From Side Project to 7,000 Orders
In January, Beijing software engineer Feng Qingyang started tinkering with OpenClaw, a newly released Chinese AI tool that can take over a device — a phone, a computer, a smart appliance — and autonomously complete tasks on the user's behalf. Within weeks, he was advertising installation support on a second-hand shopping platform. By early March, he had more than 100 employees and had completed over 7,000 orders for customers who wanted OpenClaw running on their devices but lacked the technical knowledge to set it up themselves.
Feng's story is unusual in its scale but representative of a broader phenomenon. OpenClaw has become a sensation in a country where early AI adoption is often faster, broader, and more commercially creative than in Western markets. A cottage industry of installation specialists, preconfigured hardware vendors, and training service providers has sprung up around it, creating economic opportunity while raising significant questions about security risks.
What OpenClaw Does
OpenClaw is what AI researchers call an agentic system — an AI that does not just generate text or images but takes actions in the world. Given a goal, it can operate a device's interface, fill out forms, make purchases, send messages, navigate applications, and complete multi-step tasks that previously required direct human involvement. For non-technical users — which is to say the vast majority of people — OpenClaw offers capabilities previously accessible only to those who could write code or hire developers.
The Security Dimension
The OpenClaw craze is unfolding against a backdrop of genuine and significant security risks. An AI system with autonomous control over a device — the ability to access files, send messages, make purchases, and interact with any application — is also a powerful vector for data exfiltration, unauthorized transactions, and privacy violations if it is compromised, misconfigured, or misused.
The installation services ecosystem introduces additional risks. Customers who hire third parties to install and configure OpenClaw on their devices are, in effect, granting those parties significant access to their digital lives. Chinese security researchers have flagged these concerns publicly, but commercial momentum is powerful, and the gap between adoption pace and security analysis is growing.
What This Tells Us About Global AI Adoption
The OpenClaw phenomenon illustrates something important about the current AI moment: the most consequential adoption is not always happening in the contexts that Western technology journalists pay closest attention to. China has tens of millions of citizens who are heavy users of sophisticated AI tools developed by Chinese companies with different design priorities and governance structures than their Western counterparts.
The entrepreneur class forming around OpenClaw — providing installation, training, and support services for a technology most users cannot self-serve — is a preview of economic dynamics that will play out globally as AI capabilities become more powerful and widely distributed. Wherever AI creates new capabilities requiring human intermediation to access, new business models and economic opportunities will emerge.
This article is based on reporting by MIT Technology Review. Read the original article.
