As the second half of the general large model's rapid deployment approaches, the R&D organizational structures of internet giants are undergoing an unprecedented "de-laboratory" transformation.

On March 20, an internal adjustment notice from Tencent caused a stir in the industry: the nearly ten-year-old Tencent AI Lab has been officially abolished, and its research capabilities will be integrated into the large model system as a whole. This change means that the AI Lab, once a group-level basic research institution, will bid farewell to the "independent exploration" era and fully serve the iteration and application of hunyuan large model.

R&D Path Focus: The Hunyuan Team Becomes the Only Central Hub

According to the latest organizational restructuring, some core personnel from the original AI Lab have been incorporated into the Large Language Model Department, reporting to Tencent's Chief AI Scientist Yao Shunyu. As a Princeton Ph.D. and former researcher at OpenAI, Yao Shunyu currently oversees both AI Infra (Infrastructure) and the Large Language Model, becoming the organizational core of Tencent's AI strategy.

This adjustment addresses the long-standing issue of scattered R&D within Tencent. Previously, Tencent's AI capabilities were spread across different business groups, resulting in high collaboration costs. Through this consolidation, Tencent will achieve a short-loop closed loop of research, engineering, and product feedback.

Hunyuan 3.0 is Stabilized? Full-scale Public Testing Will Begin in April

The direct result of the organizational restructuring points to the upcoming release of Hunyuan 3.0. In a recent performance briefing, Tencent executives revealed that Hunyuan 3.0 is currently undergoing internal business testing and is expected to be officially launched in April. The new version will achieve significant breakthroughs in reasoning capabilities and Agent (intelligent entity) construction.

Consensus Among Big Companies: Weaken Laboratories, Strengthen the Large Platform

Tencent's move is not an isolated case; major domestic internet companies are all undergoing similar structural reshaping:

ByteDance: The original AI Lab system has been integrated into the Seed team, pursuing a frequent iteration model similar to OpenAI.

Alibaba: The Tongyi system has recently been incorporated into the Token business group, emphasizing the direct linkage between model size and business growth.

This trend reflects the hard logic of the large model era: model training has become a massive systemic project, and a single-direction research breakthrough is no longer sufficient to independently translate into a competitive advantage. Only by deeply integrating computing power, data, engineering, and real product feedback can one maintain their position in the ever-changing AI competition.

For Tencent, abolishing the AI Lab does not mean abandoning basic research, but rather directly putting the "spark of the laboratory" into the "industrial engine" of the large model. With Tencent's AI investment plan set to double by 2026, this campaign led by Yao Shunyu and with everyone focusing on the Hunyuan main line is entering its most critical delivery phase.