meet the surge in demand presents considerable challenges. “Western competitors like Coherent and Lumentum possess advanced photonic designs, but currently lack the cleanroom capacity, automated packaging infrastructure, and yield scale required to absorb Innolight and Eoptolink’s volume within a 12-to-24-month horizon,” Neil Shah, VP research at Counterpoint, noted in an August report. This highlights a potential bottleneck in achieving swift domestic production scale.

The implementation of a ban on optical transceivers could necessitate that U.S. hyperscalers seek domestic alternatives, potentially leading to increased operational costs, according to The Economist Group’s Iyer. This would represent a direct pass-through of costs to consumers and businesses relying on these data services.

The market is already facing a significant shortfall in critical power components. Power transformers and substations are experiencing an estimated shortage of 15% and 8%, respectively, in 2026. Restrictions on units manufactured in China are poised to further complicate and worsen these existing supply chain challenges, according to research from Wood Mackenzie. “A bulk of the impacts will be centered around data centers who have been using Chinese units to minimize lead times,” Wood Mackenzie’s Boucher stated in an August blog post. “The 100 MVA+ segment is where the shortage is already most acute, and it is precisely where data centers have been turning to Chinese manufacturers to manage lead times.”

However, Iyer further observed that political will in Washington to reduce dependencies on China within the power stack is demonstrably rising. “The U.S. administration appears keen to de-risk its AI infrastructure before Chinese tech is fully embedded in it, quite like the 5G rollout where removing Chinese telecoms gear later on led to network rollout delays and rising costs for telecoms companies.” This strategic foresight aims to prevent costly and disruptive remediation efforts in the future.