Trump’s tariffs threaten US AI leadership despite push to become ‘world capital’

Representational Image

US President Donald Trump’s renewed ambition to make the United States the “world capital” of artificial intelligence is facing growing pushback from the very companies tasked with building that future. While Trump has sought to present himself as an AI champion—hosting industry leaders and launching billion-dollar plans—his escalating trade war with China is driving up the cost of building AI infrastructure and creating uncertainty that could undermine America’s global competitiveness in the field, the Washington Post reported.

Though Trump exempted certain electronics like smartphones, computers, and AI chips from his latest wave of tariffs late Friday, much of the hardware needed to construct and operate AI data centres remains affected. These include critical components like cooling systems, transformers, and backup generators, most of which are sourced from or assembled with materials from China.

Mounting costs and confusion

Executives and industry experts say the tariffs could raise the cost of constructing AI data centres by up to 20 percent. “It sounds good that chips are exempted,” said OpenAI CEO Sam Altman, “but there are so many other parts to the cost of a data centre.” Altman said his team has been working round-the-clock to assess the financial implications of the tariffs.

While semiconductors—crucial to running advanced AI models—were spared for now, Commerce Secretary Howard Lutnick confirmed on Sunday that new levies are still being planned under Section 232, a national security statute. These would go through a longer regulatory process but have already fuelled industry concerns about long-term costs and supply chain disruptions.

Josh Levi, president of the Data Center Coalition, said the AI industry “urgently needs certainty” to make large-scale infrastructure investments. Without it, companies could reconsider the U.S. as the default destination for next-generation AI infrastructure.

Stargate and sky-high ambitions

Trump’s Oval Office event last month included announcements from OpenAI, Oracle, and Japan’s SoftBank of a proposed $500 billion AI infrastructure initiative dubbed “Stargate.” Separately, Google plans to spend $75 billion and Microsoft $80 billion on AI data centres this year alone.

The AI race, initially ignited by OpenAI’s release of ChatGPT in 2022, has triggered a frenzied global buildout of data centres—enormous facilities consuming massive energy and hardware resources. AI models’ growing complexity has made these centres central to national economic and security strategies.

Don Clark, co-CEO of data centre construction firm Clark Pacific, called the current AI data centre boom “the largest market I’ve ever seen in my career.” But he and others caution that Trump’s tariffs risk slowing momentum just as companies ramp up investment.

A global shift?

Real estate consultant Jay Biggins noted that demand for key components was already stretched thin before the tariffs. “All of the essential hardware was already on 24- to 36-month backlogs,” he said, suggesting companies may soon face even higher costs—or decide to build elsewhere.

That’s a scenario Andrew Ng, former head of Google’s AI division, finds increasingly likely. “Tariffs will definitely increase the cost,” he said. “When the regulations change overnight by tweet, it’s difficult to plan.”

AI firms already maintain data centres in regions like Malaysia and Singapore, where costs are lower and regulatory environments more predictable. Given the processing latency tolerances in AI applications, the geographical proximity to customers—once a key factor in data centre placement—is becoming less relevant.

Balancing ambition and impact

Trump has positioned AI as a cornerstone of U.S. strategic leadership, touting deregulation and reduced permitting as key to unlocking growth. But by imposing sweeping tariffs—some as high as 145 percent—his administration may inadvertently empower China and other rivals in the AI arms race.

Despite appearances of relief from the recent chip exemption, the broader industry remains wary. “This is a moment of serious uncertainty,” said Ng. “And unfortunately, that’s the last thing the US needs if it wants to lead in AI.”