With less than a week to go until tariff “liberation day,” U.S. President Trump’s technology strategy might soon come into sharper focus. One potential victim of incoming tariffs is mature nodes manufactured in China. At 28 nanometers and above, mature nodes are relatively low-tech chips that nevertheless underpin a vast range of consumer and military technology.
While industry and political attention tends to center on the U.S.’s competitiveness in cutting-edge chip technology, the risk of overdependence on China for mature nodes has become a source of anxiety for American leaders. A portion of the 2022 CHIPS and Science Act was allocated to bolstering mature chip manufacturing in the U.S. In January of 2024, the Bureau of Industry and Security launched an investigation into the U.S.’s dependence on China for mature chips. The results of this investigation led former Commerce Secretary Gina Raimondo to conclude, “China is subsidizing those chips in these new fabs, dumping them into the global market and tanking the price. That isn’t fair.” The Office of the U.S. Trade Representative then initiated a Section 301 investigation into China’s policies and practices regarding mature chips in December 2024. That investigation is ongoing.
Countering China’s growing dominance in mature chip production is an incredibly tough issue to address because 1) the extent and nature of this threat are contested and 2) the tools the U.S. has at its disposal to constrain China and bolster the U.S. are faulty.
Paul Triolo, a technology expert and partner at DGA-Albright Stonebridge Group, argues that the threat of China dumping mature chips into the American market is overstated and misunderstood. First, “overcapacity” — a term often used to describe China’s substantial capability to produce mature chips — is an odd frame for a technology that is deployed at an incredible scale across many different industries (for example, the average car contains about 1,400 chips, 90% of which are mature nodes). Second, overproduction of around 15% to 20% in the chip industry is a good thing to deal with fluctuations in the market. Third, by 2030, Chinese foundries will only be able to satisfy 90% of domestic demand, which remains their focus. Fourth, since mature chips are already cheap, Triolo questions whether price cutting by Chinese manufacturers would really make that big of a difference for purchasers.
Still, “Chip War” author Chris Miller says we are already starting to see pricing pressure and risks of dependence on Chinese mature nodes. Speaking at an event in Taipei on March 25, Miller said, “I think there’s been a recognition in the U.S., in Taiwan, in Japan, and to some degree in Europe … that there [are] already major impacts.” The Bureau of Industry and Security Investigation launched by the administration of former U.S. president Joe Biden found that “72% [of chips] were cheaper at the PRC-based foundry than at a non-China alternative; the median price at PRC-based foundries was 10% lower.” U.S. companies also reported difficulty finding non-Chinese foundries that are willing to take their business, which underscores the relative lack of investment in manufacturing mature chips outside of China.
One reason Western chip manufacturers are not investing more in mature chips is because they expect that China will be able to outcompete them in the future by offering lower prices, Miller noted. This means that even if the risk of overdependence on Chinese mature nodes is overstated, merely the perception of risk could make overdependence a reality.
Matthew Turpin, who served in the National Security Council during Trump’s first administration, warns that if China takes hold of the mature node market, this could negatively impact the ability of other chip manufacturers to develop cutting edge technology. Even though the profit margin for mature chips is low, these chips are ubiquitous and can be manufactured at very high volumes. So as the cost of establishing a new fab is amortized over time, fabricating mature chips is a key revenue generator to fund R&D and upgrade the technology used to manufacture advanced nodes.
Whether the threat is genuine or exaggerated, there is not a clear path forward to constraining China’s dominance in mature chip technology because no tool currently available to the U.S. government is perfectly effective or has been tested at scale, according to Miller. Each tool has a deficiency.
Export controls
Export controls limit the export of sensitive technology from the U.S. The Biden administration took a “small yard, high fence” approach to chip controls, meaning that it focused on narrowly restricting the export of advanced chip and chipmaking technology. Miller is skeptical that export controls for mature chips would work because China may already possess the semiconductor manufacturing equipment to make this technology. In fact, Triolo argues that these controls actually accelerated China’s ability to produce mature chips by preventing Chinese foundries from focusing on advanced chip production.
Tariffs
The first Trump administration imposed 25% tariffs on mature chips imported from China in 2018, which the Biden administration raised to 50% in January. As Miller pointed out during his talk on March 25, the issue with using tariffs to restrict the trade in semiconductors is that the supply chain is incredibly complex, so the U.S. doesn’t actually import many chips directly from China. The U.S. does, however, import lots of goods from around the world that may contain chips originally manufactured in China. The Bureau of Industry and Security found that nearly half of the U.S. companies surveyed don’t know if the products they are importing contain Chinese chips.
Tariffs could be effective if they are applied broadly at the component level, according to Miller. Component tariffs target the chips themselves, as opposed to the product they are embedded in — this means that products manufactured outside of China but that contain Chinese chips would be subject to tariffs. In December of 2024, the Commerce Department said that the new Section 301 investigation might look into the viability of imposing a component tariff. But implementing this at scale would be unprecedented and an incredible logistical challenge. “Most manufacturers say it would be costly or maybe even impossible in the short run” to apply component tariffs to a component that is in every product by the dozens, hundreds or thousands, Miller said. He thinks that component tariffs are worth exploring but only over a long time horizon.
Nevertheless, given that Trump is the self-proclaimed “tariff man,” we might see him try to impose component tariffs at scale during his second term, possibly even on tariff “liberation day” on April 2. What exactly is in store for this day remains mostly unclear. Reports last weekend suggested that Trump wouldn’t be imposing blanket tariffs on industrial sectors, including semiconductors. But then on Monday Trump said that “We’ll be announcing some additional tariffs over the next few days, having to do with automobiles, cars and also to do a little with lumber down the road, lumber and chips.”
Import bans
The U.S. could ban the import of mature nodes manufactured in China, particularly for use in certain sensitive supply chains like drones or data centers. But Miller argues that these bans have limited efficacy unless there is a coalition of countries willing to implement them. (The point of a coalition would presumably be to constrain China’s semiconductor industry at a global scale and to prevent the reexport of Chinese chips to the U.S. via other countries.) Trump could pressure Europeans — who have made no movement on this front — into participating, but this would be tricky because Trump has not endeared himself to the continent and negotiating with the EU is notoriously difficult, Miller noted.
Even if one of these tools could work at scale, at the end of the day, constraining China’s ability to develop mature chips won’t fix the U.S.’s deficiency in this domain. “Fundamentally, the most significant factors impacting the health and viability of the U.S. semiconductor industry emanate from within the United States itself, rather than from developments in China,” wrote the Computer & Communications Industry Association in a letter to the U.S. Trade Representative last month.
Subsidies, like those given out through the CHIPS and Science Act, could help bolster American chip manufacturing. However, as Miller succinctly put it, “no one is going to win a subsidy race against China.” Not to mention that the political winds in the U.S. are shifting against subsidies as Trump derides the CHIPS Act and encourages Congress to repeal it.
One thing Miller is sure about is that the solution doesn’t lie in bringing manufacturing jobs back to the U.S. Job creation is good window dressing for Trump, but the future of chip manufacturing lies in automation.
Leave a Reply