This month marks a year since Washington opened a fresh Section 301 investigation into whether China has kept the promises it made in the 2020 Phase One trade deal, among them commitments on intellectual property and forced technology transfer .For anyone who thought the IP fight had faded behind the bigger tariff story, it is a reminder that it never left. It just changed shape.

One of the sharper framings of that change came from Singapore’s former foreign minister. At Asia Society in New York in June 2024 , George Yeo was asked what America should do about Chinese behaviour it regards as deal-breaking. He answered with a story from the 1990s. As Singapore’s minister for information and the arts, he told Disney that Chinese pirates were, in effect, cultivating its future market. Today, he said, Disney collects full royalties there. Intellectual property, he concluded, is “a problem which resolves itself.”

Five years earlier, in a March 2019 South China Morning Post interview, Yeo had made a related argument about Huawei: worry that 5G networks expose a country to foreign intelligence is legitimate, but banning one supplier would not remove the vulnerability . The two remarks belong together, because they separate arguments Washington often merges.

On the first, the data now favour him. Applicants in China filed about 1.8 million patent applications worldwide in 2024, against roughly 502,000 from the US . In WIPO’s international PCT system, China led again in 2025 with 73,718 filings (up 5.3%) to America’s 52,617 (down 3.0%), and Huawei was the top PCT applicant for the ninth straight year. A country whose firms sell batteries, telecom gear and AI systems abroad has plenty to lose from copying. One caution: a filing is not a granted, enforceable patent, and volume is not value.

But “IP theft” was never one complaint. It bundles piracy and counterfeiting, trade-secret misappropriation, cyber-enabled theft, pressured technology transfer, and industrial policy that favours domestic champions. The USTR’s 2018 Section 301 findings tied these together , and Phase One tried to bind China on several at once, including a ban on forcing technology transfer as a condition of market access. Piracy, the 1990s problem Yeo remembers, has receded. The rest has not. Five weeks before Yeo’s New York talk, USTR’s four-year review concluded that China had not eliminated many of these practices and was still pursuing foreign technology through cyber intrusions.

Companies on the ground paint a mixed picture. AmCham China’s 2026 survey found 73% of respondents rated their IP-leakage and data-security risk in China as comparable to or lower than elsewhere, and 37% said enforcement had improved over the past year, up five points. Yet difficulty enforcing rights through the courts and inadequate legal protection remained the most prominent IP challenges . That describes a market maturing, not a problem solved.

For the tech industry, the stakes cut both ways. Strong IP protection funds R&D, makes licensing possible and lets startups monetise inventions; US IP-intensive industries account for about 44% of GDP . But protection has costs when stretched: patent thickets around standards, licensing fees that squeeze smaller firms, and a blurring line between IP enforcement and strategic control, as export controls and investment screens begin to resemble patent policy.

The distinction matters commercially. An IP dispute can end in damages or licences, which firms can price. A security dispute ends in entity lists and market exclusion, which firms can only route around by duplicating supply chains and splitting products, and eventually R&D, by region.

The blur is Yeo’s real subject. Intelligence gathering, he said, is a “spy versus spy game” every state plays, America probably more than any other. It will never end, but it can be managed, and it must not be allowed to drive policy. His test is an immune system: necessary, but if overwrought, it damages the body it protects. In his telling, blocking Huawei denied China a capability while also curtailing the West’s own access, leaving Western 5G behind China’s. That last claim is contestable, since Huawei was badly hurt in handsets and advanced chips, and 5G comparisons depend on the metric. But the cost of treating every technology question as a security question is real.

His reasoning reflects a certain kind of vantage point. A small, trade-dependent economy with no large domestic market and no bloc to shelter in has little to gain from assuming one nationality of supplier is safe, or that exclusion delivers security. Its prosperity rests on being trusted by everyone at once. That helps explain why Singapore maintains an IP regime widely regarded as among Asia’s strongest, to reassure foreign investors, while staying embedded in Asian supply chains because its economy demands it. Yeo’s 2019 answer on Huawei and his 2024 answer on IP share one instinct: name the specific vulnerability and engineer against it.

The critique cuts both ways. Small-state realism can slide into convenient equidistance, and calm is easier to counsel when it is not your trade secrets that left the building. State-directed acquisition of technology is a legitimate grievance, not an American invention, and Yeo’s “receding problem” describes piracy far better than cyber-enabled theft.

Still, the question he poses is the right one: what exactly is the vulnerability, who benefits from fixing it, and what does the fix cost everyone else? For an industry built on globally distributed ideas, capital and supply chains, it is worth asking before the next patent dispute is filed as a national-security case.