"Everything We're Competing On"

At three in the afternoon on Monday, August 14, 2017, in the Diplomatic Reception Room of the White House, a sixty-nine-year-old trade lawyer named Robert Lighthizer stood a step behind President Donald Trump and watched him sign a single sheet of paper. Treasury Secretary Steven Mnuchin and Commerce Secretary Wilbur Ross flanked the desk. Tom Kennedy, the Raytheon chairman, would later be handed one of the signing pens. The cameras were tight on Trump. The sheet was a presidential memorandum directing the United States Trade Representative to consider opening an investigation under Section 301 of the Trade Act of 1974 into China’s policies on technology transfer, intellectual property, and innovation. “Ambassador Lighthizer,” the president said, looking up from the paper, “you are empowered to consider all available options at your disposal.” It was the most important sentence anyone had spoken in the chip industry’s politics in a decade, and almost no one outside the room noticed.

The reason no one noticed was Charlottesville. The white-supremacist rally that had killed Heather Heyer on Saturday had detonated through the news cycle, and Trump’s first statement condemning hate “on many sides” had drawn a furious bipartisan reaction. He had arrived at the China signing late, after a tense prepared statement that condemned the Klan and neo-Nazis by name. Cable networks split-screened the trade signing in a small inset while running Charlottesville coverage as the main story. Wire services compressed the day into a single line. The president had ordered an IP investigation. The next sentence was about Charlottesville. The story moved on. What had actually happened was that the United States had, for the first time in the post-Cold-War era, formally opened a unilateral case against the country it had spent two decades integrating into the world trading system. The architect of the case was the same man who had spent the early 1980s building the unilateral case against Japan.

Lighthizer had grown up in Ashtabula, Ohio, the son of a doctor in a steel and chemical town that spent his youth shedding jobs to imports. He had worked for Bob Dole on the Finance Committee in the late 1970s and arrived at USTR as a deputy in 1983, when the Reagan administration was negotiating “voluntary” steel restraints, “voluntary” auto restraints, and the early outlines of what would become the 1986 U.S.-Japan Semiconductor Trade Agreement. He had spent that period sitting across from Japanese vice ministers who told him, in flawless English, that their market was open and that the absence of foreign chips inside it was a quality problem. He had not believed them then. He did not believe their successors in Beijing now. The thirty years between government tours he had spent representing American steel companies in antidumping cases, increasingly convinced that the assumption underlying American trade policy since the early 1990s, that markets would discipline mercantilists if Washington kept the door open, had been a slow national disaster. Trump’s transition team had been sent a 2008 New York Times op-ed in which he had made that case. Trump had read it. Lighthizer was confirmed in May 2017 by 82 to 14 and given a small office at the Winder Building four blocks from the White House, the same building from which he had run his Reagan-era Japan files.

The administration around him was not unified. Mnuchin, a former Goldman partner, was trying to settle the trade dispute before it spiraled. Gary Cohn, the National Economic Council director and another Goldman alumnus, was openly skeptical of tariffs and would resign over them in March 2018. Ross leaned hardline but was an unreliable executor. McMaster was building a foreign policy team that took the rise of China seriously without quite knowing what to do about it. Steve Bannon was running an explicit “economic war” framing through Breitbart and said so to The American Prospect in an August 2017 interview that helped get him fired. And then there was Peter Navarro, a tenured economist from UC Irvine who had spent the late 2000s writing books with titles like Death by China: Confronting the Dragon, citing throughout an analyst named Ron Vara whose name was an anagram of his own and whose views were Navarro’s. Jared Kushner had found him by searching Amazon for books critical of China. He had no operational power in the White House but he had an unimpeded line into the Oval Office. When Mnuchin and Cohn talked Trump out of a tariff, Navarro could walk in afterward and talk him back into it. Lighthizer used Navarro’s intermittent victories to keep the formal investigation alive while he built it.

