Tencent's Former OpenAI Researcher Publicly Declares AGI Ambition as China Recruits From U.S. Labs

Tencent Chief AI Scientist Yao Shunyu said Friday he wants to build a “long-term AGI organization” in China, CNBC reported from a Tencent event in Beijing co-organized with local authorities. Yao left OpenAI last year and was appointed to his current role in December, according to Bloomberg. Tencent also restructured its AI operations around Yao’s appointment, as the South China Morning Post reported at the time.

The public declaration is the news. Chinese AI companies have historically focused on applications over foundational research. Baidu CEO Robin Li previously predicted AGI was at least a decade away. Yao’s remarks signal Tencent is adopting the U.S. frontier lab playbook: recruit researchers from OpenAI, Anthropic, and Google DeepMind, then redirect their expertise toward AGI-scale ambitions.

The Agent Angle

“I don’t think ChatGPT or Claude will be the only super-app,” Yao told the Beijing audience, according to CNBC. He described untapped potential in the “trillions of dollars” and said the path forward in China runs through smaller models with more consistent performance on basic tasks.

That framing matters for agent builders. Tencent operates WeChat, the most widely used messaging platform in China with over a billion monthly active users. An AGI-oriented research lab feeding into WeChat’s agent infrastructure could produce consumer agent deployments at a scale no Western platform currently matches. Smaller, reliable models optimized for task completion align with how agents run in production: consistent execution matters more than benchmark scores.

Timing and Contrast

The timing is pointed. Anthropic warned Thursday that frontier models are nearing recursive self-improvement and called for an industry slowdown. OpenAI published a federal governance blueprint this week proposing institutional monitoring of that same capability threshold.

Yao’s response, delivered at a government-affiliated event in Beijing: accelerate. The divergence between U.S. labs calling for governance guardrails and a U.S.-trained researcher in China calling for AGI organizations is the clearest signal yet that AI safety timelines are being set by the fastest mover, not the most cautious.