Introduction: Pony Ma has begun reallocating resources.

On Tencent's 28th birthday, an internal speech unexpectedly sparked outside speculation about layoffs.

On November 11, Tencent's management discussed organizational change during the 28th-anniversary Linktime online livestream, putting forward a striking idea: 50% of employees should operate like an army, holding down the existing core businesses; the other 50% should operate like a navy, embracing the AI wave, exploring AI Native, and seeking new room for development.

After the news spread, the outside world quickly interpreted it as Tencent preparing to lay off 50% of its staff, and claims of 8,000 layoffs even emerged. On November 12, Tencent Public Relations Director Zhang Jun publicly responded, saying the claims were a misunderstanding and a rumor. The so-called two 50%s refer to the company reallocating resources to adapt to the AI era, not a layoff plan.

But what is truly noteworthy about this controversy is not whether Tencent plans to lay off employees, but why this internet company—with more than 100,000 employees and a vast commercial empire built on WeChat, games, and advertising—has begun emphasizing the reallocation of organizational resources.

Over the past few years, Tencent has been talking about AI and has invested heavily in it. But from model research and development to product implementation, and then to organizational structure and staffing, AI's impact on a large internet company is clearly not just about launching a few more products.

The signal Pony Ma released this time is that Tencent wants to change not only its technical capabilities, but also the way the entire company works.

01 Re-examining organizational efficiency

This is not the first time Tencent has proposed organizational change.

In 2018, Tencent launched its third major organizational restructuring, known externally as the "930 Change." At that time, the mobile internet traffic dividend was gradually peaking, and the industrial internet began to become a new direction of development. Tencent reintegrated its original business groups and established the Cloud and Smart Industries Group (CSIG), hoping to move from consumer internet to industrial internet.

That adjustment was more about redrawing business boundaries.

This time, however, the focus of change has begun to shift to the internal organization.

According to information disclosed to the media by people close to Tencent, management clearly stated in the internal livestream on November 11 that this change is different from previous cost reduction and efficiency improvement efforts.

Past cost reduction and efficiency improvement efforts were more passive adjustments made by companies under operating pressure; this time it is a proactive change aimed at the AI era, with the focus shifting from reducing costs to creating new value.

Management hopes that frontline employees can proactively seek new opportunities, while middle management needs to become flatter and more agile.

Behind this is Tencent's rethinking of organizational efficiency at large internet companies.

The organizational structure of traditional internet companies is usually built around mature businesses. One department handles products, one handles R&D, one handles operations, and multiple management levels coordinate resources and pass down decisions.

This system was reasonable when business scale was continuously expanding.

But AI is changing the cost and speed of product development.

In the past, an internet product often required product managers, designers, engineers, testers, and operations teams to work together from conception to launch. As AI tools gradually mature, some work that previously required collaboration among many people can be completed by smaller teams.

When development efficiency improves, the original management levels and collaboration methods may become new costs.

In June this year, Tencent Senior Executive Vice President and CSIG CEO Dowson Tong said in a media interview that AI-native product R&D teams are becoming flatter, and team sizes may shrink further, with more R&D work completed by small teams of three to five people.

Liu Yi, head of CodeBuddy and WorkBuddy, also said that his team's management hierarchy has already been relatively streamlined, reducing the efficiency loss caused by information passing between different management levels.

This means that Tencent's emphasis on flattening middle management this time is not a sudden idea, but a signal that practices already adopted by some AI teams are extending to a broader scope.

For a giant organization like Tencent, the challenge brought by AI is no longer just the speed of technology R&D, but how to enable more employees to directly participate in product innovation and reduce communication costs within the organization.

02 The army holds the fort

The army proposed by Tencent first needs to hold down the existing core businesses.

This is not hard to understand.

Although AI has become the most important competitive direction in the internet industry, what truly enables Tencent to continuously generate revenue and cash flow is still WeChat, games, advertising, fintech, and enterprise services.

These businesses are not only the commercial foundation Tencent has built over more than two decades, but also an important source of funding for the company to continue investing in AI.

From the second-quarter 2026 financial report, Tencent's total revenue reached 204.8 billion yuan, up 11% year on year.

Among this, value-added services revenue was 98.4 billion yuan, up 8% year on year; marketing services revenue was 43.6 billion yuan, up 22% year on year; fintech and enterprise services revenue was 60.3 billion yuan, up 9% year on year.

Games remain an important source of revenue for Tencent. In the second quarter, domestic market game revenue was 47.3 billion yuan, up 17% year on year.

