Tencent AI Team Major Restructuring

Author | Huang Yu

Editor | Zhou Zhiyu

The wave of AI is rapidly bringing the younger generation of entrepreneurs and scientists to the forefront.

Tech giants from the traditional internet era are generously extending olive branches to young people. At the end of the year, global tech giant Meta acquired Manus, founded by a post-90s Chinese entrepreneur; not long ago, Tencent also officially announced the appointment of 27-year-old former OpenAI researcher Yao Shunyu (Vincesyao) as Chief AI Scientist.

As the global AI competition enters deeper waters, Tencent is also signaling acceleration in AI, actively competing for top AI talent while restructuring its AI-related organizational framework to ensure further enhancement of team combat effectiveness.

According to information from Tencent insiders, Tencent's TEG (Technology Engineering Group) recently underwent organizational adjustments within its AI Lab. Due to personal development reasons, former Tencent AI Lab Deputy Director Yu Dong will be leaving Tencent soon.

In this round of AI competition, compared to Alibaba and ByteDance's aggressive strategies, Tencent has given the impression of being unhurried. Especially in the third quarter of this year, Tencent's capital expenditure was approximately 12.98 billion yuan, a year-on-year decrease of 24% and a quarter-on-quarter decline of about 32%, signaling caution in AI investments.

However, from Tencent's recent series of AI team reorganizations and talent acquisition initiatives, it appears that Tencent AI is quietly accelerating.

It is reported that Yu Dong joined Tencent in 2017, serving as an outstanding scientist and Deputy Director of Tencent AI Lab, as well as Chief Scientist of Tencent Youtu Lab. Yu Dong is an expert in speech processing and deep learning, and one of the leaders in successfully applying deep learning technology in the field of speech recognition for the first time.

During his time at Tencent, Yu Dong led research teams to publish hundreds of papers at various top academic conferences and journals, and promoted the application of NLP, speech, and digital human-related technologies in Tencent's business.

With Tencent pressing the acceleration button in the AI competition, frequent organizational and personnel changes are inevitable. The only goal is to ensure Tencent firmly stands at the table.

Since the beginning of this year, Tencent has made multiple organizational adjustments in AI-related businesses. At the beginning of the year, Tencent integrated AI products and applications such as Tencent Yuanbao, QQ Browser, Sogou Input Method, and ima into CSIG (Cloud and Smart Industry Group). By April, TEG established new departments for large language models and multimodal models.

On the same day that Yao Shunyu's joining was announced, Tencent also announced an upgrade to its large model research and development structure, with TEG establishing new AI Infra, AI Data, and Data Computing Platform departments to comprehensively strengthen its large model R&D system and core capabilities.

After this adjustment, the organizational structure of Tencent's AI team is now more clearly presented to the outside world.

Currently, Tencent's AI R&D is primarily centered around Yao Shunyu and Tencent Vice President Jiang Jie It is reported that Yao Shunyu has been appointed as the Chief AI Scientist of the "CEO/President's Office," reporting to Tencent President Liu Chiping; he will also concurrently serve as the head of the AI Infra Department and the Large Language Model Department, reporting to TEG President Lu Shan. Wang Di will continue to serve as the Deputy General Manager of the Large Language Model Department, reporting to Yao Shunyu.

This flattened reporting structure undoubtedly benefits Tencent in enhancing AI research and development efficiency, as well as unifying the feedback loop for model training and application development, accelerating AI product iteration.

In addition, Liu Yuhong will serve as the head of the AI Data Department, and Chen Peng will serve as the head of the Data Computing Platform Department, both reporting to Jiang Jie.

Jiang Jie joined Tencent as early as 2012 and, as the Vice President of Tencent's Corporate Development Group, is fully responsible for the technical management of Tencent's advertising platform products; he also concurrently serves as the Vice President of TEG and is the head of Tencent AI Lab.

From the main tasks of each department, the AI Infra Department, as an important part of Tencent's large model system, will be responsible for the technical capability construction of large model training and inference platforms, providing stable and efficient technical support and services for large model algorithm research and business scenario implementation.

The AI Data Department and the Data Computing Platform Department will be responsible for the construction of large model data and evaluation systems, as well as the construction of a data intelligent integration platform for big data and machine learning.

Insiders at Tencent revealed that this upgrade of the large model research and development architecture aims to further strengthen Tencent's engineering advantages while enhancing AI large model research capabilities, focusing on the company's AI strategic layout, and improving the research and development efficiency of AI large models.

The latest organizational structure involving Tencent AI Lab was established in 2016, focusing on fundamental research in artificial intelligence fields such as computer vision, speech recognition, natural language processing, and machine learning.

At that stage, AI was more viewed as a technological reserve rather than a plug-and-play productivity tool. The criteria for measuring achievements were often the number of papers at top conferences, rankings in competitions, and specific breakthroughs in particular scenarios.

With the wave of large AI models rising over the past three years, global competition has become increasingly fierce. The AI battlefield in China in 2025 has evolved from a battle of model parameters to a comprehensive game concerning capital efficiency, infrastructure, and traffic entry points.

Even if Yu Dong's choice to leave is a personal decision, it reflects a reality: AI talent is accelerating its movement, and the entire industry is undergoing a restructuring of technology, capital, and talent.

To emerge as a leader in this global AI competition, talent can be said to be the most important competitive advantage.

Tencent will inevitably need to attract more AI talent. On April 17, Tencent announced the launch of its largest employment plan in history, aiming to add 28,000 internship positions over three years and increase conversion to employment, with 10,000 campus recruitment interns expected in 2025, 60% of which will be open to technical talent.

Tencent stated that against the backdrop of accelerating the implementation of large models, it has increased recruitment efforts for technical positions in artificial intelligence, big data, cloud computing, game engines, digital content, and other fields, with an unprecedented expansion of technical positions.

On June 12, Tencent launched the "Qingyun Plan," aimed at globally recruiting top technical students, focusing on talent cultivation in ten cutting-edge technology fields Recently, there have been reports that Tencent is offering double salaries to top researchers in ByteDance's AI department (especially in the large model and multimodal teams) and has successfully recruited some key personnel. This undoubtedly reverses the previous trend of talent flow where ByteDance was "poaching" talent.

At this three-year mark, internet giants represented by Tencent, Alibaba, and ByteDance have become the focus of attention in the current AI arena, but so far, the three have chosen distinctly different strategic paths for AI development.

Alibaba has opted for heavy asset investment, accelerating growth in the B-end market share and fully launching C-end AI applications; ByteDance has chosen a strategy of breaking through traffic and feeding back applications to the underlying infrastructure. Recently, Alibaba and ByteDance have also become the focus due to their heavily promoted Qianwen APP and Doubao mobile assistant, respectively.

Unlike Alibaba and ByteDance, Tencent has maintained a consistent restraint in its AI strategy, preferring to integrate large model capabilities into its vast ecosystem, which has led to external perceptions of it "falling behind."

In the competition among Tencent, Alibaba, and ByteDance, whether Tencent can deliver stronger AI products next year will be key to reversing the situation