Analysis

The CCP blocked Meta from acquiring Manus, an AI agent company that relocated from Beijing to Singapore ahead of Meta’s proposed acquisition. Blocking the acquisition demonstrates how the CCP’s priority of political control over commercial prosperity is seeping into global markets. Citing export control violations, the CCP decision effectively ends not only Meta’s acquisition; it also effectively severs all future US-China tech deals. 

Manus has catalyzed a change from “Singapore Washing” (新加坡洗白), whereby companies register in Singapore to disguise their ultimate national ownership, to “China Shedding” (去中国化), where companies flee the PRC to avoid CCP scrutiny and control. In some ways, these are two sides of the same coin. Chinese companies either wishing to escape the CCP’s clutches or to engage globally without the baggage of being a “Chinese” company, use locations like Singapore, Malaysia, or elsewhere to provide both regulatory protection and neutralize concerns about national origin. In other words, “Washing” is the mechanism by which Chinese companies achieve “Shedding.” 

Chinese talent, capital, and IP are themselves products of the PRC state-controlled system. If they can all relocate, then that is an issue of lost resources and a potentially devastating technology transfer to China’s adversaries. Tech innovation is crucial to China’s superpower ambitions, which means that the risk of losing tech (and expertise) is a direct challenge to the CCP’s legitimacy. For the CCP, preventing “shedding” is therefore a function of national security, with all of the legal implications that entails for Chinese corporations. 

The consequences of this decision will be far-reaching. Chinese startups will likely try to launch outside the PRC to preemptively “shed,” balancing the potential for lost preferential access to capital and Chinese research environments against future potential to sell the company to foreign bidders. To prevent this, as happened with Manus’s executives, the CCP will therefore rely on the levers it has to prevent PRC citizens from leaving the country, or, as with Shein, force them to relocate back to the PRC. The CCP’s willingness to compel PRC citizens to “return” (or be unable to leave) demonstrates how simply founding a business outside the PRC will be insufficient if the party demands “reshoring.” The CCP also might lean on other tech firms to “unwind” their overseas operations, although that would undo much of China’s economic clout overseas. In sum, the CCP’s view of “China Shedding” is unlikely to limit the CCP’s reach; if a company is founded by PRC citizens, regardless of where in the world the company is registered, then it is fair game for the CCP. 

Foreign companies, particularly in the tech sector, will likely see these moves and decide to accelerate decoupling from the PRC to avoid similar risks created by removing assets from the PRC. Foreign companies will similarly have to consider more carefully how to protect their PRC citizens employees, who, like Manus’s former employees, may face legal jeopardy from the CCP for their involvement in foreign tech innovation. Collectively, the fallout from Manus stands to not only sever the potential for outreach between Silicon Valley and China’s tech industry; more broadly it signals how the CCP’s grip over Chinese companies (and their employees) is intensifying amid the deeper geopolitical rivalries of the Sino-American Cold War.

Analysis

If the race for nuclear and space technology defined the Cold War, then the race for AI is defining the new Cold War. The fusing of tech and national security is accelerating worldwide: China’s blocking of Meta’s acquisition of Manus on export control grounds highlights one manifestation in the PRC; the introduction of multiple chip export control bills in the US Congress is another. Beyond the US and China, the European Union is similarly pushing its own tech platforms as insurance against US companies having overt leverage over European critical infrastructure. Collectively, tech innovation and AI in particular are moving away from the purview of private companies and into the realm of public-private partnerships.

With the rise of corporate solutions like Anthropic’s Mythos, the trend towards “AI nationalism”--that is, governments seeing AI innovation as “belonging” to a country or government, rather than a private company–will likely accelerate. Last week, the Pentagon announced that US AI companies would assist with technology innovation relating to national security. Such assistance will enhance the US government’s existing tech capabilities. This assistance will also further cement the international perception that these companies are American, rather than global companies. 