The advertising business performed even more prominently.

Tencent's second-quarter marketing services revenue grew 22% year on year, and the company attributed the growth to AI-driven optimization of advertising recommendation models, upgrades to intelligent delivery products, and the integration of closed-loop marketing capabilities within the WeChat ecosystem.

This actually already reflects the relationship between AI and traditional businesses.

AI does not necessarily have to appear in the form of an independent product; it can also directly enter the production processes of existing businesses.

For example, in the advertising business, AI can help systems more accurately identify user interests and improve advertising delivery efficiency; in the gaming business, AI can assist content production, game operations, and user services; in the WeChat ecosystem, AI has the opportunity to change the way users obtain information, handle affairs, and use services.

These changes may not immediately form new revenue items, but they may improve the operating efficiency of existing businesses.

Tencent disclosed in the second quarter that total user time spent on Video Accounts increased by more than 20% year on year, and advertising recommendation models and marketing capabilities within the WeChat ecosystem are also continuing to upgrade.

This means that Tencent does not need to abandon existing businesses in order to embrace AI.

On the contrary, how to embed AI into WeChat, games, and advertising may be more important than simply launching a new AI application.

This is also where the value of the army lies.

For Tencent, WeChat has a huge user base, games have a mature content and payment system, and the advertising business has relatively well-developed commercialization capabilities. These are resources that emerging AI companies will find difficult to replicate in the short term.

But the problem lies here as well.

Mature businesses usually have their own operating inertia. User scale, revenue growth, profit margin, and market share are all indicators that management has focused on for a long time.

But the changes brought by AI often require renewed trial and error.

If all teams work according to the assessment methods of the original businesses, it may be difficult for anyone to be willing to invest resources in exploring new products that have not yet generated revenue.

Therefore, what Tencent needs is not only to make existing businesses use AI, but also to change the way it evaluates businesses and employee value.

03 The navy sets out

If the army represents the commercial territory Tencent has already built, then the navy represents the new markets the company hopes to open up.

Since the beginning of this year, Tencent's layout in the AI field has clearly accelerated.

From foundation models to AI applications, and then to enterprise office tools, Tencent has gradually formed a business system based on the Hunyuan large model and jointly advanced by multiple application products.

On November 4, Tencent Hunyuan released the official version of Hy4.

Compared with previous versions, Hy4 added visual understanding capabilities and, for the first time, supports Computer Use, enabling it to perform computer operation tasks through graphical interfaces and command lines.

This means Tencent hopes that large models will no longer just answer questions, but be able to take on more practical work.

At the application level, Tencent is also advancing multiple products.

WorkBuddy targets office scenarios, CodeBuddy targets software development, Yuanbao targets individual users, and Xiaowei, the AI assistant within WeChat, seeks to further integrate AI capabilities into the WeChat ecosystem.

Among them, WorkBuddy is one of the products Tencent has focused on promoting recently.

Tencent disclosed on October 29 that since its launch in March this year, WorkBuddy has completed more than 9 billion task requests, covered more than 50 industries and more than 50% of the Fortune China 500, and has more than 200,000 Skills.

These data at least show that Tencent's AI products are no longer confined to the laboratory stage, but have begun to enter real office and enterprise business scenarios.

But it should be noted that the number of task requests and enterprise coverage do not equal actual paid revenue, nor do they mean the products have formed a stable profit model.

For Tencent, whether AI products can truly become new business growth points still needs to be tested by commercialization.

This is also the biggest difference between the navy and the army.

The army needs to do existing businesses well, and the assessment standards are relatively clear; the navy needs to enter markets that are not yet fully mature, find user needs, and explore product forms and charging methods.

For example, can WorkBuddy make enterprises willing to pay continuously? Can Yuanbao become an AI entry point that users use daily? Can Hunyuan find a balance between model capabilities and commercialization? And can the AI assistant within WeChat create new value without damaging the original product experience?

None of these questions has ready-made answers.

Tencent needs to give new businesses more room for trial and error, and also needs more flexible organizational forms.

This may be exactly why management proposed a 50% navy.

In the past, innovation at large internet companies often relied on resource allocation within business groups. A new project had to go through multiple layers of approval, and after the product launched it still had to face coordination among different departments.

But the competitive pace of AI-native products is clearly faster.

If a small team can use models and agent tools to complete work that previously required dozens of people to collaborate on, then the organizational advantage of large companies is no longer just headcount scale and financial strength.

