--- title: "China is stirring up an OpenClaw storm." type: "News" locale: "en" url: "https://longbridge.com/en/news/278246541.md" description: "China is experiencing a surge in AI technology with Tencent's OpenClaw initiative, where users can install AI agents for free. Major companies like Xiaomi are also developing their own AI systems. The competition is not just about AI tools but also about creating new revenue streams from computing power. As traditional user interactions fail to generate sufficient cash flow, local agents like OpenClaw are designed to maximize API usage, creating a continuous demand for cloud services. This shift is crucial for tech giants to maintain their investments in computing power and to gather valuable trajectory data for future AI development." datetime: "2026-03-08T03:58:49.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/278246541.md) - [en](https://longbridge.com/en/news/278246541.md) - [zh-HK](https://longbridge.com/zh-HK/news/278246541.md) generator: "portal-rs" --- # China is stirring up an OpenClaw storm. Author:Song He; Source: All-Weather Technology In early March, at the Tencent headquarters in Shenzhen, Tencent engineers set up a stall in the north square of the building, like a market, to install "Lobster" OpenClaw for users for free.The line was long and continuous, with some carrying NAS, some carrying MacBook, and others carrying mini PCs, resembling a geek gathering of Android system flashing ten years ago.In fact, many large companies are intensively promoting their own "Lobster". Xiaomi has begun internal testing of MiclawAgent, hoping to embed AI agents into its "human-vehicle-home ecosystem," making smartphones, cars, TVs, and home appliances execution nodes for AI. Cloud vendors have also entered the fray. With major terminal manufacturers starting to cram agents into their operating systems, this "lobster" storm has begun, ushering in the second half of the big model competition. This is not simply a battle over AI tools, but a covert war over the next generation of "super gateways." ## **One ​​** **Cash flow from selling tokens** Currently, a dilemma faces all players: the simple "Chat" model simply cannot generate a healthy business model. Over the past two years, domestic cloud vendors and tech giants have been embroiled in a long-term arms race, with tens of thousands of high-end computing cards being systematically deployed to data centers. In 2026, ByteDance, Alibaba, and Tencent's combined capex exceeded $60 billion. However, if users don't utilize this computing power, it remains idle, incurring significant depreciation daily. The reality is that relying solely on user-facing interactions not only fails to utilize such a massive computing power reserve but also fails to generate revenue from users accustomed to free services. Users occasionally asking AI to write emails or draw diagrams; these single interactions consume low amounts of tokens, insufficient to cover the depreciation and operating costs of the underlying massive computing power clusters. To keep this expensive computing power running and generating real cash flow, these giants urgently need a "token black hole" that can continuously and automatically consume computing power. Locally deployed agents like OpenClaw have emerged to fill this role. When a user issues a complex command, OpenClaw breaks down the task, searches the internet, calls local software, identifies errors, and self-corrects and retryes. Each of these steps sends a request to the cloud's API interface. The token consumption for a complex task is hundreds or even thousands of times that of a normal conversation. An AI analyst pointed out to Wall Street Insights: "OpenClaw adopted the Chinese open-source model primarily because of its high cost-effectiveness. Compared to overseas competitors, the lower cost allows for more frequent API calls, which directly translates into cash flow for cloud vendors, avoiding the waste of huge investments in computing power." This is why cloud vendors like Tencent are willing to subsidize manpower to set up offline "stalls" to help users deploy open-source agents, and why Alibaba strongly promotes OpenClaw's one-click cloud deployment. Each deployment is like burying a 24/7 "computing power pump" in the user's local or cloud computer. Regardless of whether the front-end runs an open-source model or not, as long as the APIs called by inference and tools point to their own cloud services, the massive number of tiny requests will eventually converge into considerable B2C and B2B cash flow. Under the current capital market's stringent scrutiny of the commercialization of large models, this API revenue driven by agents is a crucial lifeline for giants to maintain their computing power expansion. Beyond the first layer of cash flow, the second goal of giants pushing local agents has reached the ceiling of large model development: the depletion of high-quality training data. For the past few years, the core resources for large-scale model competition have been computing power and training data. However, as model capabilities continue to improve, another resource is becoming increasingly important: trajectory data. The current consensus is that high-quality publicly available text data on the internet (Wikipedia, news reports, books, and academic papers) has already been largely consumed by large models. If they continue to be fed only this static text, large models will only become more erudite "bookworms," ​​unable to move towards true AGI. What does the next generation of large models need? They need to know how humans "take action" in this digital world. This is what the industry desperately craves: "trajectory data." When a user instructs AI to perform a task, the AI ​​goes through a series of steps. From understanding needs to searching for information, then to using tools, filling out forms, and completing payments, every action leaves a record. These records form a complete task chain. For agent models, this data is more valuable than ordinary text because it reflects the logic of actions in the real world. This is precisely the data that giants previously found most difficult to obtain. This data is hidden deep within countless fragmented software programs, closed apps, and corporate intranets, rendering even search engines with massive web crawler ecosystems powerless. OpenClaw deployed on user terminals and miclaw at the system level act as "data detectors" penetrating deep behind