REPORT: China seizes its AI engineers — talent as a militarized state asset

The Bloomberg report of May 26, 2026 and its revelations

On May 26, 2026, Bloomberg revealed what many observers had anticipated but Beijing had not officially confirmed: China was imposing travel restrictions on its top engineers and researchers in artificial intelligence working at private companies, notably Alibaba and DeepSeek. These professionals — startup founders, senior researchers, chief technical officers — must now obtain formal approval from the relevant government authorities before any international travel.

The measure represents a major qualitative shift in Chinese government policy: until now, exit restrictions were reserved for civil servants, nuclear scientists, and executives of state-owned enterprises. Extending them to the private AI sector — including companies and individuals that developed within a nominally commercial framework — marks the moment when Beijing officially decides that its best minds in artificial intelligence are national assets, not free professionals.

The scope of the restrictions: Broader than the headlines suggest

According to Bloomberg's sources — who requested anonymity given the sensitivity of the subject — the restrictions do not apply based on professional title or hierarchical level, but on an assessment of the strategic importance of an individual's research to national technology ambitions. A mid-level engineer working on a cutting-edge language model can be subject to restrictions, while a CEO whose work is more administrative may not be.

The affected companies include giants like Alibaba and DeepSeek, but also other players in large language models, advanced robotics, and autonomous driving systems. In some documented cases, engineers have had their passports confiscated by their employer — a practice that creates a legal gray zone between company policy and state coercion.

What is happening here is fascinating and alarming at once: China is treating its AI engineers the way the USSR treated its nuclear physicists. It is an admission that artificial intelligence is now perceived in Beijing as a strategic weapon comparable to the atomic bomb — and that those who develop it can no longer be ordinary citizens with ordinary rights.

From informal advice to mandatory control

The trajectory of this policy is documented and reveals a deliberate escalation. In March 2025, according to the Wall Street Journal, Chinese authorities had begun to informally discourage AI founders and researchers from traveling to the United States — a guidance, not an obligation. In February 2026, DeepSeek founder Liang Wenfeng declined an invitation to the Paris AI Summit — a signal that the pressure was already real.

The Manus crisis in March 2026 served as a decisive accelerator. When the AI startup Manus attempted to relocate to Singapore and be acquired by Meta for $2 billion, Chinese authorities barred its two co-founders from leaving the country during the investigation into export control implications. This is the precedent that normalized the use of exit restrictions as an industrial and security policy tool in the private sector.

The June 1, 2026 regulation: Institutionalizing the restrictions

On June 1, 2026, the Chinese State Council published new rules strengthening control over outbound investment in key sectors including AI and national security-related technologies. These regulations formally allow the government to restrict the travel of Chinese engineers out of the country — a tool that Beijing had deployed ad hoc against Manus and that is now anchored in a formal legal framework.

The Foundation for Defense of Democracies (FDD) noted in an analysis dated June 3, 2026 that these new rules come with a broader arsenal: China can now cut commercial ties, revoke visas, cancel market access agreements, or impose sanctions on entities and individuals that "endanger the sovereignty, security, or development interests of China" — a deliberately vague definition that gives authorities considerable room for maneuver.

The rhetorical wave of "national security" as a universal justification is not a Chinese exclusivity — democracies use it too, sometimes excessively. But there is a crucial difference: in a democracy, a citizen can challenge a restriction on their fundamental rights before the courts. In China, that recourse does not really exist. National security is a final justification that tolerates no appeal.

DeepSeek: From global prodigy to untouchable national asset

DeepSeek created a worldwide sensation in January 2025 with the launch of its R1 model, which demonstrated performance comparable to the best American models at a fraction of the development cost. This breakthrough simultaneously impressed the tech world and alarmed Beijing: if a model that capable could be developed with limited resources, its creators represented an extraordinary strategic asset that no American competitor could be allowed to recruit.

Key engineers at DeepSeek were among the first to have their passports held by their employer — a practice presented as a company human resources policy, but which, according to Bloomberg's sources, applies under government pressure. International AI conferences — including NeurIPS, ICML, and ICLR — which represent the most important knowledge-sharing spaces in the field, will likely see a reduction in the presence of top Chinese researchers.