The ability to make decisions quickly, iterate quickly, and find users quickly may become more important.

What Tencent may need is not to build another enormous AI business group, but to enable more small teams to have the ability to independently seek market opportunities.

04 An expensive AI bill

Behind the organizational change there is also a more realistic account: AI investment has begun to affect Tencent's financial performance.

In the second quarter of 2026, Tencent's capital expenditure reached 52.8 billion yuan, up 176% year on year.

This expenditure is closely related to the company's expanded computing power procurement and construction of AI infrastructure. Tencent stated in its financial report that the related investment is used to support the upgrade of the Hunyuan model, inference demand for WorkBuddy and CodeBuddy, WeChat AI initiatives, and the construction of AI capabilities for other products and services.

However, Tencent's capital expenditure also includes data centers, information technology equipment, software, and other infrastructure investment, so not all capital expenditure can simply be regarded as AI investment.

From the perspective of profit performance, investment in new AI products has also begun to show.

Tencent's second-quarter non-IFRS operating profit was 75.6 billion yuan, up 9% year on year. If the impact of new AI products is excluded, this indicator was 86.1 billion yuan, up 19% year on year.

The difference between the two measures is 10.5 billion yuan.

This shows that Tencent's current new AI products and related investment have already had a significant impact on operating profit for the period. But this does not mean that the entire 10.5 billion yuan is AI R&D expense, nor can it simply be understood as an independent loss of the AI business.

In terms of cash flow, Tencent's second-quarter free cash flow was negative 13.8 billion yuan.

The company also explained that operating cash flow for the quarter included large AI-related prepayments. If prepayments for computing power procurement are excluded, free cash flow was 37.6 billion yuan.

This set of data reflects that Tencent is investing substantial resources in AI infrastructure and new products.

But what the capital market really cares about is not how much Tencent is willing to invest, but when these investments can be converted into revenue and profit.

If AI can only boost the efficiency of existing businesses without generating new revenue streams, it functions more as a cost-optimization tool than a growth driver.

Only when AI can stand as independent products, attract new customers, and establish a sustainable charging model can it potentially become Tencent's growth engine for the next stage.

This is also why organizational adjustments and capital expenditure need to be assessed together.

The money has already been invested, and the models are constantly iterating. What comes next is the need for more teams to translate technical capabilities into commercial value.

From this perspective, the "army and navy" proposed by Tencent represent not simply a division of personnel, but an attempt to accomplish two things simultaneously within one company: on one hand, to improve the efficiency of mature businesses; on the other, to seek new growth space.

05 The common AI challenge facing big tech is laid bare by Tencent

The reason Tencent's internal speech on its 28th anniversary drew so much attention is that it addressed a problem that large internet companies commonly face in the AI era.

In the past, the core competitiveness of internet companies mainly came from user scale, traffic entry points, technology R&D, and commercialization capabilities.

But AI is changing the way these capabilities are formed.

On one hand, AI can improve the production efficiency of existing businesses, enabling mature companies to complete more work at lower cost; on the other hand, AI may also give rise to new product forms and change the way users interact with internet services.

For Tencent, these two changes are happening at the same time.

WeChat, games, and advertising businesses need to continue generating cash flow to provide the company with a stable operating foundation; AI products such as Hunyuan, Yuanbao, and WorkBuddy need to find new user needs and business models.

If it only holds the army, Tencent may miss the new markets brought by AI; if it puts all resources into the navy, it may weaken the competitiveness of its original businesses.

The real difficulty is how to balance the two.

This also means that the success or failure of Tencent's organizational change this time should not be measured by how many employees were redeployed or how many AI teams were established, but by two results.

The first result is whether AI can truly improve the efficiency of existing businesses. Can advertising delivery, game R&D, content production, and WeChat ecosystem services create more revenue or reduce unit operating costs because of AI?

The second result is whether new AI products can establish independent commercial value. Can models, agents, and office tools move from user growth to payment, and then from payment to sustained profitability?

Tencent will announce its third-quarter 2026 results on November 18. By then, the market can further observe the relationship among AI investment, enterprise services revenue, and overall profitability.

For Tencent, the organizational change proposed on its 28th anniversary may be just the beginning.

In the past, Tencent built its internet empire on WeChat and games. Now, it hopes to use AI to reorganize the company's resources.

Half to hold the existing commercial territory, half to seek new markets that have not yet been defined.

But for a large internet company, the real challenge has never been deciding where to put resources, but whether it can ultimately turn those inputs into new productivity and profit.