Inside the National Security Council, the actual intellectual work on the China brief was being done by Matt Pottinger. A former Wall Street Journal China correspondent, he had been roughed up by state security agents in a Beijing Starbucks in 2005, walked away from journalism for the Marines, served in Iraq and Afghanistan as an intelligence officer, and arrived on the NSC in early 2017 as senior director for Asia. He had drafted, with McMaster, the section of the December 2017 National Security Strategy that called China a “strategic competitor” and a “revisionist power.” In the prose of national security documents, that was the equivalent of regime change. The strategy Trump signed on December 18, 2017, declared that “China seeks to displace the United States in the Indo-Pacific region, expand the reaches of its state-driven economic model, and reorder the region in its favor.” The shift was less a policy than a permission slip. It told the agencies that the interagency caveats of the previous twenty years, about engagement, about not provoking Beijing, about preserving cooperation on transnational issues, no longer applied.

The intellectual cover for the shift had appeared in print earlier that year, in a Foreign Affairs essay by Kurt Campbell and Ely Ratner titled “The China Reckoning.” Campbell had been Hillary Clinton’s assistant secretary of state for East Asia, the architect of Obama’s “pivot to Asia.” Ratner had been Joe Biden’s deputy national security adviser. Their essay argued that nearly half a century of American China policy had rested on a faulty premise: that economic engagement would push China toward political openness and a market economy. Neither had happened. Engagement had failed. The piece was not a Trump talking point and the authors were Democrats, but it served as a structural endorsement, from the heart of the policy establishment, of the diagnosis until then associated with Lighthizer, Navarro, and Bannon. By the spring of 2018, the question of whether engagement had worked produced the same answer at Cato, Brookings, the Council on Foreign Relations, and Heritage. It had not. The next question, what to do, was the open one.

The Section 301 investigation, formally launched by USTR on August 18, 2017, was Lighthizer’s answer. Section 301 had been the same statute under which Reagan-era USTR Clayton Yeutter had pried open the Japanese semiconductor market in 1986. The provision authorized the U.S. government to take “appropriate action” against trade practices that were “unjustifiable” or “unreasonable and discriminatory” and that burdened American commerce. It had not been used in a major way against any country since the 1990s, partly because the WTO’s dispute settlement system had been understood to supersede it. Lighthizer did not accept that understanding. He had spent his entire career arguing that multilateral institutions were a tool for the United States, not a constraint on it, and that the WTO had become the latter. He used Section 301, with eyes open, to rebuild the legal architecture of unilateral American trade action.

For seven months his office took submissions, held public hearings, and built a case file. The final report, delivered on March 22, 2018, ran to 215 pages and was the most detailed indictment of any country’s trade practices the U.S. government had ever produced in peacetime. It documented what nearly every Western multinational operating in China had complained about privately for two decades and almost none had been willing to attach their name to in public. China used foreign-ownership restrictions and joint-venture requirements to compel American firms to transfer technology. It used a discriminatory licensing regime that imposed worse terms on foreign technology than on domestic. It used state-directed acquisitions of foreign firms, often through opaque ownership chains, to acquire technologies the Chinese economy could not generate organically. And it sponsored, through the Ministry of State Security and elements of the PLA, a sustained campaign of cyber theft directed at American industrial information. The report named names. It cited cases. It tied each finding to Chinese industrial policy, above all the “Made in China 2025” plan under which the State Council had set targets to dominate ten advanced industries, semiconductors among them, by the middle of the next decade.

The same day as the report’s release, Trump signed a second memorandum directing Lighthizer to draw up a list of Chinese imports for tariffing. “Technology,” Lighthizer told the assembled cameras, “is really the backbone of the future of the American economy. And, frankly, the future of the world economy. This is the future. This is what we want to have for our children and our grandchildren.” The framing was deliberate. Section 301 was not, in Lighthizer’s hands, about widgets and trade deficits. It was about who would own the industries that the rest of the twenty-first century would be built on, and the case rested on an argument that the United States could not let the country it now considered a strategic competitor write the rules of those industries.

