Leadership Reshuffle at Tencent's Core AI Team Hints at Strategic Pivot
A significant personnel change has occurred within Tencent's Hunyuan large model team. Hu Han, formerly in charge of multimodal understanding, has departed to pursue entrepreneurial ventures. Tian Yonglong, a researcher with experience at OpenAI, will now lead the development of vision-language models.
Departure of a Key Figure: From Microsoft to Startup
Hu Han joined Tencent in early 2025, bringing with him substantial research experience from Microsoft Research Asia. During his tenure, he worked within the frontier research group of the large language model department, overseeing multimodal understanding and contributing to exploratory projects like world models. His move to found a startup reflects the ongoing circulation of top talent within the AI industry.
The Successor's Profile: Infusing OpenAI Expertise
The appointment of Tian Yonglong is notable for his background as a former OpenAI researcher. He will now helm VLM development, continuing to report to Yao Shunyu at Tencent. This move introduces a fresh, internationally-informed perspective to a critical competitive area for the company.
Future Team Focus: Shifting Toward Foundational Research
With Tian Yonglong taking over VLM duties, Hu Han's former team is expected to reorient its priorities. Sources suggest the group may increasingly concentrate on more forward-looking domains, such as world model research. This indicates Tencent's aim to balance ongoing VLM development with strategic investments in foundational AI capabilities.
Changes in core research leadership often signal subtle shifts in a company's technical strategy. Through this adjustment, Tencent is not only incorporating external expertise but also potentially strengthening its long-term positioning in generative AI's foundational layers.