This could be good or bad, depending on context. For some countries, like Taiwan, this conflation between governments and companies is not a bad thing. In Taiwan, AI is existential not only because of its potential threat if weaponized, but also because Taiwanese chip manufacturing is so central to global AI capacity. Taiwan’s “Silicon Shield” (硅盾), whereby security guarantees from the United States are at least partially premised on maintaining access to vital chips, means that a continuing AI race enhances Taipei’s security. 

For US companies, however, close association with the US government akin to the Cold War will require a shift in thinking. As national security concerns seep into corporate decision making, the presence of PRC nationals in US companies will face greater scrutiny and will likely decline. Because PRC brain power has been so integral to US tech development, the reverse brain drain may make US companies less competitive and less able to attract top global talent. 

US tech companies may also face market access restrictions overseas as foreign governments question the pervasiveness of US government-aligned industries in their societies. The EU’s recent attempt to ban Meta’s WhatsApp from civil servant devices, for example, is likely the first domino in what could be a cascade of regulations that limit US tech companies’ access within certain markets. Indeed, sources in US tech already tell us that European employees and governments increasingly ask whether US tech is politically reliable, given the US administration’s hostility to Europe. While increasing ties with the US government is enticing, US tech companies therefore need to plan for the additional consequences that such arrangements entail.

Analysis

With a Trump-Xi meeting two weeks away, the US and China are both laying the groundwork for strong negotiating positions. The outcome of these talks has the potential to change tariff rates and export controls, redefine supply chains, and determine whether the current trend toward US-China decoupling will stabilize or accelerate 

For China’s part, new PRC government restrictions on companies that try to withdraw from Chinese markets in response to foreign pressure have already made clear the potential cost of trying to cut China out of supply chains. At the same time, Beijing’s decision to unwind the sale of Chinese AI company Manus to Meta has demonstrated the lengths to which China will go to secure its interests (see “From Zhongnanhai” above). China has also addressed overproduction and dumping accusations by making aggressive efforts to curtail the surplus in its solar industry. Beijing also removed tariffs from 53 African countries (the entire continent, barring Eswatini, which recognizes an independent Taiwan) in an effort to secure unfettered access to additional natural resources. In short, the CCP is making clear the cost of the US of taking a hard line on trade, while shoring up supply chains to insulate itself from potentially unsuccessful negotiations. 

The US is betting on its advantage in the production of advanced semiconductor and chip technologies to carry the day. After producing 25 times more advanced chips than China last year, it is clear that the Chinese drive for self-sufficiency has yet to close the gap with the US in critical sectors. While loopholes and bad actors have still enabled China to acquire US advanced tech, the proposed MATCH Act, if passed, will cement the US’s lead by denying China access to both US and our allies’ technology. 

The looming threat of new rules of origin that would exclude many Chinese-made goods from the US market could also further compel Beijing to give way to US demands, as overproduction and deflationary concerns in China would be exacerbated by the removal of China from more global supply chains. The US is therefore likely to offer increased but incremental access to US advanced technology for reciprocal gains in Chinese rare earths, which by now is a variation on a theme for negotiations.

While one cannot predict the outcome of these negotiations in advance, there is likely to be no significant change to the status quo. China’s desire to be seen as a stabilizing force, relative to a more impulsive US foreign policy, suggests they will wait for the US to make the first misstep  before making any significant demands. It also implies that this round of negotiations is as much about proving the US to be an unpredictable or unreliable actor as it is about securing Chinese access to advanced US technologies. This is also true because Washington’s and Beijing’s respective needs for rare earths and advanced tech make some level of compromise a nearly foregone conclusion. A US win in negotiations will therefore also largely be a matter of reputation manipulation, insomuch as it needs to make China look like it is unreasonably hoarding resources to spur the global community to rally around the US cause. 

For US companies, the performative outcomes of the negotiations will likely be less important than the expected trade compromises. Even a relatively neutral negotiation outcome could result in greater opportunities to sell in Chinese markets and increased access to raw materials, relaxing strained supply chains if only for a limited time. The likely continuation of trade barriers, including tariffs, export controls, etc., would still suggest an ultimate trajectory of decoupling, but the focus on stability will slow this process to a point that companies can manage future bifurcations of the global market into increasingly separate Chinese and American spheres.

