中国AI企业赴硅谷招募顶尖人才面临薪酬与前沿模型吸引力双重挑战

Chinese AI companies are hiring Silicon Valley-based engineers to work on their models, potentially tapping into the talent at American labs like OpenAI and Anthropic.

MiniMax, the Hong Kong-listed Chinese AI startup, currently has around 16 San Francisco-based job openings listed on its website, including for a research lead focused on developing its core AI model. The company is also recruiting specialists in finance, law, life sciences, and cybersecurity, as well as a lead recruiter to attract top AI talent in Silicon Valley and Seattle.

Moonshot AI, the Chinese company behind the Kimi AI model, is looking for U.S.-based  “solutions architects,” who will “help enterprise customers integrate its AI models into their businesses.”

American AI companies, with their deep pockets, have attracted global talent, including many Chinese engineers, by offering high compensation and opportunities to work on the world’s best models. In comparison, Chinese companies have so far largely relied on homegrown talent. Hiring in the U.S. reflects a strategic shift by Chinese companies as they look to compete and stay close to one of their most lucrative markets, experts said.

“Wanting to establish research teams in Silicon Valley makes perfect sense for Chinese AI companies seeking access to talent and the latest developments in the industry,” Cheryl Wu, a researcher in the economics of AI at Yale University and global think tank GovAI, told Rest of World.

Good Work and $$$

More than half of the world’s top AI researchers complete their undergraduate studies in China, a September study by research center Carnegie China said, based on the educational backgrounds of people who published papers at NeurIPS, one of the world’s leading AI conferences.

Yet the U.S. remains a global magnet for AI talent, the study concluded.

In 2025, the U.S. saw a net gain of 2,145 researchers while China recorded a net loss of 1,729, according to the analysis. 

China’s AI companies have been trying to lure top Chinese researchers at American AI labs to return home. One high-profile example is Tencent, which recruited former OpenAI researcher Yao Shunyu to lead its AI research.

Chinese companies are seeking “little geniuses” who have worked at companies like OpenAI and Anthropic, a headhunter recruiting for Chinese labs in the Bay Area told Rest of World, requesting anonymity as she is not authorized to speak to the media.

These attempts, however, have not had much success because Chinese companies often struggle to match the salaries paid by their Silicon Valley peers.

“It’s money. Chinese companies simply don’t have enough of it to offer competitive wages for talent,” said Wu, who published a report examining Chinese AI companies’ financing with research institute Epoch AI. “They have only recently begun raising more substantial funding and still lag far behind their U.S. competitors.”

There is a stark gap between what Chinese and American AI companies pay employees.

MiniMax spent $84.3 million on employee remuneration in 2025, equivalent to roughly $203,000 per person, based on its year-end head count of 415, according to its financial filings. The annualized remuneration costs of Z.AI, another Hong Kong-listed Chinese AI lab, in the first half of 2026 amounted to $140,000 per employee. Neither company disclosed a breakdown of pay for engineers or researchers, and both figures include share-based compensation and staff across all departments.

In comparison, OpenAI advertises annual base salaries of $250,000 to $445,000 for research engineers in San Francisco, before equity awards. Three Chinese researchers working in Silicon Valley say it’s common for their peers to get an annual compensation package worth more than $1 million.

As of September, China’s six leading AI firms collectively generated only about 10% of the combined AI-related revenue of OpenAI and Anthropic, according to Wu’s research.

“Career prospects are another factor. Even senior Meta engineers might accept a lower-ranking position to join Anthropic because of its work on frontier AI,” Wu said. “That makes it even harder to persuade top talent to join a lab whose models are just not at the cutting edge.”

Chinese tech companies are now ramping up their salary game. If an AI researcher had offers from OpenAI or Anthropic, ByteDance’s frontier model division Seed would try to lure them with a competing offer that’s 1.5 times higher, according to a person familiar with the team.

Wu Yonghui, a veteran AI researcher who worked at Google for 17 years, joined Seed as the head of research in early 2025. Seed currently lists 71 positions in San Jose, San Diego, and Seattle. 

But even with matching pay, three Chinese researchers working in the U.S. told Rest of World they preferred American companies for better work-life balance and the chance to work on frontier models. 

“I think most of the top Chinese talent would feel pretty comfortable in Silicon Valley; it’s the only other place where “996” isn’t really frowned upon,” Damien Ma, director of Carnegie China, told Rest of World, referring to the demanding work culture common in China’s tech industry, where employees work from 9 a.m. to 9 p.m., six days a week.

Extension of the U.S.-China AI war

Washington has spent years restricting Chinese companies’ access to advanced Nvidia chips, and is scrutinizing their use of overseas data centers and cloud services. But no comparable restrictions prevent Chinese AI companies from establishing U.S. operations and hiring engineers there. 

Both Washington and Beijing are increasingly seeking to protect their most advanced AI technologies from each other.

The Chinese government proposed export controls on some of the country’s most advanced AI models, particularly those trained using the country’s “important data,” despite specifications.

U.S. officials and American AI companies, meanwhile, have accused Chinese AI labs of using unauthorized distillation techniques to replicate their models’ capabilities. On other sensitive tech fronts, like the semiconductor industry, the U.S. has already restricted some American engineers from working in Chinese facilities without a license.

While neither country has imposed a blanket ban on AI companies hiring researchers across borders, the growing competition over advanced technology could make talent recruitment another flashpoint.

“Given it is already so competitive, it doesn’t take much to imagine U.S. companies at some point capitalizing on U.S.-China competition and making national security arguments that will make it more difficult for Chinese labs to operate or to attract the best talent. Conversely, it’d be a similar dynamic if U.S. labs were operating in China,” Ma said. “This elite talent is by definition scarce; we are talking [about] like the top 5%–10% of people. Who wouldn’t want to hoard the best and the brightest for themselves?”

Chinese tech companies have previously faced U.S. restrictions that disrupted their American research operations. In 2019, telecom giant Huawei cut more than 600 jobs at its U.S. research arm, Futurewei, after Washington placed the company on a trade blacklist.

Not all Chinese AI companies want to hire in the U.S. China’s AI poster child, DeepSeek, continues to advertise research and engineering positions in China. It recently announced about 150 positions for engineers who would expand AI agentic and back-end infrastructure. In a 2024 interview, DeepSeek founder Liang Wenfeng said he hired mostly Chinese-educated people and believed the company itself could cultivate elite AI researchers.

Yale researcher Wu is skeptical that Chinese startups could attract enough top researchers to compete with America’s leading AI labs. “They may have plans to hire, but that doesn’t necessarily mean they’ll be able to find the people,” she said.