The first tariff list landed on July 6, 2018. It targeted 818 product categories worth roughly $34 billion in annual trade and applied a 25 percent ad valorem duty. USTR’s accompanying notice said the list had been chosen to focus on “industrially significant technologies” benefiting from Chinese industrial policy: aerospace, information and communications technology, robotics, industrial machinery. Consumer goods with visible price effects on American shoppers had been excluded. Beijing retaliated within hours with 25 percent tariffs on $34 billion of American imports, focused on agriculture. Trump, undeterred, ordered List 2, $16 billion at 25 percent, in August; List 3, $200 billion at 10 percent rising to 25, in September; and List 4 a year later, $250 billion more. By the end of 2019, average effective duties on Chinese imports had risen from around 3 percent to roughly 21 percent. The tariffs drew the headlines. The more consequential half was the part the headlines mostly missed.

That part had begun, in earnest, the month before the first tariffs landed. On April 16, 2018, the Bureau of Industry and Security at the Commerce Department, a small agency that almost no one in the press covered, reactivated a previously suspended denial order against ZTE, the Shenzhen-based telecommunications equipment maker that was China’s second-largest networking firm. ZTE had pleaded guilty in March 2017 to violating American sanctions on Iran and North Korea and had agreed to a $1.19 billion penalty and a probation regime. BIS now alleged that ZTE had lied about disciplining the executives responsible. The denial order banned every American firm from selling any product to ZTE for seven years. Within forty-eight hours, ZTE’s manufacturing lines, dependent on Qualcomm modems, Intel processors, Microsoft software, and dozens of smaller American components, came to a stop. The company suspended trading on the Hong Kong stock exchange and announced that “major operating activities” had ceased. A flagship Chinese company, ranked among the country’s national champions, had been killed in a week by a single Commerce Department order.

The denial was eventually relaxed. Trump, after a call from Xi Jinping in May, ordered BIS to find a face-saving exit. A settlement with a further $1 billion penalty, $400 million in escrow, and a wholesale replacement of ZTE’s board allowed the company to resume operations in July. Lighthizer’s wing of the administration was furious, seeing it as Trump giving away leverage. But the technical demonstration was unmistakable. The Korean and Taiwanese fabs, the Dutch lithography monopoly, the EDA tools made by three California companies, the dependence of every Chinese cell phone on at least a dozen American chip designs: all of it had become, in the space of one denial order, an architecture of leverage that any future American president could pull on at will. The chokepoints worked.

The legal infrastructure to operate those chokepoints at scale was being built simultaneously, in a piece of legislation almost no one outside the trade bar had heard of. The Export Control Reform Act of 2018, signed on August 13 as part of the John S. McCain National Defense Authorization Act, was the first substantial overhaul of American export-control law since 1979. It made permanent the Commerce Department’s authority to maintain the Export Administration Regulations. More importantly, in Section 1758, it directed the executive branch to identify and impose controls on a category the law called “emerging and foundational technologies” essential to American national security. The list was open-ended. The criteria were vague. The point was that the United States could now place export controls on any technology a Commerce Department interagency review decided was strategically important, including technologies that did not yet exist when the law was written. It was a license, drafted carefully and passed quietly, for a future generation of officials to weaponize the global supply chain in ways the architects of that supply chain had never contemplated.

Companion legislation, the Foreign Investment Risk Review Modernization Act, was signed into law the same day. FIRRMA expanded the Committee on Foreign Investment in the United States, the interagency panel that reviewed foreign acquisitions of American companies, into a much broader screen for investments in critical technologies and infrastructure. CFIUS had already blocked, the previous September, a Chinese-backed acquisition of Lattice Semiconductor, an Oregon FPGA maker. In March 2018 Trump had blocked, on CFIUS recommendation, Singapore-based Broadcom’s hostile bid for Qualcomm, a deal between two non-Chinese parties the panel had nonetheless judged would weaken American leadership in 5G. FIRRMA codified the practice and extended it. After August 2018, every foreign investor who wanted to put a check into an American chip or AI company knew that an interagency body in Washington could and would say no.

