ByteDance's AI Strategy Shifts: Doubao Becomes Central to Company Development

Embodied AI has become the most prominent race in robotics in 2026. At the WAIC exhibition last month, demonstrations of humanoid robots dancing and fighting gave way to displays of robots operating on real production lines, performing assembly, sorting, and parts picking tasks on full-scale factory replicas.

Investors closely monitored this area: in the first half of 2026, over 300 deals in embodied AI and related supply chains in China exceeded 90 billion yuan. Furthermore, the number of employees involved in embodied AI roles increased fourfold in just the first four months.

However, beneath all this activity lies a common problem—data. Models with limited computational power and algorithms with limited capabilities are now facing the fact that data is what restricts their development potential. Robots cannot automatically absorb accumulated internet content, such as text, images, and videos. They require a different array of information: an understanding of friction between materials, forces generated when handling different weights, and the causal relationship between grasping and an object—all the subtle physics that websites do not provide.

According to industry experts, the global supply of high-quality manipulation data amounts to only a few hundred thousand hours, whereas supporting a general-purpose robot likely requires tens of millions of hours.

Teleoperation allows for the collection of high-precision data, but it is a costly process. First-person videography scales cheaply, but it lacks detail regarding joints, forces, and trajectories. Pure simulation, on the other hand, fails when transferred from the virtual environment to the real world. Liu Shenxiang, founder and CEO of WUWENAI, who previously led autonomous driving data and testing projects at Baidu, views these methods not as alternatives, but as a unified pipeline.

The system includes multimodal real-world data collection complexes for continuous acquisition of physical interaction data. Concurrently, a generative world model expands scenarios and edge cases in simulation, and both elements are continuously used for training and further data generation. WUWENAI calls this a 'closed loop for merging virtual and real data,' built upon three components: the Data Factory, the World Model, and the World Simulator, which enables large-scale training, evaluation, and feedback.

Liu believes the company goes beyond simply being a data provisioning platform. Once a high-precision world simulator can reliably reproduce real physics, a robot will be able to perform an AlphaGo-level maneuver, conduct millions of inexpensive trial-and-error cycles in simulation, and transition from imitation learning to reinforcement learning. WUWENAI began supplying this Real2Sim2Real stack several months before SceniX was acquired by World Labs.

Liu emphasizes the parallel as a signal to the industry: world models only become infrastructure when integrated with real-world data, robotic simulation, scaled training, and real-world feedback. Video and 3D generation alone are insufficient.

The company's bet is structural, not product-based. Embodied AI models are limited not so much by the number of parameters as by the volume of data they consume, and the data industry is transforming from simple raw data collection to the creation of orchestrated infrastructure. The next test for the Chinese startup WUWENAI is whether it can become a world-class company in physical AI infrastructure amidst giants like World Labs, Tesla with Optimus, and Google DeepMind, which are developing their own stacks in parallel. The road is being paved regardless, and WUWENAI has decided to take on the role of the asphalt paver.

This website employs a specialized security service designed to protect against malicious bots. This page is displayed during the verification process to ensure that the user is a human and not an automated bot.

Security Check in Progress

The website's security service functions to prevent attacks from malicious bots. The display of this page is part of a procedure that confirms the visitor is not a bot.

Three major Chinese tech giants—Alibaba, Tencent, and ByteDance—have simultaneously focused their efforts on one strategic goal: the desktop office AI agent. According to a report by the analytics firm Analysys for the second quarter of 2026, covering 17 main desktop AI agents, the total number of visits in June 2026 reached 60.62 million. Of this, the three leaders—Alibaba, Tencent, and ByteDance—collectively accounted for 56.22 million of these visits, leaving only 4.4 million for all other competitors.

AI Agent Market Dynamics

This data confirms that the platform economy applies to office AI agents, just as it did with browsers and search engines: users tend to install one primary desktop agent, which forms a natural oligopolistic dynamic.

Restructuring Company Portfolios

In July 2026, all three companies restructured their AI agent portfolios. Alibaba merged three existing products—QoderWork (for local desktop operations), Wukong (for corporate security and permissions), and MuleRun (for complex workflow templates)—into a single product tentatively named Qianwen Office. This project is led by Chen Yusen, the new CEO of DingTalk, a technical founder born in 1992 whose startup Chaitin Technology was acquired by Alibaba Cloud.

Tencent reportedly moved its QClaw product center team to the Cloud Products Department 6 and integrated it with WorkBuddy under unified management. ByteDance, meanwhile, transformed Doubao into its flagship office agent brand.

Prospects and Player Advantages

The Chinese AI agent market is projected to grow from 80.4 billion yuan in 2025 to 696.8 billion by 2030. The most convertible segment consists of single-employee companies requiring AI assistance as an assistant, accountant, or designer. Nearly 60% of users are willing to pay because increased office productivity provides a direct return on investment: saved time translates into earned money.

Alibaba's strategy under Chen Yusen involves integrating three layers: QoderWork acts as the interaction layer for natural language desktop control, Wukong serves as the permission layer operating within corporate security boundaries, and MuleRun functions as the execution layer for reusable workflow templates. This approach is similar to Anthropic's methodology, where Claude Chat, Claude Code, and Claude Cowork utilize one base model under a unified interface. The goal of the unified product is to eliminate user confusion, which previously forced customers to choose between three incompatible Alibaba products.

Tencent's advantage lies in its deep integration with the WeChat ecosystem and cloud services. WorkBuddy already maintains a close connection with Tencent Docs, WeCom, and Tencent Meeting, offering a corporate workflow deeply embedded in China's dominant business communication ecosystem, WeChat. ByteDance Doubao leverages the company's experience in recommendation algorithms and content generation. This three-way competition boils down to which company can offer the most comprehensive office experience, from the start of user intent to task completion. The outcome of this battle will likely determine the direction of enterprise software in China for the next decade, as the desktop agent becomes the primary interface for knowledge workers interacting with digital tools, replacing the traditional office suite model with an AI-centric workflow paradigm.