enemy lines. Alan Feng, OpenClaw's China community manager, points out: "Users often expect magical automation after installing OpenClaw, but its true value lies in clearly defined tasks. Trajectory data feedback allows the model to continuously optimize, which manufacturers can then use to enhance the agent's capabilities." When users run the agent locally, allowing it to perform operations on their behalf, the agent records every user's intention and software interaction trajectory. The intensive promotion of agent applications by major domestic companies is essentially a distributed, unprecedented-scale data crowdsourcing effort. Users believe they are getting free AI labor; in reality, by guiding and correcting the agent's errors, users are providing the giants with the highest quality reinforcement learning fine-tuning data for free. Once this "trajectory data" flows back to the cloud, it will become a core barrier for major companies to train next-generation agent models with strong logical reasoning and execution capabilities. This is similar to how Tesla collected real-world road condition data from millions of electric vehicles on the road, ultimately feeding back into its FSD (Full Self-Driving) algorithm. An insider from Alibaba's Qwen project told Wall Street Insights, "The probability of China leading the new paradigm is less than 20%, but through agent trajectory data, Alibaba can quickly iterate its models and narrow the gap." Now, giants are turning users' computers and mobile phones into "data collection vehicles" for the AI ​​era. Whoever controls the most trajectory data will be the first to train a truly "armed" super model. From this perspective, major companies promoting local agents is not just about creating a new tool. They are still vying for the operational entry point in the AI ​​era. The battle for entry points in the Chinese internet has actually gone through several typical rounds of entry point wars. Early portal websites competed for homepage traffic; in the search era, Baidu became the information entry point; in the mobile internet era, the user entry point became apps, with WeChat, Alipay, and Douyin gradually becoming traffic centers. But the emergence of AI is changing this structure. Alibaba's Qianwen platform continues to invest in "AI-powered services," allowing users to place orders with just a sentence; Xiaomi is internally testing miclaw, deeply embedding it into the underlying system of mobile phones. These actions signal that in the future, the user interface for interacting with the digital world will be restructured. When users become accustomed to expressing their needs in a single sentence, the operational path will change. Users will no longer actively open an app, but instead entrust tasks to AI. AI will decide which platform to use, which service to call, and which payment link to complete. Therefore, in such a system, the status of apps will change. They will still exist, but will become more of a service node. The real entry point is the Agent that helps users complete tasks. In this new context, "grabbing app entry points" is outdated. The real battle is to become the "underlying agent" that directly obeys users and controls the overall situation. If tech giants can allow their agents to dominate users' devices, they will wield the most powerful force in the business world—the power to distribute intent. They can easily redirect food delivery orders to their affiliated companies and travel requests to their payment ecosystem. In this new "walled garden" built by agents, once-dominant super apps will be reduced to mere "pipelines" providing only basic service interfaces, completely losing the opportunity to directly engage with users and losing brand and traffic premiums. This is why major companies are so sensitive to agents. Everyone wants to be the platform that controls the agents. The explosive popularity of OpenClaw may just be a signal. The real change is that AI is transforming from a "talking tool" into a "system that can do things." Over the past two years, the core goal of the large model industry has been to improve intelligence levels, but now more and more companies are beginning to consider another question: how to give AI the ability to act. Once AI can reliably perform tasks, the structure of the internet will change. Many applications may retreat to the background, and users will only need to interact with an agent to complete most of their digital life operations. In this world, the agent acts as a new operational layer, connecting users with all services. Looking back at technological history, every platform-level change often begins with a seemingly insignificant beginning. Android was initially just a system for geeks to customize their devices, WeChat's official accounts were initially just a simple content tool, and mini-programs were more like lightweight web pages. But these products later became new platforms. If AI truly enters the agent era in the future, then OpenClaw today is likely to be one of the first names to be remembered. What the Chinese internet is experiencing may be the eve of this storm. ### Related Stocks - [KWEB.US](https://longbridge.com/en/quote/KWEB.US.md) - [159998.CN](https://longbridge.com/en/quote/159998.CN.md) - [TCTZF.US](https://longbridge.com/en/quote/TCTZF.US.md) - [512720.CN](https://longbridge.com/en/quote/512720.CN.md) - [CLOU.US](https://longbridge.com/en/quote/CLOU.US.md) - [MCHI.US](https://longbridge.com/en/quote/MCHI.US.md) - [512380.CN](https://longbridge.com/en/quote/512380.CN.md) - [TCEHY.US](https://longbridge.com/en/quote/TCEHY.US.md) - [XIACY.US](https://longbridge.com/en/quote/XIACY.US.md) - [01810.HK](https://longbridge.com/en/quote/01810.HK.md) ## Related News & Research - [Tencent Details Progress Toward Carbon Neutrality Goals and Outlines AI-Era Priorities | TCEHY Stock News](https://longbridge.com/en/news/296040172.md) - [Another AI lab is burning through cash as Tencent earnings rise](https://longbridge.com/en/news/295641594.md) - [The push for AI watermarks is spawning a new wave of tools to remove them](https://longbridge.com/en/news/296262434.md) - [Argentum AI Announces $10 Billion+ in Contracted Revenue and Targets 1GW of AI Infrastructure Capacity in 2026](https://longbridge.com/en/news/296261035.md) - [Tranchi AI Launches an Always-On AI Acquisition Employee for Real Estate Investors](https://longbridge.com/en/news/296524606.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**