Former OpenAI Researcher Joins Tencent, Escalating US-China AI Competition
China's Tencent has lured Yao Shunyu from OpenAI, which is not only a strategic move at the corporate level, but also an important step for China in the global AI sovereignty competition.
Chinese tech giant Tencent recently reported that it has successfully recruited top AI researcher Yao Shunyu from the American artificial intelligence company OpenAI, which has drawn widespread attention in the industry. This is one of the most closely watched cross-border talent movements against the backdrop of rising US-China tech tensions, highlighting Chinese companies' determination to accelerate their layout in the global AI talent race.
Bloomberg cited sources who said that Yao Shunyu has joined Tencent, based in Shenzhen, to integrate artificial intelligence technology into its existing service system. The source said that Tencent offered Yao Shunyu a compensation package worth up to RMB 100 million, although specific conditions and details were not disclosed.
Yao Shunyu previously worked at OpenAI and had significant research experience at Google and Princeton University. He graduated from Tsinghua University in China and earned a Ph.D. in computer science from Princeton University. His expertise lies in language agents and cognitive architectures, and he has written or co-written about ten academic papers to date.
In response to external attention, Tencent did not officially confirm Yao Shunyu's employment, but published an article on its WeChat official account to deny some media reports, without specifying which part of the information was inaccurate. OpenAI and Yao Shunyu himself have not commented on this.
As artificial intelligence technology accelerates, global tech companies are fiercely competing for top AI talent. According to reports, US companies such as Meta have offered signing bonuses of up to $100 million to try to attract talent from competitors. Yao Shunyu's move to a Chinese company symbolizes that despite increasing restrictions imposed by the US on China's tech industry, Chinese companies still have appeal.
OpenAI CEO Altman recently told the media in San Francisco that the US may have underestimated the complexity and seriousness of China's progress in the field of artificial intelligence, and stated that export controls alone may not be a reliable solution.
He pointed out that the US-China AI competition is no longer just about who is ahead, but involves a multi-layered deep contest, including reasoning ability, research capabilities, and productization capabilities. "I don't think it's just a question of who runs faster."
Altman also expressed doubt about Washington's policy of using export controls to hinder China's development of AI. "You can restrict the export of a certain technology, but it may not be the key one... people may build their own chip factories or find other alternatives," he warned.
Aside from the talent battlefield, the competition between the US and China over chip independence is also intensifying. According to the US tech media "The Information", Chinese tech giants Alibaba and Baidu have started to train AI models with self-developed chips. Alibaba has been deploying its "Zhenwu" processor for small models since early 2025, while Baidu has tested its upgraded ERNIE model with the "Kunlun P800".
This trend is seen as a specific manifestation of China's strategy to promote "technological self-reliance and controllability." In the face of continued tightening of US export restrictions under the Trump administration, Chinese companies are gradually reducing their reliance on American chips such as those from Nvidia.
Under the trend of US-China tech decoupling, many countries around the world are accelerating their "AI sovereignty" strategies, aiming to reduce dependence and gain control over data, computing power, and technological leadership.
South Korea is focusing on developing local AI infrastructure with core domestic tech groups such as Samsung and LG; India is focusing on developing large language models (LLMs) suitable for local languages to serve practical needs in agriculture, healthcare, and education.
Original: www.toutiao.com/article/1843125087416332/
Statement: The article represents the views of the author.