Alibaba and the Chinese private AI ecosystem

Alibaba, through its Tongyi Qianwen laboratory, has developed some of the most advanced Chinese language models. Alibaba researchers regularly appear in the most prestigious international academic publications. Travel restrictions do not cut them off from global research — they can still publish, collaborate online, participate in virtual conferences — but they limit direct recruitment and the physical transfer of knowledge that occurs in face-to-face meetings.

Other companies cited in analyses such as Moonshot AI, StepFun, and ByteDance face additional constraints: they must obtain government approval before accepting American capital, according to a Bloomberg report from April 2026. The combination of travel restrictions and inbound investment controls creates a wall around the Chinese AI ecosystem in both directions.

Here is what Beijing is trying to avoid: the Manus scenario. A brilliant Chinese startup moves to Singapore, gets acquired by Meta, and its technologies and brains pass under American control. From that perspective, the restrictions are coherent with a logic of economic security. The problem is that the best minds, if imprisoned in their country, eventually stop innovating freely — or find creative ways to leave anyway.

An inverted mirror of American export controls

China's policy of restricting AI engineers' travel is analytically symmetrical to American export controls on semiconductors. Since 2022, the United States has imposed growing restrictions on the export to China of advanced chips, semiconductor manufacturing equipment, and AI technologies. The American objective is to prevent China from accessing the components needed to develop advanced AI systems and weapons.

China's response is to protect what it already has — no longer the hardware components it cannot produce in sufficient quantities, but the human capital it has succeeded in training. This is the tech cold war viewed from its most human angle: individuals whose skills have become pieces of a geopolitical chessboard whose rules they do not control.

The appeal of the West and the brain drain risk

The travel restriction policy is also a response to a genuine concern of Beijing: for several years, Chinese AI engineers and researchers have been actively recruited by American companies like OpenAI, Google DeepMind, Anthropic, and Meta AI. The salaries offered are often several times higher than those in China, and research freedoms are generally greater.

The risk for Beijing is not merely brain drain in volume — it is the flight of the most decisive minds for its military AI ambitions. A researcher working on large language model architectures for general use can have direct implications for military targeting systems, surveillance systems, and autonomous weapons. This is why the selection of individuals subject to restrictions is based on the strategic value of their work rather than their hierarchical position.

There is a deep irony here: Beijing trained these brilliant engineers partly thanks to a partial international openness from the 1980s-2010s — Chinese students were welcome at the world's best universities, many did internships at American companies. Now that these minds are trained and productive, Beijing wants to keep them behind a border. This is a form of retroactive nationalization of human capital.

Science without borders: An ideal under strain

The global AI research community operates on an open-sharing model that has been extraordinarily productive: publications in open-access journals, code sharing, international conferences where researchers of all nationalities exchange freely. China has benefited greatly from this system — and contributed to it with high-quality publications that have advanced the field globally.

Travel restrictions threaten this ecosystem in a way that hardware export controls did not directly reach. If the best Chinese researchers can no longer physically attend major global academic conferences, the informal exchanges that often precede major collaborations will be compromised. The predictable result is a growing bifurcation of AI research between a Sino-aligned bloc and a Western bloc, with divergent architectures, standards, and development ethics.

Silicon Valley facing a shortage of specific talent

Paradoxically, Chinese restrictions could in the short term benefit American companies by reducing competition for certain technical profiles, but in the long run they could deprive Silicon Valley of an extraordinarily productive talent pool. Chinese engineers represent a significant fraction of teams at major American AI companies — many have become American citizens or permanent residents, and are not directly affected by Chinese restrictions.

What changes is the renewal flow: the young Chinese talents who would have done post-docs in the United States, or who would have applied at OpenAI after a few years at Alibaba, will now be retained or discouraged from taking that path. American startups specializing in recruiting Chinese AI talent have already reported a slowdown in their candidate pipelines.

The global technological bifurcation I observe being built is not merely an abstract geopolitical problem. It is a reality that will affect the speed of progress in fields like medicine, climate change, and education — areas where AI offers real promise. Fragmenting the global scientific community in the name of strategic rivalry is shooting ourselves in the collective foot.

