Several Chinese technology hubs are racing to build an ecosystem around OpenClaw, an open-source AI agent that automates digital work such as booking travel, managing emails, and coordinating tasks across software platforms.
Cities including Shenzhen, Wuxi, Hefei, and Suzhou have begun offering subsidies, computing resources, and office support to startups building products around the technology. Some programmes promise incentives worth up to 10 million yuan for companies adopting the platform.
Local governments see a clear opportunity. AI agents promise to compress the size of a typical startup team. A single founder could rely on automated tools to manage marketing, coding, and operations. That model has attracted policymakers eager to create new innovation clusters.
The idea resembles a career decision many professionals recognise. A worker with the right digital tools can suddenly handle tasks that once required an entire department. Governments appear to believe AI agents could produce the same effect at a national scale.
The platform’s popularity has grown quickly among developers and entrepreneurs.
OpenClaw’s rise also carries risks. Cybersecurity experts warn that poorly configured systems could expose sensitive information or grant attackers access to internal networks.
One security specialist highlighted the danger of granting AI agents extensive system permissions.
OpenClaw’s “super permissions” and “super capabilities” could bypass human-set safeguards if misconfigured or manipulated, creating significant risks, according to security expert Wang Liejun.
Chinese regulators have already raised concerns about data security and cross-border transfers as adoption accelerates. Authorities are exploring compliance centres and stricter controls on how the software interacts with sensitive data.
Developers continue to experiment regardless. The platform’s open-source nature allows companies to customise it for different industries, from logistics planning to automated customer service.
The tension reflects a broader policy challenge. Governments want to accelerate AI innovation, yet they must manage the risks created by autonomous software with access to real systems.
China’s strategy appears to balance both goals: encourage rapid experimentation while tightening oversight as the technology spreads.
That raises a difficult question for the global AI industry. If autonomous agents can run entire workflows, will the next generation of startups consist of small teams augmented by powerful software—or will security failures slow that vision before it fully takes hold?
Author: Pishon Yip