Chinese artificial-intelligence companies are establishing footholds in the United States, with startups like MiniMax and Moonshot announcing U.S. outposts. Their goal is to attract engineers who have worked on leading models at firms such as OpenAI and Anthropic, but they confront substantial obstacles related to compensation, reputation and geopolitical tension.
MiniMax, a Hong Kong-listed AI venture, lists roughly sixteen openings in San Francisco, including a lead researcher for its core model. The firm also seeks specialists in finance, law, life sciences, cybersecurity, and a senior recruiter to source talent across the Bay Area and Seattle, signaling a broad hiring push beyond pure research roles.
Compensation remains a decisive factor. MiniMax reported $84.3 million in employee remuneration for 2025, averaging about $203 000 per staff member, while Z.AI’s first-half-2026 costs were around $140 000 per employee. By contrast, OpenAI advertises base salaries between $250 000 and $445 000 for research engineers, and three Chinese engineers in the valley note that many peers receive total packages exceeding $1 million.
Experts say the move reflects a strategic shift. Cheryl Wu, an AI economics researcher at Yale and GovAI, told Rest of World that establishing research teams in Silicon Valley “makes perfect sense” for Chinese firms seeking talent and the latest advances. Data from a Carnegie China study show a net gain of 2 145 researchers for the United States in 2025 versus a loss of 1 729 for China.
Chinese firms are also courting Chinese nationals already at U.S. labs. Tencent recently hired former OpenAI researcher Yao Shunyu, and ByteDance’s frontier-model unit Seed reportedly offers 1.5 times the pay that OpenAI or Anthropic would provide. Wu Yonghui, a veteran from Google, joined Seed as head of research in early 2025, and the division now lists 71 positions across San Jose, San Diego and Seattle.
Despite higher offers, several Chinese engineers say they favor American companies for better work-life balance and the chance to work on cutting-edge models. Damien Ma, director of Carnegie China, explained that Silicon Valley is the only place where the demanding Chinese “996” schedule is not the norm, making it more attractive for top talent.
Washington has limited Chinese access to advanced Nvidia chips and scrutinises overseas data-center usage, yet it has not imposed comparable hiring bans. U.S. officials accuse Chinese labs of unauthorized model-distillation techniques, while the Chinese government has proposed export controls on its most advanced AI models trained on “important data.” The 2019 Huawei Futurewei layoffs after a trade blacklist illustrate how policy can disrupt cross-border research.
Not all Chinese AI firms pursue U.S. recruitment. DeepSeek continues to post roughly 150 engineering roles in China, focusing on building AI agents and backend infrastructure. Its founder Liang Wenfeng told a 2024 interview that the company mainly hires Chinese-educated staff and believes it can develop elite researchers internally, a view echoed by Yale researcher Wu, who doubts that hiring drives will close the talent gap with U.S. labs.