Singapore, Dubai, San Francisco: Destinations of the tech exile

Despite passport restrictions and pressure on employers, Chinese AI engineers continue to leave China — sometimes via circuitous routes. Singapore has become the primary hub for this tech mobility: dozens of startups founded by Chinese engineers have been registered there since 2024, some with ghost offices in Beijing serving as cover for teams actually working from the island. Dubai and the United Arab Emirates play a similar role for the Sino-Arab technology community.

Silicon Valley remains the most coveted destination, but American visa restrictions for Chinese nationals in technology have made this path more difficult since 2023. This situation has created a geopolitical paradox: American entry restrictions and Chinese exit restrictions combine to produce a group of high-level technology talent in a geographic gray zone — neither fully in China nor fully integrated into the Western innovation economy.

The human cost of retention policies

Behind the technology mobility statistics lie concrete human trajectories. Engineers who accepted jobs abroad and can no longer honor their contracts without legal risk in China. Families separated when a spouse cannot obtain exit authorization. Researchers who must choose between their professional ambitions and their access to family and culture. These human realities are the invisible price of tech talent retention policies.

The accounts circulating in Chinese diaspora tech communities reveal a growing anxiety: even engineers not working in areas considered sensitive report difficulties renewing their passports or being questioned upon returning to China. The atmosphere of uncertainty is itself a tool of control — it discourages mobility even when it would be legally possible.

There is something profoundly sad about watching brilliant engineers — often trained at the world's top universities, having contributed to fundamental research benefiting all of humanity — treated as national resources to manage rather than human beings with legitimate aspirations. China is shooting itself in the foot in the long run: the minds you imprison eventually find other ways to leave, and those who remain in a growing control environment innovate less freely.

DeepSeek, Baidu, Alibaba: Navigating between Beijing and global markets

Major Chinese tech companies — DeepSeek, Baidu, Alibaba, ByteDance — find themselves in an increasingly uncomfortable position. On one side, Beijing imposes national security obligations including talent retention and cooperation with intelligence services. On the other, their international growth ambitions require an image of reliability and openness that clashes with these same policies. TikTok is the most visible example: its ownership by ByteDance has become an existential obstacle to its growth in democratic markets.

Travel restrictions imposed on engineers create concrete operational problems for these companies. Collaborating with international academic partners, presenting research at global conferences, recruiting foreign talent — all of this becomes harder when employees cannot travel freely. DeepSeek, which surprised the tech world with its performance in January 2025, operates in this contradictory environment: seeking global recognition while subject to domestic constraints that limit its international reach.

The international listing dilemma

Several Chinese tech companies were considering initial public offerings in New York or Hong Kong before the passport restriction policy complicated their situation. Western institutional investors now ask questions about governance and national security obligations before committing to Chinese companies. How do you value a company whose key human resources are subject to state-imposed mobility restrictions? How do you assess the regulatory risk of a company the government could potentially subject to data transfer requirements at any time?

These questions have no simple answers, and they are part of why the flow of foreign direct investment into China continued to decline in 2025 and 2026. The talent retention policy, presented as a national security measure, has real economic side effects that weaken the very companies it claims to protect.

Chinese tech companies find themselves in a position that Beijing's strategists had probably not fully anticipated: their national security policies harm the international competitiveness of their own tech champions. This is the fundamental contradiction of a state that simultaneously wants world-class international companies and total control over their human resources. These two objectives are structurally incompatible.

The CHIPS Act program and its limits

The American response to China's technological rise took the form of two principal instruments: the CHIPS and Science Act of 2022, which subsidizes semiconductor production on American soil, and export controls on advanced chips to China. These two instruments have real merits — they have effectively slowed Chinese access to cutting-edge semiconductor technologies. But they also have important limits: export controls create pressure on American allies who do not all share the same view of the Chinese threat, and CHIPS Act subsidies have complex redistributive effects on the global semiconductor industry.

The more fundamental question is whether a restriction strategy is sufficient or whether it must be complemented — or even replaced — by an attraction strategy. The United States has long been the global destination for the best technology talent precisely because it offered the freedom, resources, and innovation environment needed. Restricting visas for Chinese nationals in technology, as some political circles demand, risks cutting American access to a fraction of the world's best talent — exactly the opposite of what technological competition requires.

