(ECNS) -- OpenClaw, an AI agent, has quickly become one of the most popular projects on the code-sharing platform GitHub. Unlike traditional conversational AI systems, OpenClaw, nicknamed "lobster" by users -- is built to take action.
Zeng Jianqiu, a professor at Beijing University of Posts and Telecommunications, told China News Network that OpenClaw was developed by Austrian programmer Peter Steinberger and represents a typical open-source AI agent. Its goal, he said, is to solve highly complex problems with the simplest possible tools.
More like a digital human than a chatbot, OpenClaw can learn, carry out tasks, and directly assist users with real-world operations. Online users have shared videos showing it organizing computer desktops and processing data.
According to Zeng, OpenClaw's rapid rise reflects a new phase in AI development. He said that while AI is expected to evolve from weak AI to general AI and eventually strong AI, today's question-and-answer systems are still in the very early stage of weak AI. In his view, the emergence of OpenClaw suggests the industry is beginning to move toward truly intelligent systems.
The growing enthusiasm for "raising lobsters," the term users use for deploying OpenClaw -- has also turned into a new battleground for major tech companies.
OpenAI has made one of the boldest moves so far. Its CEO, Sam Altman, not only announced plans to deploy OpenClaw internally, but also recruited its creator.
Chinese tech companies have moved quickly as well. The Baidu App has added one-click access to OpenClaw, while Alibaba Cloud, Tencent Cloud, and China Unicom Cloud have all launched enterprise deployment solutions.
Small and medium-sized businesses are also seeing benefits. Xiang Ligang, chairman of the Zhongguancun Information Consumption Alliance, said his team has already begun installing OpenClaw to improve efficiency in marketing, customer service, and product development.
So why has OpenClaw attracted so much attention?
On the surface, the race is about controlling user entry points. In 2026, AI agents were included in China's government work report for the first time, with a target of reaching more than 90 percent penetration by 2030. Whoever builds the universal "super entry point" could shape the next era of digital traffic distribution.
At a deeper level, the competition is about data resources. The International Data Corporation (IDC) predicts that the number of active AI agents in Chinese enterprises will exceed 350 million by 2031, with a compound annual growth rate of more than 135 percent. This expansion is expected to drive annual token consumption to more than 30 times its current level, what experts describe as the "digital feed" of the AI era.
Even more valuable is user trajectory data. When OpenClaw books a flight for a user, for example, every click, input, error, and correction becomes part of a detailed behavioral dataset. Such data could be extremely valuable for training AI decision-making systems.
Zeng said the rise of OpenClaw is a reminder that future AI agents will not simply be upgraded tools, but an entirely new form of existence. They may be able to communicate, collaborate, and evolve on their own, he said, which means humans will need to move beyond current ways of thinking.
At the same time, risks are emerging alongside this rapid growth.
Recently, China's Ministry of Industry and Information Technology warned that some open-source OpenClaw deployments could pose significant security risks if configured improperly, potentially leading to cyberattacks or data leaks.
To reduce those risks, OpenClaw can still be deployed effectively, but only with proper safeguards. Experts say it is best to begin with limited permissions and gradually expand access, balancing convenience with security.
(By Gong Weiwei)
京公网安备 11010202009201号