Analysis

China’s 2026 Spring Festival–a celebration of the lunisolar new year and China’s most important holiday–had its usual suspects of happy dancers and fireworks. Among the festivities, however, the audience was treated to a synchronized dance of humanoid robots in a display of technological prowess from three state-backed manufacturers: Unitree, Galbot, and Magic Lab. More recently, a robot from Chinese smartphone brand Honor won the Beijing Half Marathon (while several other entrants tripped and/or shattered) in a marked improvement from last year’s showing. Collectively, China is publicly broadcasting its new obsession with robotics as not only a new, China-made tech product, but as a potentially revolutionary new “embodied AI” industry (具身智能) that China intends to dominate. 

Innovation in robotics promises multiple advantages to sectors like manufacturing. Ideally, and already in practice in some manufacturing sectors like automobiles, robots enable increased production at high quality, greater precision, and at a lower cost compared to humans. For countries with declining or aging populations, robotics also serves to buoy productivity in areas with shrinking labor forces. For China, in particular, robotics are therefore a potential much-needed economic boost to areas experiencing flatlining economic growth. 

Robotics are central to the CCP’s efforts to achieve “general artificial intelligence” and to create an “intelligent economy” ahead of the United States. These endeavors combine overlapping state and private sector entities into an “embodied artificial intelligence ecosystem” (具身智能生态系统‌), with Chinese AI software technology integrated into physical hardware that can interact with humans. This means that China’s ambitions for AI are not limited to software. They are intended as an extension of hardware, with Chinese-made products intended to dominate entire industries from manufacturing to transport to healthcare. Robotics’ integration is therefore as much a function of China’s national security as it is a manifestation of tech innovation prowess. 

This integration of CCP’s political ambitions into the robotics industry means that Chinese robotics manufacturers face a Faustian choice: to enjoy the benefits of massive state support, these firms must comply with the party’s demands. In part, this is because robotics also has a dual-use function for the Chinese military, whereby robotics innovation is one way in which the PLA intends to out-innovate and deter and defeat the United States. 

Regardless of why the CCP is intensifying its political pressure on robotics manufacturers, the result for Western companies is the same: engaging with Chinese robotics companies is now an even higher risk for facilitating CCP political and military goals. Chinese firms, driven by party demands, will be incentivized to even more aggressively pursue IP theft, talent poaching, and data exfiltration from US companies and consumers. 

Moreover, outside of the PRC, US adversaries are likely to be the first recipients of Chinese offers to wholly install complete robotics systems and to displace US tech companies from markets. Collectively, the surge in political interest in robotics in China, combined with the potential for Chinese intelligent systems to dominate global markets, means that China’s dancing robots may pose a greater threat to US economic and national security than many US companies realize.

Analysis

Chinese firms are now selling high resolution, sub-meter imagery to Iran. (Sub-meter imagery refers to satellite photos in which each pixel represents less than one square meter on the ground). This capability allows Iran’s military and Revolutionary Guard to track detailed movement of US forces and thus reduces the ability of the US military to act against Iranian forces and targets without being detected.

As with other frontier technologies, the CCP is making aerospace innovation an economic priority and seeking to dominate global markets. Chinese companies like Chang Guang Satellite Technology and Earth Eye now sell satellite systems and imagery access to foreign clients, including militaries. The reported $36 million sale of the TEE-01B satellite to Iran shows how a mid-tier power can acquire  imagery capability without developing its own program. 

Yet despite its growing and impressive capabilities, commercial imagery has limitations. Unlike the PRC’s military Synthetic Aperture Radar assets—which provide all-weather, 24-hour observation—most commercial satellites (Chinese or otherwise) rely on optical sensors that cannot see through clouds or smoke. Furthermore, dependence on ground stations for data transmission introduces latency and creates stationary physical targets and digital links susceptible to jamming or kinetic strikes. Finally, without the software, expertise, and experience to filter, integrate, and interpret raw imagery, its utility is limited. Still, China’s sale of intelligence capability to Iran is another example of how the CCP will use its so-called private sector to both generate profits and harm the US.