CHIPS Act leaves chipmakers facing choice between U.S. and China
Bill's funding comes with overseas investment restrictions attached
TAIPEI/PALO ALTO, U.S. -- Chipmakers are cheering Washington's long-awaited passage of a bill to provide funding for the U.S. semiconductor industry, but accepting those subsidies could tie their hands when it comes to future investments in China.
Is it a price that the likes of Intel, TSMC and Samsung will be willing to pay?
The CHIPS and Science Act, a $280 billion package passed by both houses of Congress last week, contains $52 billion in support for the semiconductor industry. But it also specifies that companies accepting federal subsidies will be restricted from making any "significant transaction" to materially expand their chipmaking capacity in China or any other foreign country of concern for 10 years.
While there appears to be some leeway for chipmakers to protect their existing business in China, analysts and lawyers told Nikkei Asia that the terms of the legislation create a minefield for companies and could, in effect, force them to choose between the U.S. and China.
Semiconductors are a key battleground between Washington and Beijing. They serve as the brains of electronic devices, from smartphones to laptops to data centers, and they also play a crucial role in sophisticated weapons systems, such as the Javelin antitank missiles the U.S. is supplying to Ukraine.
The CHIPS+ Act, as it is also known, contains exceptions that may allow chipmakers to continue investing in China if those investments are aimed at protecting existing and significant business interests in the country. But these exceptions only apply to expanding existing facilities, and only for "legacy semiconductors." Intel and Qualcomm, for example, are the dominant players in the market for premium core processors for Chinese computer and handset makers, respectively.
Legacy semiconductors include 28-nanometer or older generation chip technologies, according to Tan Albayrak, an export controls and sanctions lawyer with Reed Smith LLP.
Nanometer size refers to the line-width between transistors on a chip. Generally speaking, the smaller the number, the more advanced the chip. For example, iPhone 13 core processors use 5-nm nodes.
Albayrak said the act's exceptions for investments in China are aimed at avoiding sudden disruptions for U.S. business.
"This does not weaken the U.S. guardrail, but rather serves as a necessary balancing act in pursuing the policy goal of keeping a technological edge, all the while recognizing business realities and minimizing collateral damage to the U.S. sector," he said.
Legacy chips are used in a wide variety of electronics devices, from smartphones and computers to home appliances and connected cars. They are generally viewed as companion chips and are usually needed in much larger numbers than more cutting-edge processor chips.
Many global chip titans produce such chips in China. Taiwan Semiconductor Manufacturing Co. has a plant in Nanjing that currently makes 16-nm and 28-nm chips, while Samsung has a large production footprint for memory chips in Xi'an, in Shaanxi Province. SK Hynix has memory chip facilities in Wuxi and Dalian. Intel and Micron, both of the U.S., have chip packaging and testing facilities in China.
The CHIPS+ Act restrictions may not appear onerous at first glance -- few if any major chipmakers plan to make their most cutting-edge products in China, at least not in the foreseeable future.
TSMC and Intel had no specific comment on the China investment restrictions, though Intel said that in passing the CHIPS in a bipartisan vote, Congress had recognized that the U.S. is competing against other countries for technology leadership.
SK Hynix told Nikkei Asia: "Our company complies with regulations in countries where we do our business. We will respond flexibly to market circumstances in [our] investment and production plan."
Samsung declined to comment for this piece.
Nevertheless, lawyers warn that companies should take the investment restrictions seriously. "Penalties can include not only losing the funding, but also other penalties 'in the national interest,' which is quite broad language. And then, of course, there is the reputational impact of a company misusing these funds," said Clinton Yu, a Washington-based partner specializing in international trade and export control regulations at business law firm Barnes & Thornburg.
Another cause for concern, according to Martijn Rasser, a senior fellow and director at the Center for a New American Security, is that the restrictions could be a prelude to even more scrutiny of outbound investments by chipmakers.
"These restrictions underscore that U.S. policymakers are increasingly concerned about U.S. money supporting a buildup of capabilities in China," Rasser told Nikkei Asia. "These restrictions are an initial sketch of elements you would expect in an outbound investment review framework."
Publicly, chipmakers have welcomed passage of the act. A number have already committed to investing in the U.S. -- TSMC to the tune of at least $12 billion in Arizona, and Samsung $17 billion in Texas. SK Hynix last week announced a $15 billion investment plan for the U.S. American players Intel and Micron are also ramping up investment.
Several chipmakers warned that failure to pass the CHIPS Act would impact investment plans, and some, like TSMC founder Morris Chang, have criticized what they see as the small size of support being offered.
But subsidies aside, there is another reason chip companies may be keen to expand in the U.S.: China's own chip ambitions: Beijing has laid out plans to raise the country's share of homemade chips to 70% by 2025.
"In the long run, China's ambition is really to replace all the cutting-edge chips with domestically made chips. So there will be a question as to whether TSMC and Samsung can really continue to dominate high-end manufacturing in China," said Louis Lau, director of investment at Brandes Investment Partners. "I think the U.S. is more of a captive customer because the domestic manufacturers are very weak and they've long lost the edge in manufacturing."
China responded to the passage of the CHIPS act by saying it is firmly against provisions that restrict "normal sci-tech" collaboration between the two countries. But despite its strong words, few expect Beijing to retaliate immediately, given its reliance on foreign semiconductors.
"I think China is not in a position to sanction companies yet. Maybe after five, 10 years down the road ... if the tables have turned in a sense that Chinese companies are supplying most of the semiconductors for China, and they don't need to rely on TSMC and Samsung, then they will be in a position to sanction companies that take the U.S. government money," Lau said.
In the meantime, chipmakers face a more immediate question: In an industry where investments are often measured in billions of dollars, is $52 billion enough?
Only $39 billion of the CHIPS Act funding will go to the chip manufacturing sector, while $13 billion will be allocated to research and development, and innovation. However, building semiconductor factories is not only about chip facilities, it also involves a huge supply network: hundreds of chemicals and materials, as well as chipmaking equipment and consumable parts.
Consultancy Bain, meanwhile, estimates that increasing U.S. chip capacity by just 5% to 10% would require some $40 billion. Funding the next 10 years of development of new technologies would be even more expensive, costing some $110 billion.
"So $52 billion would certainly help fund both of these efforts, but not pay fully for either," said Peter Hanbury, a partner with the consultancy who specializes in the chip and manufacturing industries.
"Rebuilding the semiconductor ecosystem is an expensive proposition," he said.
Additional reporting by Kim Jaewon in Seoul