Europe as an alternative destination and a strategic stake

Europe can play a unique role in this competition for technology talent: it can be the destination that offers both the innovation environment, academic freedom, fundamental rights, and a cultural proximity that does not force Chinese engineers to choose between their heritage and their professional ambition. Countries like the Netherlands, Germany, and France have developed tech talent attraction programs that are beginning to bear fruit. Mistral AI in Paris has recruited researchers of several Asian nationalities.

But for Europe to become a genuine alternative, it must resolve its own deficits: still-heavy visa procedures, venture capital ecosystems less deep than in the United States, and language barriers that complicate integration in some countries. Europe has a unique window to attract Chinese tech talent who can no longer easily go to the United States and no longer want to stay in China. This window will not stay open indefinitely.

If the West is serious about winning the tech cold war, it must bet on its deepest comparative advantage: freedom. The freedom to innovate without ideological constraint, the freedom to collaborate without censorship, the freedom to live without permanent surveillance. These are the freedoms that made Silicon Valley what it is. Restricting visas, erecting walls, treating foreign researchers as potential spies — that is throwing this advantage in the trash.

When people become national resources

The policy of restricting AI engineers' travel illustrates a fundamental transition in the nature of contemporary geopolitical competition: it is no longer only armaments, territories, or natural resources that define national power — it is the concentrated cognitive capacity of a few thousand highly qualified individuals in critical technological fields. China has understood that in the AI war, losing its best minds to the United States could be as damaging as losing a province.

This logic is understandable from a short-term strategic perspective. It is profoundly problematic from a fundamental rights standpoint — the right to mobility, to choose one's workplace, to participate freely in a global professional community. Individuals treated as national assets lose their fundamental freedom of choice. This is a sign of the direction in which the Chinese regime intends to develop its technological state.

The lesson for democracies

The correct response of democracies to this policy is not to imitate it. It is to strengthen their attractiveness for global talent — by offering research conditions, academic freedoms, and professional prospects that neither China nor any authoritarian regime can match. If the West wants to win the AI race, it must remain the place where the best minds in the world want to live and work. This is its most durable comparative advantage — and it must protect it by being open, not by raising its own walls.

I will close with a question I pose to my readers as much as to myself: in a world where AI engineers have become soldiers of the tech cold war, which society do we prefer to build? One that treats its minds as national resources to control, or one that treats them as free human beings whose talent flourishes in freedom? I know my answer. And that answer is also a geopolitical argument.

By Maxime Marquette, columnist

Editorial stance

I am not a journalist but a columnist and analyst. My expertise lies in observing and analysing the geopolitical, economic and strategic dynamics that shape our world. My work is to take political strategies apart, to understand global economic movements, to put the decisions of international actors in context and to offer analytical perspectives on the changes that are redefining our societies.

I do not claim the cold objectivity of traditional journalism, which stops at reporting facts. I claim analytical clarity, rigorous interpretation and a deep understanding of the complex issues that concern us all. My role is to give meaning to the facts, to place them in their historical and strategic context, and to offer a critical reading of events.

Method and sources

This text keeps the basic distinction between verified facts and interpretive analysis. The rule never changes: a factual claim is published only when a verifiable source supports it, and the sources this article actually used are listed under Sources, never here.

Primary source families used by the publication, when the story has them: official statements from governments and international institutions, public statements by political leaders, reports from intergovernmental organisations, dispatches from recognised international news agencies.

Secondary source families: specialised publications, internationally recognised news media, analyses from established research institutions, reports from industry organisations.

When an article cites statistical, economic or geopolitical data, it comes from the institutions that produce it (intergovernmental organisations, central banks, national statistics offices), and the exact institution is named under Sources.

Nature of the analysis

The analyses, interpretations and perspectives in the analytical sections of this article are a critical, contextual synthesis based on the information available, the trends observed and the expert comments cited in the sources consulted.

My role is to interpret these facts, to set them in the context of today's geopolitical and economic dynamics, and to give them a coherent meaning within the larger story of the changes shaping our time. These analyses reflect expertise built through continuous observation of international affairs and an understanding of the strategic mechanisms that drive global actors.

This article describes a documented state of affairs on its publication date, not a prediction: later developments may shift its outlook. No update is promised in advance; when an article is corrected or completed, the change is dated in the text.