China Poaches U.S. AI Talent as Tencent Eyes AGI Race
CNBC reported Friday that China AI talent recruitment from U.S. firms is intensifying. The clearest signal: Tencent’s newly appointed chief AI scientist is a former OpenAI researcher now openly pursuing artificial general intelligence on Chinese soil.
Tencent’s AGI Ambition Takes Center Stage
Tencent Chief AI Scientist Yao Shunyu joined the Chinese tech giant within the past year after departing OpenAI. Speaking at a Tencent-hosted event in Beijing on Friday, he laid out an ambitious agenda. He told the audience his personal mission is to build a long-term AGI organisation within China.
Yao argued that existing platforms like ChatGPT and Claude will not be the only dominant AI applications. He described untapped market potential in the “trillions of dollars.” His preferred roadmap for China involves leaner AI models with more reliable performance on everyday tasks, rather than racing to scale frontier systems.
The Beijing event, co-organised with local authorities, drew a senior city official to deliver opening remarks. It underscored how closely China’s government is aligned with its tech sector’s AI push.
Background: A Diverging Race
For years, U.S. labs including OpenAI, Anthropic and Alphabet’s DeepMind treated AGI as the central goal. China’s AI ecosystem, constrained by U.S. chip export controls, largely focused on applied uses. Those ranged from factory automation to consumer devices.
Baidu CEO Robin Li once projected AGI was at least a decade away. That cautious posture is now giving way as Chinese firms absorb talent carrying a more aggressive U.S.-style vision.
Other high-profile moves reinforce the pattern. Alibaba reportedly brought on a Google DeepMind researcher to support its Qwen model programme. ByteDance hired a Google DeepMind vice president to lead research at its Seed division. Moonshot AI, behind the Kimi model, was founded by a former Meta AI and Google Brain engineer.
Caution Rising in the U.S. as China Leans In
The timing is notable. Anthropic on Thursday warned publicly that cutting-edge models are approaching a threshold where they could improve themselves without human oversight. The company called for a deliberate slowdown in frontier model development. Critics have questioned whether those safety warnings serve competitive as much as ethical purposes.
Meanwhile, uncertainty around U.S. immigration policy has given Chinese researchers additional reason to return home. Beijing has also pledged heavier investment in foundational research over its next five-year cycle, making the talent pull stronger.
The net effect is a convergence. China’s AI agenda is beginning to look less like an application layer built on Western breakthroughs and more like an independent race toward the same long-horizon goals.
Read Next: OpenAI, Anthropic and the Battle Over Who Controls the AI Safety Narrative