New York (CNN) — If there’s one thing the White House, Wall Street and Silicon Valley can agree on, it’s that artificial intelligence is a top priority.
Tech giants are pouring billions into new data centers and infrastructure to support the technology. The White House came out with an AI action plan in July to boost America’s leadership in the space, underscoring the tech’s importance to the administration. Wall Street keeps pushing AI-related stocks like Nvidia (NVDA) to new records.
But President Donald Trump’s trade war has raised questions about whether the administration’s policies could work against its big AI push. Certain tariffs could raise the costs of materials and components necessary to support those AI models.
For example, the president said on August 6 that he would issue a 100% tariff on semiconductors imports, although he added that companies that have committed to expanding their manufacturing operations in the US would be exempt. (He did not give an exact timeline for when those tariffs would start.) And in late July, he imposed a 50% tariff on copper, which is used in electronic components such as printed circuit boards and chips.
But while tariffs could stoke uncertainty around costs, experts say they won’t slow technological advancements, primarily because the stakes are simply too high to fall behind in the global AI race.
For large tech companies like Meta and Microsoft, losing in AI would be a higher price to pay than any additional costs from tariffs. Dallas Dolen, the US technology, media and telecommunications lead for PricewaterhouseCoopers, said these types of companies likely view the AI boom as an “existential moment” for their businesses.
“Cost, if you have enough money, is not the most important variable that you take into account when you’re told it’s an existential threat,” he said to CNN.
Big Tech is pouring billions into AI
When Meta, Microsoft and Google reported earnings in late July, one message rang loud and clear: Big Tech is spending big on AI, and it’s starting to pay off.
Meta spent $17 billion in capital expenditures for the quarter that ended in June, and it saw its earnings per share go up 38% compared to a year ago. Capital expenditures typically refer to money spent on things like data centers and infrastructure, likely a sign that Meta is investing more in the servers needed to power its burgeoning AI services.
Wall Street cheered the results; Meta shares (META) rose 9% in after-hours trading when it posted the results on July 30, and shares are up roughly 30% year to date.
Microsoft (MSFT) also posted strong results thanks to its cloud computing business. It spent $24.2 billion in capital expenditures during its most recent quarter, and it plans to spend another $30 billion in the coming months, the company said in late July. Microsoft became the second company to reach a $4 trillion valuation last month, following Nvidia, and its shares are up about 26% so far this year.
And Google parent Alphabet increased its capital expenditures for 2025 to $85 billion because of demand for its cloud products. The company said its cloud services are used by “nearly all gen AI unicorns,” referring to privately held companies worth $1 billion or more in the generative artificial intelligence space. Alphabet shares (GOOG) are up nearly 7% year to date.
That additional infrastructure may be essential; Goldman Sachs estimates that global power demand from data centers will surge 50% by 2027 and 165% by 2030 compared to 2023 because of AI.
“We have barely scratched the surface of this 4th Industrial Revolution now playing out around the world led by the Big Tech stalwarts such as Nvidia, Microsoft, Palantir, Meta, Alphabet, and Amazon,” Wedbush Securities analyst Dan Ives said in a research note following the companies’ earnings results.
How tariffs could impact the AI push
Trump’s rapidly changing tariff policies have made it difficult to estimate how exactly the levies could impact the cost of building and operating data centers.
But PwC’s Dolen said he’s seen estimates indicating that tariffs could increase construction costs by 5% to 7%. The National Association of Manufacturers’ outlook survey also found that trade uncertainties and increased costs of raw materials were the top business challenges for manufacturers in the first quarter of 2025.
However, big tech companies are likely to eat any additional costs related to AI infrastructure because “demand is so strong,” said Michelle Brophy, director of research for tech, media and telecom at market intelligence firm AlphaSense.
It’s a different story for smaller companies that don’t have billions to spend each quarter. They also typically have private investors demanding a fast return on investment, and data centers are long-term bets that could take years to show value in a meaningful way.
Between 2015 and 2020, it took one to three years on average to construct a data center, according to commercial real estate services firm CBRE. And a data center is useful for 25 years to 30 years on average, McKinsey & Company senior partner Pankaj Sachdeva said in October 2024.
Because data centers are long-term projects, “the degree of uncertainty will have a larger impact in terms of, you know, committing to something that will take multiple years to execute,” said Laurence Ales, a professor of economics at Carnegie Mellon University.
It’s also unclear whether Trump’s semiconductor tariffs will raise the cost of future data centers. The president said companies that have “committed” to building in the US won’t have to pay a levy on semiconductors.
“But the good news for companies like Apple is, if you’re building in the United States, or have committed to build, without question, committed to build in the United States, there will be no charge,” he said on August 6 during an event announcing Apple’s $100 billion initiative to produce iPhone parts in the US.
Trump didn’t specify which companies would be exempt, but chipmaking giants Nvidia and TSMC have both said they would expand their US operations.
Experts believe more collaboration between the White House and Silicon Valley is likely to come, possibly easing any potential tariff-induced costs for tech giants. Trump showed his willingness to negotiate with tech leaders earlier this week: He allowed Nvidia and AMD to sell their AI chips to China as long as they provide a 15% cut to the US government in exchange for export licenses. And the White House is reportedly discussing taking a stake in chipmaker Intel.
Building AI infrastructure is a key part of the White House’s AI action plan, which includes policy recommendations for streamlining permits for facilities like data centers and semiconductor manufacturing facilities.
The United States already has more data centers than any other country, according to data from Cloudscene, a platform that connects businesses with cloud services, compiled by Statista. Many of the world’s largest cloud providers, like Microsoft and Amazon, are American companies.
“We need to be mindful that this is an area in which we have an advantage,” Matt Pearl, director of the strategic technologies program at the Center for International and Strategic Studies, said to CNN. “And we don’t want to give that up.”
The-CNN-Wire
™ & © 2025 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.
Originally Published: