Trump's Semiconductor Tariffs Could Cost US $90B Annually, Threaten AI Infrastructure Buildout

AI industry says Trump plans to tax chips in the "single dumbest way imaginable

Donald Trump may be preparing to announce sweeping new semiconductor tariffs at the absolute worst time, the tech industry fears. On Thursday, Politico reported that a wide range of new semiconductor tariffs, which the tech industry expects will "doom" artificial intelligence innovation in the US, could be imposed in the coming "weeks or months." About eight people familiar with the Trump administration's plans were granted anonymity to discuss how the tariffs might work. They suggested that the framework for tariffs may change as it becomes finalized, but one approach under consideration could "dramatically expand the number of tech products subject to the duties, hitting not just chips but potentially many of the goods made with them, such as gaming consoles or the servers that fill data centers." That scenario is the tech industry's nightmare. Since Trump took office, trade groups have warned that imposing tariffs on both semiconductors and the many downstream products that use semiconductors -- potentially even taxing used or refurbished products that contain chips -- would be ruinous for the economy. Most absurdly, given the priorities of Trump's AI Action Plan, that tariff approach would cost the US about $90 billion annually in GDP losses and cause about 20 percent of data center projects planned through 2030 to be delayed or canceled, the Computer and Communications Industry Association (CCIA) estimated in June. The tariffs could even drive more data center development outside the US, the CCIA warned, which seems counterproductive, since Trump is imposing them in order to force more development into the US. Beyond disrupting the momentum of data centers, the tariffs could have wide-ranging ripple effects if certain products aren't exempted, the CCIA said in a May letter to Treasury Secretary Scott Bessent, which was cosigned by about 20 trade groups. For consumers, prices of "everyday tools," like smartphones, laptops, tablets, smartwatches, connected devices, and vehicles, could increase at a time when US households are already budget-strained. Tariffs could also limit technology choices for Americans by delaying new product launches, including devices featuring the latest AI technologies. And any dampened demand for popular tech risks further limiting innovation while seemingly working in lockstep with tariffs to slow AI adoption in the US, the letter warned. "Consumer devices are the primary interface through which Americans access AI-powered tools. AI only delivers on its promise when people can actually use it -- and tariffs that price consumers out of the device market would slow AI adoption at the very moment the United States is positioned to lead," the letter said. To shield AI firms, the Trump administration is mulling some tariff relief, but sources told Politico that it would likely be tied to foreign firms investing in US chip manufacturing, like Taiwan Semiconductor Manufacturing Co. Apparently, that's the approach to relief that Commerce Secretary Howard Lutnick most favors. Perhaps mindful of the data center development timeline -- and possibly the upcoming holiday season, as Trump previously has exempted goods from tariffs to avoid consumer backlash that negatively impacted his approval ratings -- the administration may impose tariffs in phases to avoid the worst impacts hitting all at once. Industry talks with Trump turn "negative" The global data center buildout already has the tech industry broadly scrambling for access to high-end semiconductors, which are forecasted to remain in short supply well into 2027. Just yesterday, Gartner forecasted that due to the shortage that's driving price increases, global semiconductor revenue is expected to reach $1.6 trillion in 2026 -- much sooner than expected. Economists expect that tariffs would only further increase prices for chips, which Politico reported could "hammer US chip designers such as Nvidia and Advanced Micro Devices, which rely on overseas manufacturers to produce their chips." Tariffs could also hurt companies like Apple, which competes with foreign rivals overseas that wouldn't have to deal with the tariffs. Chinese firms could benefit, Politico suggested, as chip suppliers potentially avoid tariffs by increasing business in China. Politico's sources confirmed that the tech industry is so panicked by the risk of semiconductor tariffs that it "launched a lobbying blitz" to guarantee that data centers at the very least would be exempted. For the tech industry, the hope is that Trump will recognize an inherent tension at the heart of his semiconductor tariffs plan: Taxing data centers on chips during the key time they need to acquire as many as possible to scale AI infrastructure will not increase demand for chips produced domestically for a simple reason. Domestic chip plants take years to build, and the US can't rush that timeline, so firms will remain heavily dependent on importing semiconductors. As a detailed report from The Next Web summarized the problem: "Taxing the imports in the meantime raises the cost of the thing the administration also says it wants, which is American AI infrastructure at scale. There is no version of the timeline in which domestic supply arrives before the buildout needs the chips." "This may be the single dumbest way imaginable to pursue American dominance in AI," said one tech official who, Politico reported, is from a major industry group and also served in the first Trump administration. "It's like kneecapping yourself at the starting line." However, even though industry representatives and lobbyists have met with Trump officials "with increasing frequency since the start of summer," sources suggested that this time, the tech industry may not get the exemptions that they're seeking. The talks have recently "trended in a negative direction," sources told Politico. Lutnick appears to be standing in the way. According to Politico's sources, the Commerce Secretary thinks the US needs to apply the tariffs broadly to prioritize reshoring the domestic supply chain. Four sources suggested that Lutnick plans to hold his ground and announce a system where the "US would let a set volume of chips enter the country duty-free, with the size of that allowance tied to how much companies pledge to produce on American soil." Individual countries could also be hit with different semiconductor tariff rates, Trump officials have weighed, further complicating the math for companies trying to guess how tariffs might shake up their supply chains. Critics are worried that the structure that Lutnick likes for tariffs "risks widening the gap between the supply of tariff-free chips and the volume US companies need," Politico reported. One tech representative suggested more exemptions would be needed, since "the volume they're talking about granting duty-free wouldn't cover the hyperscalers alone, let alone the rest of the industry. Those are chips we physically can't buy here, because the capacity doesn't exist yet." "The math literally just does not work," that source said. What does the tech industry want? Earlier this year, Trump imposed a narrow set of semiconductor tariffs that pointedly exempted data centers. As recently as May, US Trade Representative Jamieson Greer has said that "having semiconductor tariffs is important," but they must be "on the right timing and in the right amount." At that time, Greer said that the government wanted to ensure that "there are no immediate tariffs on companies that are producing semiconductors" and that the US "will allow companies to import an unspecified amount during that 'reshoring phase.'" It's unclear what may have changed to shift the Trump administration's position since then. However, The Next Web analyzed the upcoming phase of tariffs and noted that the Commerce Department submitted a report on July 1 that would determine if the data center exemption would survive. That report remains unpublished but is likely at the center of officials' discussions with tech lobbyists. The CCIA's report suggested that the next round of tariffs could avoid the worst consequences by carving out exemptions for semiconductors used in AI servers and by lowering the tariff rate from a presumed 25 percent to 10 percent. That would supposedly ensure the "most fragile" projects aren't disrupted and that GDP losses are minimized. The best solution, their letter said, would be to avoid imposing tariffs on semiconductors and derivative products entirely. But if the administration must proceed, then trade groups hope they will consider applying tariffs in "a nuanced and targeted way," their letter said. That means perhaps excluding products from tariffs "with minimal semiconductor content or products whose semiconductor inputs are incidental to their principal use." Or possibly setting a "de minimis" threshold where any product under a certain weight or value might be exempted. And at minimum, they should acknowledge that sellers of secondhand products may not have information on hand to establish the provenance of the chips contained inside. To ensure that there aren't shipment bottlenecks, any tariff policy should only require paperwork that firms already have on hand, the letter said. Finally, and likely the tech industry's biggest ask: If tariffs are inevitable, don't stack tariffs so that firms must pay a double tax on both the semiconductors within the derivative products and the products that contain them. For the tech industry, the risk of tariffs coming at this key moment likely feels like Trump is adding a roadblock where he could be focused on cutting red tape that could expand US chipmaking capacity, data center development, and AI innovation. Right now, investments in US chipmaking capacity are at a "historic high," CCIA's letter to Bessent said, and we need "policy measures that expand domestic capacity, accelerate permitting and infrastructure deployment, support advanced packaging and workforce development, and strengthen trusted supply chains [to] better advance US resilience without imposing economy-wide costs on downstream innovation."

President Trump's Love of Tariffs Comes Into Conflict With His Love of Data Centers

Since returning to office in early 2025, Trump has imposed sweeping tariffs that have disrupted nearly every sector of the global economy. Now they've found an unlikely new target: artificial intelligence, an industry the president has vowed to protect. According to new reporting from Politico, the Trump administration is discussing a round of tariffs on semiconductors, despite objections from Silicon Valley. Citing anonymous sources, Politico reports that the Trump administration is considering tariffs that would impact not just the flow of foreign-made chips into the country but also the countless Silicon Valley products that rely on these chips. Data centers are one obvious item, but it will also likely hit consumer electronics like laptops, which are already suffering from notable price hikes due to a global supply crunch on memory chips. That supply crunch has been driven by a surge in demand for AI, leading to an unprecedented AI infrastructure buildout that has eaten up chipmaking capacity. The AI-driven data center buildout in the United States, and much of the rest of the world, depends on crucial chips and parts coming from a handful of Asian companies, chief among them the Taiwanese semiconductor giant Taiwan Semiconductor Manufacturing Company, better known as TSMC. The company accounts for 73% of the global semiconductor foundry market share, and even supplies American leader Nvidia with processors. President Trump has made "winning the AI race" one of his administration's top priorities. In practice, that has meant clearing away virtually anything standing between the AI industry and its goal of putting a data center on every street corner. But the President has another top priority that, in this instance especially, is colliding with the other: to bring chip manufacturing stateside. In the pursuit of this goal, the Administration struck multibillion-dollar investment deals with foundry giants like TSMC and has taken government stakes in once-glorified but now-struggling chipmakers like Intel. Trump's commerce chief Howard Lutnick is reportedly gunning for tariffs that would allow foreign companies to evade these tariffs, but only if they invest in U.S. chip manufacturing capacity. Under Lutnick's vision, the amount of chips that can enter duty-free will reportedly be based on how big an investment the company makes. That arrangement might work better in theory than practice, though, experts warn. Building a foundry business from scratch is already a daunting task, particularly given the limited infrastructure and expertise available in the United States. It is likely to cost American businesses an absurd amount of time and money to bring chip manufacturing to the capacity that is needed to be the majority supplier of the hefty infrastructure buildout goals laid out by the AI industry, while also making sure other tech and adjacent sectors that rely on these chips can get enough supply to keep their businesses running. The timeline is in years, or even decades, according to some experts and tech industry insiders who spoke to Politico. Until then, American tech companies will have to continue to depend on a flow of chips from abroad. The American chipmaking industry is unlikely to scale up fast enough to make up for lost supply if that flow gets hindered by Trump's potential tariffs, potentially causing chaos not just for the AI industry but also across other sectors that rely equally heavily on semiconductors, like banking, automotive and even healthcare. "This may be the single dumbest way imaginable to pursue American dominance in AI," an anonymous tech lobbyist and former Trump official told Politico.

What the US semiconductor tariff does and does not currently touch

One study puts the cost of removing it at $90bn a year and 243,000 jobs, with roughly $450bn of planned capacity cancelled, delayed, or built elsewhere. The 25% tariff the United States imposed on advanced semiconductors in January comes with a carve-out that has kept it from touching the AI buildout, exempting covered products destined for American data centres. That exemption was always provisional, and the Commerce Department report that determines whether it survives was due on 1 July, in a policy environment that has already produced a draft ban on Chinese equipment in data centres. The report has not been published. Until it is, every operator planning American capacity is working from an assumption rather than a settled rule. The tariff itself is narrower than the headline rate suggests. It applies to three tariff subheadings covering data processing machines and their parts, and only to logic integrated circuits within specific performance and memory bandwidth bands, which is a deliberately surgical way of hitting AI accelerators without catching laptops. Seven end-use exceptions sit alongside it, covering repairs, research, startups, public sector use, consumer electronics, civil industrial applications, and data centres. The data centre one is the expensive exception, and it is the one under review. The proclamation also reserved authority to go considerably further. It explicitly contemplates tariffs on semiconductors, semiconductor manufacturing equipment, and their derivative products, which is the language that would reach servers rather than just the chips inside them. For an operator, the difference between the two positions is not marginal. A tariff on the accelerators alone is painful, whereas one that reaches assembled servers changes the arithmetic of where a facility gets built. An analysis published in June by the Computer and Communications Industry Association put a figure on removing the carve-out. Applying the 25% rate to data centres would cost the American economy about $90bn a year and put 243,000 jobs at risk. The arithmetic behind that is not complicated, which is both its strength and its weakness. Trevor Wagener, the association's chief economist, multiplies the share of data centre spending that goes on computing equipment, put at 78%, by the share of that compute which is imported, put at 80%, by the tariff rate. That produces an effective tax of 15.6% on building a data centre in the United States. The study then applies a standard multiplier for knock-on effects and converts the output gap into jobs, which is conventional methodology rather than anything exotic. The projection that will get quoted is the capacity one. About 20% of planned American AI data centre capacity between 2026 and 2030, roughly $450bn of capital expenditure, would be cancelled, pushed beyond 2030, or relocated abroad. Whether 80% of American data centre compute is imported is the assumption doing the most work in that chain. It is a defensible figure given where advanced packaging and assembly happen, but it is an estimate rather than a customs return. The CCIA is a technology industry trade association whose members include the companies that would pay the tariff, so the figure is an advocacy document with a methodology attached rather than a neutral assessment. The tension the policy is trying to resolve is real enough on its own terms. Tariffs are meant to pull semiconductor manufacturing onshore, and the fabs that would satisfy that goal take years to build, while the data centres are being built now. Taxing the imports in the meantime raises the cost of the thing the administration also says it wants, which is American AI infrastructure at scale. There is no version of the timeline in which domestic supply arrives before the buildout needs the chips, and Taiwan remains the epicentre of that supply regardless of what any proclamation says. A tariff offset programme for companies investing in American semiconductor manufacturing capacity was mentioned in the January proclamation, without details. Those details have not appeared either. Meanwhile, the competitive backdrop has not paused for the review. China has drafted a $295bn data centre plan built specifically to exclude Nvidia, which is a reminder that the cost of American capacity is not being decided in isolation.