China’s OpenClaw Boom Is a Gold Rush for AI Companies
George Zhang, working in cross-border ecommerce in Xiamen, China, became fascinated by OpenClaw after seeing a viral video demonstrating its potential to autonomously manage stock portfolios and make investment decisions. Despite not fully understanding how the AI agent software worked, Zhang installed OpenClaw in late February, joining a wave of Chinese users caught up in the recent excitement surrounding the technology. Across China, workshops teaching OpenClaw usage have attracted hundreds, and major tech companies are swiftly integrating it into their platforms. Local governments have even announced subsidies for entrepreneurs developing products based on OpenClaw, with images going viral of elderly people lining up to install the software.
Zhang rented a cloud server from Tencent and subscribed to the Chinese large language model Kimi, enabling him to interact with his OpenClaw agent, or “lobster” as it’s colloquially known. Initially impressed by the AI’s ability to generate detailed market analyses swiftly, Zhang soon found his agent’s performance declining, offering only basic trend outlines and repeatedly claiming to be “working on it” without delivering results. He concluded that OpenClaw was poorly suited for those without coding skills, as it frequently prompted him to configure technical settings beyond his expertise. Ultimately, Zhang abandoned using the agent for stock trading, instead repurposing it for aggregating AI industry news to fuel social media content production on WeChat.
Discussion among other OpenClaw users in China reveals a clear split: technologically adept adopters find the agent a transformative productivity tool, while nontechnical users feel misled and disappointed by its capabilities after investing in costly cloud servers and large language model tokens. The true momentum behind OpenClaw’s popularity lies with Chinese companies like Tencent, Alibaba, ByteDance, Minimax, Moonshot, and Z.ai, which benefit considerably as ordinary users start paying for AI services. A single active OpenClaw instance can consume significantly more tokens daily compared to typical chatbot use, which explains why Tencent engineers have been offering free installation help to encourage adoption.
Song Zhuoqun, a Chinese college student interning in social media at an AI startup but lacking programming knowledge, found OpenClaw installation baffling and frustrating. Despite seeking help via AI chatbots to guide her, she encountered incomprehensible code and repeated errors, ultimately gaining little from the process. Her experience mirrors many others who joined the OpenClaw trend under promises of user-friendly AI tools, only to face technical barriers. Even prominent figures like Changpeng Zhao expressed skepticism on social media, warning that users spend more time troubleshooting than benefiting from the AI.
Startup founder Rain Miao advised that users unable to navigate installation and permissions should avoid OpenClaw altogether, suggesting alternatives like Claude Cowork, which have seen less attention in China. Most nontechnical users rely on rented cloud servers and paid API access to run OpenClaw, as local hardware often lacks the necessary compatibility or power. Zhang estimates his expenses for setup and subscription approaches $30, rising significantly with complex tasks that consume many tokens. Skilled users might reduce costs by delegating demanding tasks to more capable models like ChatGPT and running simpler processes locally, but this requires technical expertise.
Recent social media jokes in China highlight that while OpenClaw costs money for tokens, unpaid interns can be employed instead for free labor, underscoring token expenses. The broader significance of the craze is that ordinary Chinese are willing to pay for AI, a shift from a cultural norm of free software supported by data and ads. Companies have seized this opportunity, offering help, tutorials, and launching customized OpenClaw versions—Tencent’s QClaw, ByteDance’s ArkClaw, Moonshot’s KimiClaw, and Z.ai’s AutoClaw—aiming to lock users into their platforms. The original OpenClaw founder publicly criticized these companies for copying the software without supporting its development.
Promoters of OpenClaw in China often overlap with crypto event organizers, now pushing Web4.0 concepts centered on autonomous AI agents rather than Web3.0 narratives. Technically adept engineers have monetized the trend by installing OpenClaw for others, reportedly performing thousands of such installations at a fee. Local governments, despite security concerns tied to OpenClaw, have offered developer grants to capitalize on the hype, signaling eagerness to attract tech talent, although the practical outcomes remain uncertain. This pattern echoes earlier government efforts promoting virtual technologies like the Metaverse through subsidies, with limited success in drawing user engagement.
