China Restricts Overseas Travel for Top AI Talent at Alibaba, DeepSeek

Liang Wenfeng, founder of DeepSeek, did not attend the AI summit in Paris this February. He declined the invitation, and at the time, it was attributed to scheduling conflicts. Now, Bloomberg has connected the dots: China is tightening controls on the overseas travel of top AI talent, requiring individuals deemed "strategically important" at private companies like Alibaba and DeepSeek to obtain government approval before leaving the country.

Not by Rank, but by What You Know

The logic behind this control is unique. It is not based on whether you are an executive or the size of the company, but on an assessment of your "strategic value" by relevant authorities. Researchers, founders, and technical leads—anyone whose knowledge is considered too sensitive to be taken abroad—can be placed on the list.

The timeline unfolds as follows:

  • March 2025: Authorities merely "advised" against traveling to the US, a soft approach.
  • December 2025: DeepSeek's parent company, High-Flyer, began withholding some employees' passports—note, this was at the company level, not a government ban.
  • May 2026: The policy hardened into a unified pre-approval requirement for all related companies, regardless of destination.

A landmark event occurred in March this year: Manus AI CEO Xiao Hong and chief scientist Ji Yichao were barred from leaving the country by the National Development and Reform Commission. This followed Meta's reported $2 billion acquisition attempt of Manus in December 2025, a deal that was blocked on April 27.

What Beijing Fears

Simply put, Beijing fears talent leaving with technology. These individuals are viewed as national security assets, and the core goal is to prevent expertise from flowing to competitors. The legal basis is Article 7 of China's 2017 National Intelligence Law, which requires all organizations and individuals to "support, assist, and cooperate" with national intelligence work.

Why now? Stanford's 2026 AI Index Report provides a telling figure: the performance gap between top US and Chinese AI models has narrowed to just 2.7%, down from 17.5 to 31.6 percentage points in 2023.

The smaller the gap, the more valuable the talent. With parity in sight, no one wants key individuals to slip away at this critical juncture. The funding gap remains, however: in 2025, US private AI investment totaled $285.9 billion, compared to China's $12.4 billion. Models are nearly catching up, but capital depth lags by an order of magnitude.

The Cost of This Move

In the short term, controlling passports stabilizes the talent pool. But AI thrives on international exchange—publishing papers, attending conferences, recruiting, and collaborating with overseas teams all depend on mobility. If researchers must report every time they travel abroad, their engagement with the global frontier will thin out.

China has built models that can rival those of the US, but now it chooses administrative measures to confine the people who built them within its borders. This is a difficult puzzle: preventing outflows may also block inflows. Whether it will actually drive talent away is a longer-term question that will take years to answer.

Sources: CocoLoop, China Limits Overseas Travel for AI Talent at DeepSeek, Alibaba, Private Firms (Bloomberg); China AI Travel Curbs Reach Alibaba, DeepSeek (TechTimes)