China’s ‘Lobster’ AI Craze Draws Security Warnings as Operating Costs Rise
Udumbara
.
An open-source artificial intelligence tool known as OpenClaw—nicknamed “lobster” by Chinese users—has surged in popularity across China’s tech community in recent months, becoming one of the country’s most talked-about AI trends. However, increasing operating costs and regulatory warnings about cybersecurity risks could slow the technology’s rapid expansion.
The Chinese regime has issued warnings that the software could pose cybersecurity and data-leak risks, highlighting the tension between Beijing’s push to accelerate AI development and its concerns over information security.
An AI-Powered ‘Digital Worker’
OpenClaw was first developed by Austrian software developer Peter Steinberger and released in November 2025. Its icon—a red lobster—quickly inspired Chinese users to nickname the software “lobster,” and the phrase “raising lobsters” has since become slang for running the program.Unlike traditional AI chatbots, OpenClaw functions as an autonomous AI agent. After receiving user authorization, it can carry out complex tasks independently, such as reading documents, searching for information, writing code, and sending emails.
Since it can complete multi-step assignments with minimal human input, OpenClaw is often described by Chinese software engineers as a form of “digital worker.”
Major Chinese cloud providers—including Tencent, Alibaba Group, and Baidu—have already integrated OpenClaw deployment options into their platforms, thereby accelerating its adoption among developers and businesses.
Xue Zongzhi, a former procurement manager at Taiwan Semiconductor Manufacturing Company, said that OpenClaw essentially functions as a much more advanced version of a digital assistant.
“It’s like having your own personalized Siri,” he told The Epoch Times. “OpenClaw will help you ask questions and analyze your needs. It’s an algorithm that adapts to you and keeps learning.”
Hidden Costs
Despite the hype, some experts warn that operating AI agents like OpenClaw can be far more expensive than users expect.Running the software requires continuous computing power and large volumes of data processing, which can quickly escalate costs, according to Xue.
Zhao Junshuo, host of the Chinese-language geopolitical analysis program “U.S.-China-Taiwan Situation Room,” told The Epoch Times that executing complex AI tasks consumes substantial computing resources.
“Whenever the AI is asked to handle complicated problems, it has to perform large-scale searches and calculations,” Zhao said. “And that [computing power] costs money.”
While promotional pricing may make the service appear inexpensive at first, he said, widespread adoption could create financial strain for smaller companies.
Reports from Chinese media suggest those concerns are already emerging. China’s National Business Daily reported that some Chinese tech companies are abandoning plans to automate operations with OpenClaw after the platform’s Application Programming Interface (API) usage fees surged unexpectedly.
Beyond cost issues, users have also raised concerns about security risks.
Chinese social media users reported instances in which the software accidentally deleted files, exposed private information, or performed unintended transactions during automated tasks.
Those risks prompted Chinese regulators to intervene.
China’s Ministry of Industry and Information Technology warned through the regime’s mouthpiece, People’s Daily, that some OpenClaw deployments contain security vulnerabilities, particularly when default configurations are used. The ministry said the software could increase the risk of cyberattacks or the leak of sensitive information if security controls are insufficient.
