Moonshot AI's CEO turned down Apple and built Chinese rival to Anthropic

Yang's journey reflects brain drain that threatens US tech competitiveness

RYOTARO YAMADA August 11, 2026 23:52 JST

PALO ALTO, California -- Chinese startup Moonshot AI's CEO, Yang Zhilin, could have worked at any top tech companies in the U.S. after receiving a PhD from a leading American university. Instead he returned to China to found a company that eventually grew to rival Open AI and Anthropic.

Yang's story is a prime example of how a Chinese researcher receives an education in the U.S. and returns to China to create a formidable business that takes on U.S. counterparts.

Yang pursued his doctorate under Ruslan Salakhutdinov, a professor of machine learning at Carnegie Mellon University and an Apple alumnus. Around seven years ago, an executive from Apple who reported directly to CEO Tim Cook reached out to Salakhutdinov asking if Yang could come work at the company.

Born in 1992, Yang graduated from China's prestigious Tsinghua University before going to the U.S. to obtain his doctorate degree in a short four years.

As a student, Yang stood out for his research in artificial intelligence, receiving numerous job offers from Apple, Google and other U.S. tech companies. He could have easily stayed in the U.S. if he wanted to, according to .

"He saw excellent research directions, but he could also go and do like a very deep coding and optimization," said Salakhutdinov, adding that Yang could have joined a major tech company or pursued an academic career at a top university.

Yang did neither. In 2023, after returning to China, Yang cofounded Moonshot AI, a company that shocked the U.S. tech industry in July by releasing AI model Kimi K3, whose performance approaches that of the top models from Anthropic and OpenAI.

In March, Yang visited California's Silicon Valley and spoke at an AI conference hosted by Nvidia.

"Open models cannot be just open," said Yang. "They have also to be great." The release of Kimi K3 came soon after.

Seeing Yang's success, many U.S. tech investors lamented America's failure to convince him to stay in the country.

Josh Wolfe, cofounder of the venture capital firm Lux Capital, wrote on X that the U.S. needs to attract the best talents and keep them. Wolfe also said support should be provided for Chinese talent, such as allowing their families to join them.

Currently, the U.S. surpasses China in terms of the sheer volume of AI infrastructure, such as semiconductors and data centers. However, when it comes to the talent that is crucial to developing AI, the U.S. may be at a greater disadvantage than it realizes.

A report by Stanford University's Hoover Institution, analyzing the talent pool of Chinese AI startup DeepSeek, pointed out that not only is it becoming difficult for the U.S. to retain AI talent from China, but China is already succeeding in raising enough talent in the country.

A review of researchers whose names appeared in seven papers published by DeepSeek between 2024 and April 2026 revealed that one-third of the roughly 30 core researchers, and the majority of the approximately 270 researchers whose backgrounds were known, were individuals with no recorded experience at overseas research institutions.

"It shows a Chinese Al workforce that is growing fast and increasingly producing frontier-model contributors without routing them through American institutions during their careers as much as they once did," the report said. The findings suggested that "any U.S. strategy built on the assumption that China lacks the talent or know-how to surpass the U.S. on its own was already out of step with the evidence."

If Moonshot and DeepSeek succeed in going public in China or Hong Kong in the future, they would become a role model for entrepreneurs, bolstering the trend of Chinese AI talent aiming for domestic success rather than going to the U.S.

Examples beyond Yang are already springing up. In 2025, Yao Shunyu, previously a leading researcher at OpenAI, became the chief AI scientist at China's Tencent Holdings. In March this year, a former Google executive's move to the AI division of Alibaba Group was reported.

"We will have to compete in AI talent between China and US. There's definitely going to be competition," Salakhutdinov said. "Just maintaining and retaining the top talent is definitely something that U.S. will have to do. Five, maybe 10 years ago, that was not the case."