(SINGAPORE 2026.3.12) Over the past two months, “raising lobsters” has become a blazing tech buzzword across China. The “lobster” refers to OpenClaw, an open-source AI agent created by an Austrian engineer in November 2025. Functioning as a digital worker, it can operate a user’s computer, write code, respond to emails, and even trade stocks. But few could have predicted its frenzied adoption by Chinese users, who rapidly turned it into a tool for income generation and routine works.
On platforms like Xiaohongshu and Douyin, tutorials on “raising lobsters” have proliferated. From Beijing’s hutongs to Hangzhou’s West Lake, offline meetups on OpenClaw have sprung up. At a free installation event at Tencent’s Shenzhen headquarters, nearly a thousand people—from children to seniors—lined up for a chance to harness the technology.
Initially, Chinese authorities encouraged OpenClaw’s widespread use, whose red-lobster logo has become familiar across the tech landscape. But this week, the brakes were suddenly applied. On March 8, the Ministry of Industry and Information Technology announced that some OpenClaw instances, when deployed with default or improper configurations, posed significant security risks. The perils obviously include putting China’s top leaders in danger.
In response, Tencent began tapping SkillHub on March 11, a community designed to optimize AI for Chinese users. Focused this time on OpenClaw’s open-source ecosystem, the platform offers local configuration services to mitigate potential risks. According to 36 Kr’s All-Weather Technology (全天候科技） column, this marks another step in Tencent’s counteroffensive against disruptive AI ecosystem.
Tencent also set up “Lobster Installation Stations” at its Shenzhen headquarters, where queues grew so long that signs read, “All numbers have been taken.” Meanwhile, Xiaomi released China’s first mobile version of the “lobster,” kicking off a new wave of smartphone-based AI agents, reported 36 Kr’s Ijincuodao (金错刀）column.
The OpenClaw boom is not a regular phenomenon. Chinese developers use OpenClaw to write code, build websites, and generate short-form drama content. Entrepreneurs scramble to leverage the technology: one bought eight used MacBook Airs to run a separate OpenClaw agent on each to form a 24/7 AI content factory. Others deploy arbitrage bots on prediction platform Polymarket, scanning market movements every five seconds.
During Chinese New Year, Fu Sheng (傅盛), founder of Cheetah Mobile (猎豹移动）, reportedly used eight “lobsters” to complete the work of an entire team: six daily WeChat articles, personalized greetings to 611 employees, a viral Twitter post with over one million clicks, and short videos exceeding 300,000 views. Fu also launched EasyClaw, promising “installation in 3 minutes with one click, self-configuration, and automatic connection to AI service credentials.”
While casual users treat AI as an opportunism tool, tech veterans adopt an “employee mindset,” training AI agents like staff. OpenClaw’s rise suggests that companies now need only an idea and an AI executor. Editors and creative professionals are already calculating how soon AI might replace them entirely.
Hence, the installation frenzy in China is only the beginning. From Xiaomi to Alibaba, companies race to produce “AI employees.” Unlike conversational systems such as ChatGPT or Claude, OpenClaw directly manipulates operating systems, runs tasks independently, and operates continuously, fundamentally changing AI’s role.
The buzz spread rapidly. Of about 142,000 publicly visible OpenClaw agents worldwide, nearly half were created by Chinese users. Within 100 days of its launch, OpenClaw surged on GitHub, a web-based platform for software development, attracting massive attention and contributions from Chinese developers, creating a “digital tidal wave.”
China News (中国瞭望）， a Chinese media outlet outside China, reported that the Chinese government in principle views AI as a new engine of economic growth capable of energizing China’s sputtering economy and fostering innovation. Local governments have joined the effort of promoting OpenClaw. Shenzhen’s Longgang district and Wuxi High-Tech Zone offered subsidies, computing resources, and data support—sometimes up to 5 million yuan (S$950,000).
Local authorities are keen to promote the concept of “one-person factories,” where AI agents deliver highly efficient output. In Shenzhen’s Futian district, civil servants deploy “government lobsters” to handle public complaints automatically, producing analytical reports and reducing manual workload.
Then, out of the blue, the Ministry of Industry and Information Technology recently warned of vulnerabilities: OpenClaw can be manipulated, misconfigured, or maliciously take over a system, potentially leading to information leaks or unauthorized operations. State-owned enterprises and government agencies have begun prohibiting OpenClaw on official devices, with personal tools also restricted from connecting to office networks. Staff who have installed it must report immediately.
Speculation arises that Chinese authorities fear OpenClaw’s capabilities in infiltrating Zhongnanhai, where China’s top leadership resides, risking exposure of sensitive information. OpenClaw’s deep integration with underlying operating systems allows autonomous control, file access, and command execution, making it a powerful threat to leadership security.
Now, as people fly in from Hangzhou and Hong Kong to Shenzhen to queue for Tencent engineers to install the little red “lobster,” one official signal has become clear: OpenClaw in China must not go beyond being an investor’s or gambler’s tool. It may flow into the computers, phones, and daily lives of ordinary Chinese, but tech giants like Tencent are monitoring and managing its adoption, standing guard to curb its “subversive” side.
