As an unprecedented protest movement quickly spread to dozens of Chinese cities last fall, calling into question Chinese leader Xi Jinping’s authority, the country’s Orwellian surveillance system went into high gear to scrutinize video footage, track phone records, and identify those involved. Yet China’s success in neutralizing these and other threats to the regime rests not on technology developed in Beijing, but in Silicon Valley.

Protests in China are not particularly rare. The number of “public order disturbances” rose tenfold from about 8,700 per year in the 1990s to around 90,000 in the early 2000s, according to Chinese government statistics. Beijing stopped reporting protest data in 2006, although credible studies suggest such figures have hovered near historic highs, with one such study estimating nearly 180,000 protests in 2010 alone. But whereas most Chinese protests involve specific material issues, such as local pay disputes and environmental causes, last December’s demonstrations centered around overturning Xi’s harsh zero-COVID restrictions—with some protestors even calling for Xi’s ouster.

Thousands of people from all parts of society risked their lives to pressure the Chinese government into relaxing mandatory COVID testing requirements, quarantines, and stringent lockdowns. The public anger was understandable: Xi’s pandemic restrictions wrought havoc on China’s public finances and exacerbated many of the structural imbalances that have long plagued Chinese society. The result has been a sharp uptick in urban youth unemployment, a record $1.1 trillion Chinese budget deficit, and a 26 percent drop in land sale revenues—a key driver of local government spending. Spooked by the scale of the demonstrations and the intensity of citizens’ grievances, Xi did the unexpected: he hastily retreated and ordered the rollback of nearly all of China’s pandemic-related restrictions.

Seemingly overnight, China’s protest movement fizzled out, too—but not because Xi relented. Instead, within hours of the first demonstrations, Chinese authorities began knocking on protestors’ doors and demanding to know their whereabouts during the unrest. Many demonstrators received threatening text messages about their participation in “illegal riots,” whereas others were ordered to report to the nearest police station for questioning. Some simply vanished.

China’s successful protest crackdown was no fluke—it was by design. For three decades, China has been developing a complex web of high-tech surveillance systems with ominous names like “Sharp Eyes” and “Golden Shield.” Their first objective was to establish a unified network of spy cameras around the country, many featuring facial recognition software linked to China’s national registry database. These systems were later bolstered by corresponding cloud-based databases and geolocation platforms capable of analyzing terabytes of information generated hourly by Chinese traffic cameras and telecommunications networks, as well as data gleaned by China’s army of internet censors. Under Xi’s leadership, China nearly doubled its annual spending on “public safety” to $210 billion—a figure that exceeds the country’s defense budget.

The beneficiaries of Xi’s surveillance spending spree included Silicon Valley, which explains why U.S. software and equipment from companies such as IBM, Microsoft, and Oracle is deeply embedded throughout China’s police state. Beyond aiding in the development and maintenance of government-controlled systems used to squash last December’s protests, some U.S. firms also actively collaborate with Chinese companies that publicize their links to China’s surveillance apparatus and human rights atrocities.

A typical example of these problematic partnerships is U.S. semiconductor manufacturer Intel’s years-long relationship with Tiandy, a Chinese firm recently blacklisted by Washington for aiding Beijing’s persecution of Uighurs and other minorities in Xinjiang province. Tiandy produces surveillance cameras and networked surveillance systems equipped with facial recognition technology that Chinese authorities have deployed throughout China, including in cities where last month’s protests occurred. The company also sells torture devices known as “tiger chairs” used by Chinese officials to interrogate, sometimes for days, political dissidents. An investigation by tech watchdog IPVM further revealed Tiandy sold its surveillance systems to the sanctioned Iranian Revolutionary Guard Corps, possibly for use against peaceful protesters there.

Tiandy hardly hid its ties to China’s police state. In fact, the company championed such collaboration in its advertising materials and on its website. Tiandy CEO Dai Lin even hosted meetings with former Deputy Communist Party Secretary of Xinjiang Province Wang Junzheng, whom the U.S. government sanctioned for his role in orchestrating China’s Uighur genocide. And yet, Tiandy also maintained a top-tier alliance with Intel, whose processors power Tiandy’s genocide-enabling products. Even worse, IVPM alerted Intel about Tiandy’s activities in 2021 and Intel elected to maintain the partnership, only severing it when Tiandy was added to Washington’s export control list in December.

