Chinese government agencies are now requiring select researchers, founders, and executives at private AI companies — including Alibaba Group and DeepSeek — to obtain prior approval before traveling abroad, according to a Bloomberg report published May 26. In many cases senior engineers are being asked to surrender their passports to their employers, an arrangement that functions as an informal exit ban with no judicial review.
What's actually changing
Travel controls on personnel tied to state labs and defense-adjacent research are not new in China. What's new is the reach into the private sector's frontier-model teams. Bloomberg reported comparable restrictions on some DeepSeek executives back in December 2025, and earlier reporting indicated DeepSeek staff began handing over passports around the R1 release in early 2025. The May 26 reporting extends the pattern to Alibaba and, per additional reporting, other private labs.
The official rationale is that this work could expose individuals to information classified as a state or commercial secret. Neither Alibaba nor DeepSeek commented publicly, and there was no immediate market reaction.
Part of a broader clampdown
The travel curbs land alongside capital controls. In late April, China's National Development and Reform Commission directed leading AI firms to reject US-origin investment without prior clearance — a move that already complicated secondary-market and venture flows into Chinese labs. Read together, Beijing is treating frontier AI talent and capital as strategic assets to be contained at home.
The timing tracks a closing capability gap. Stanford's 2026 AI Index put the distance between the best US and Chinese models at 2.7%, down from a 17.5–31.6 percentage-point spread in mid-2023. China now files roughly 69.7% of global AI patents and produces about 23.2% of AI publications, and installs industrial robots at nine times the US rate. Inbound AI-talent migration to the US has reportedly fallen sharply since 2017.
What it means for builders and enterprises
For anyone collaborating with or hiring from Chinese AI labs, the practical effects are immediate: fewer researchers at international conferences, harder in-person due diligence, and friction on the cross-border joint research that has historically been a strength of Chinese academic AI. Recruiters targeting senior Chinese talent should expect departures to get slower and more legally fraught.
For the labs themselves, the constraint cuts both ways. Containing talent reduces leakage risk but raises a brain-drain question: if top researchers read passport surrender as career-limiting, the firms most exposed to these rules may struggle to retain the very people who narrowed the gap. Watch researcher departures, publication output, and conference participation as the leading indicators of whether containment is working or backfiring.
