TLDR: Chinese technology powerhouses Tencent and ByteDance are significantly ramping up their efforts in the burgeoning field of AI agent development. This strategic pivot reflects a broader trend in China’s AI industry, moving from foundational model creation to the deployment of practical, task-specific AI applications and the establishment of robust AI-driven ecosystems. Other major players like Alibaba are also heavily invested, as the market for AI agents is projected to see substantial growth, driven by both corporate innovation and government initiatives.
Chinese technology giants Tencent and ByteDance are at the forefront of a new wave of innovation, aggressively entering the competitive arena of AI agent development. This strategic shift, observed throughout 2025, signifies a maturation of China’s artificial intelligence sector, moving beyond the initial focus on large foundational models to practical, application-oriented AI agents capable of automating complex tasks.
Tencent Holdings recently open-sourced its new ‘Youtu-Agent’ agentic framework, developed by its AI research department, Youtu Labs. This framework was released on Microsoft’s open-source code-hosting platform GitHub. An AI agent built on the open-source DeepSeek-V3.1 model using Youtu-Agent achieved a score of 71.47% on WebWalkerQA, a web traversal benchmark. The company has also unveiled a series of AI agents designed to assist businesses and consumers, offering solutions for online marketing, coding, and automating workflows. Tencent’s AI app, Yuanbao, is being heavily promoted across its ecosystem, including WeChat, to gain a leading edge in the AI race.
ByteDance, the parent company of TikTok, has also made significant strides by open-sourcing its AI agent development platform, ‘Coze Studio,’ in July 2025. This platform reportedly supports millions of developers in creating AI tools. ByteDance’s Douyin is integrating its own AI large model, Doubao, into its platform, marking a shift from a standalone super-app strategy to deeper ecosystem integration.
Not to be outdone, Alibaba Group Holding has also intensified its AI agent efforts. The company open-sourced its ‘Qwen-Agent’ framework in March 2025 and introduced an enhanced AI model, ‘Qwen3-235B-A22B,’ which is noted for its capabilities in task planning and execution. Alibaba’s Quark has been rebranded as an ‘AI assistant for 200 million users,’ with its ‘Deep Thinking’ AI search feature entering the market in late February 2025.
This intensified competition reflects a broader industry trend in China, where companies are pivoting from developing raw AI capabilities to creating tools that solve specific business problems. Liu Dingding, a technology industry observer, highlighted that major tech companies possess a distinct advantage in deep reasoning AI due to their extensive ecosystems, enabling seamless integration of AI into their services to enhance user experiences. Conversely, standalone AI firms may face challenges without such ecosystem support.
The market for AI agents in China is experiencing rapid expansion. Projections from Hap Academy, a Beijing-based tech industry think-tank, indicate that the market is expected to quadruple from 2024 to reach 10.9 billion yuan (approximately S$1.96 billion) in 2025. The Chinese government is actively supporting this growth, having released a roadmap in late August 2025 to accelerate AI integration across the economy and society, an initiative dubbed ‘AI Plus.’ The government aims for the adoption of AI agents and intelligent devices to exceed 70% by 2027 and 90% by 2030. A May 2025 survey by Accenture revealed that 53% of Chinese companies are integrating AI across multiple workflows, 11 percentage points higher than the global average.
Also Read:
- Colgate-Palmolive Accelerates Growth Strategy with Major Agentic AI Investment
- Guangxi Solidifies Position as Key AI Innovation Hub for China-ASEAN Collaboration
Geopolitical factors have also played a role in this landscape, with Chinese startup Butterfly Effect, known for its Manus AI agent, relocating overseas, further fueling the domestic push for AI agent development.
