🧠 Food for thought

Implications, context, and why it matters.

Beijing’s warning contrasts with multi-million-yuan local subsidies for OpenClaw

• China’s central authorities raised security alarms while several local tech hubs backed an OpenClaw surge 1.
• Shenzhen, Wuxi, and Hefei drafted policies with subsidies up to 10 million yuan (US$1.4 million) plus free computing resources for firms building OpenClaw-focused AI ecosystems 2.
• The packages support “one-person companies” (small firms run by a single founder using AI tools) and local “AI plus” industrial plans (programs to apply AI across traditional industries) 1.
• In February 2026, the Ministry of Industry and Information Technology (MIIT) warned that default or poorly configured OpenClaw deployments can create security risks 3.

The crackdown underscores a broader struggle over control of agentic AI

• After the warning, Tencent gave up much of its earlier rise, and MiniMax reversed gains then traded more than 1% lower [summary].
• MiniMax has plugged its models into the OpenClaw ecosystem 4.
• The dispute ties to a fight over who steers agentic AI tools.
• The Doubao AI phone case involves an AI agent built into a smartphone operating system that can move through apps with system-level permissions. OpenClaw runs on computers, and regulators warned that its access to private data can raise data-security risks 5.
• Arguments over agentic AI often focus on data access and control of traffic (how users and information flow between apps and services). They also cover permissions that let agents work across fragmented app ecosystems. Similar issues surfaced in China’s Doubao AI phone controversy and around OpenClaw 5.