Intel is not alone, and its links to Tiandy go well beyond any due-diligence snafu. Instead, the company’s approach is broadly representative of Silicon Valley’s culture of compliance in China, where U.S. tech firms often outsource human rights and end-user risk assessments to third-party suppliers, many themselves linked to China’s military-industrial complex. In passing the buck, U.S. multinationals can have it both ways. They can claim to be vetting their supply chains, but also maintain some plausible deniability about what derogatory information may or may not have been shared with them by these third-party entities.

To be fair, most tech sales to Chinese entities are not prohibited by U.S. law, particularly because Washington often issues licenses explicitly permitting such transfers. But, the provision of any technology that optimizes the performance or sustains the day-to-day operations of Chinese entities involved in human rights violations is, at a minimum, ethically problematic. This moral dilemma is compounded further because many U.S. tech firms, including Intel, tout their commitment to environmental, social, and governance (ESG) initiatives even as they sell their products and services to autocratic regimes. At the same time, investment research firms, like Morningstar, have acknowledged how gaps in their ESG ratings systems can inadvertently give a free pass to companies linked to China’s censorship and surveillance apparatus.

For instance, it was Cisco that provided the routers and know-how necessary to construct China’s “Great Firewall”—used by Beijing to block Western websites and censor data—even though, since 1994, Chinese leaders have been transparent about using internet gateways to control China’s masses. As recently as 2017, Apple also helped China shore up the firewall by deleting five dozen apps that Chinese citizens could use to circumvent government internet filters.

Just as troubling, Public Security Bureaus in Beijing, as well as Fujian, Guangdong, and Henan provinces, all rely on IBM, Microsoft, and Oracle products to maintain their monitoring and censorship operations, according to public contract documents from as recently as 2020. These products enable, among other things, internet traffic monitoring, as well as facial and license plate recognition. Meanwhile, Microsoft’s Windows operating system is widely used by Chinese police departments, which the U.S. State Department has cited for their “routine, arbitrary, and unlawful interference” in Chinese citizens’ lives.

The U.S. tech sector’s links to the Uighur genocide are even more glaring. For instance, the Chinese entity responsible for constructing the Uighur detention facilities in Xinjiang, Zhongke Fuxing, counts IBM, Intel, Hewlett Packard, Microsoft, and Oracle among its trusted partners. For their part, Intel and Nvidia chip’s power the Urumqi Cloud Computing Center in Xinjiang, which, since its inception in 2015, has been cited by Chinese state media for its ties to China’s security apparatus. More specifically, this computer center analyzes peoples’ daily patterns to support China’s so-called predictive policing, in which anyone can be pre-emptively arrested for crimes they have not even committed.

In all, far too many U.S. companies, as well as well-respected U.S. universities like the Massachusetts Institute of Technology, have engaged in wide-ranging, unregulated, and often dual-use research with Chinese firms whose business models revolve around furthering repression and undermining freedom. Even when those Chinese companies have been cited by the U.S. or other governments as national security threats, some multinationals look the other way.

Case in point: Amazon Web Services—which maintains billion-dollar contracts with the U.S. Defense Department and intelligence community—hosts streaming services for Hikvision and Dahua, two surveillance firms blacklisted by Washington for supporting China’s Uighur genocide. Both companies are banned from selling their products in the United States. Hikvision, in particular, developed police tools to track protest activities in China, including system-wide alarms that are triggered when crowds “disrupt order in public places.” For its part, Amazon contends that these relationships are in “in full compliance with the law.”

Some U.S. tech giants, like Dell, have signaled plans to draw down their China-based operations amid concerns over Sino-U.S. tensions. But if Amazon’s legalistic retort is any indication, hard-hitting new laws and export control regulations will be needed to compel tech firms to decouple from China’s surveillance state. Indeed, it took an outright ban on the importation of Chinese goods made via forced labor for clothing manufacturers to finally get serious about mapping their distributor networks and shifting supply chains. Similar measures—in which the legal onus falls on tech companies to prove that their third-party vendors are not linked to Chinese law enforcement agencies—should be considered.

That said, the hope of better due diligence in the future will be cold comfort to Chinese citizens arrested for participating in last month’s peaceful protests, including demonstrators identified with the use of Silicon Valley’s latest wares. For America’s tech giants, that’s just business.