The political ratification came on October 4, 2018, when Vice President Mike Pence walked onto the stage at the Hudson Institute and delivered a speech that, in retrospect, looked like the doctrinal anchor of the whole shift. It had been personally edited by Trump and blessed by Pottinger and McMaster’s successor, John Bolton. Pence’s argument was that China was now the principal challenge to the United States across every dimension America had built its postwar primacy on. Beijing was conducting influence operations in American politics, persecuting Christians and Uighurs, militarizing the South China Sea, and engaged in “the wholesale theft of American technology.” Through Made in China 2025, the Communist Party “has set its sights on controlling 90 percent of the world’s most advanced industries.” And, in the line the China-watching community would quote for the next five years, Beijing was attempting “to win the commanding heights of the 21st century economy.” The underlying contest, the speech argued, was about who would design and build the technologies on which everything else in the next century depended.

The pivot was bipartisan even where the tactics were not. Senator Mark Warner, the Virginia Democrat who vice-chaired the Senate Intelligence Committee, had spent 2017 and 2018 leading hearings on Chinese telecommunications equipment and was among the first congressional voices to publicly call Huawei a national security threat. FBI Director Christopher Wray, in February 2018 testimony, described Chinese state intelligence as “a whole-of-society threat,” a phrase he would later regret for its overreach but that captured the change in framing. Chuck Schumer cheered Trump’s tariffs, breaking with most of his caucus, who were uncomfortable with the legal mechanics but agreed with the diagnosis. By the 2018 midterm campaigns, the political question on China was no longer whether to compete but how. The era of engagement had ended, not as a single decision but as a slow change of mind across two political parties whose foreign-policy elites had spent a generation insisting on the opposite.

What had not yet happened, by the end of 2018, was the application of these tools to the chip industry directly. The Section 301 tariffs hit chip-adjacent products but spared the bare semiconductor wafers and packages American firms imported from Taiwan and Korea, because doing so would have raised costs for the U.S. companies the policy was meant to protect. ECRA’s emerging-and-foundational list had not yet been populated. CFIUS was still operating case by case. The export controls that would, two years later, choke off Huawei’s access to TSMC, and the rules that would, four years later, deny Beijing the most advanced lithography on the planet, were still in the future. What 2017 and 2018 had done was build the legal scaffolding inside which those later moves became thinkable. The hard part, in Washington’s experience, was always the framing. Once Beijing was a strategic competitor and the chip industry was the commanding heights, the rest was implementation.

Lighthizer himself, in his later memoir No Trade Is Free, compressed the period into chapters that spent more time on NAFTA than on the Section 301 file. But for him the throughline was obvious. The names across the table had been Japanese; now they were Chinese. The product had been DRAM; now it was 5G base stations and graphics chips. Washington’s instinct, in 1985 and again in 2018, had been to treat the dispute as a problem about a particular industry rather than about whether the United States would remain capable of making the things it depended on. The instinct had been wrong both times. The country on the other side of the table was not, this time, an ally. It was building, in parallel with its trade case, an army.

The shift between the summer of 2017 and the fall of 2018 had been a shift not in the tools, which had existed for decades, but in the willingness to use them. The next decade of chip policy in both capitals would unfold inside the perimeter Trump had set on a Charlottesville-shadowed afternoon, when he told Lighthizer he was empowered to consider all available options at his disposal. The options had turned out to be more than anyone in the room had imagined. The era of free trade in semiconductors was ending. The era in which a chip would be treated, by the country that had invented it, as something closer to a weapon was beginning.