China Is Upgrading the World’s Largest Surveillance Network — With AI

Cameras that think, predict, and act — before humans do

From Hardware to Intelligence: A Surveillance System Reborn

China is not just maintaining the world's largest surveillance network — it is rebuilding it from the ground up. According to a Financial Times investigation, local governments across China are deploying new AI-powered systems that significantly enhance capabilities to track individuals, analyze behavior, and anticipate potential unrest in real time — representing the most significant modernization of the network in years. The original infrastructure, built roughly a decade ago, had begun to show its limits: older hardware and fragmented software were no longer sufficient for the state's ambitions.

What is replacing it is fundamentally different in nature. An FT analysis of more than a dozen procurement documents found that Chinese local governments are deploying AI-powered surveillance with computer vision and large language models running directly on cameras — enabling text-prompt video search and real-time behavioral analysis at the edge, without the need to transmit data to a central server. In practical terms, this means a camera can now be queried like a search engine — and respond in seconds.

Predicting Unrest Before It Happens

The ambition goes well beyond identifying suspects after the fact. China's State Council stated in August 2025 that its "AI+ initiative" would include building an "early warning system" where AI and humans work together for a stronger national security framework — flagging incidents as soon as they begin to develop, giving police advance warning.

The ideological framing is revealing. Rather than representing a technological break with the past, AI in this context serves primarily to entrench governance ideas that are decades old — the difference being that the state now has tools capable of executing those ideas at a scale and speed no human bureaucracy could match. In Xinjiang, the implications are already concrete. Surveillance in the region in 2026 is described by human rights researchers as more sophisticated, more integrated, and equally coercive compared to 2019 — with expanded biometrics, drone networks, and AI-powered case management systems now standard.

A Global Export Product in the Making

The international dimension of this story is perhaps its most consequential aspect. Beijing has invested hundreds of billions of dollars into AI-related businesses, making major strides in research and development — and the authors of a recent report warn that the implications are both broad and deep, strengthening China's power overseas as a global exporter of surveillance technology.

Meanwhile, the US Congress is growing increasingly alarmed about the hardware layer underpinning all of this. Nvidia CEO Jensen Huang has been invited to testify before the Senate Banking Committee on June 11, as lawmakers warn that advanced US chips could be used by China to strengthen its military and surveillance capabilities. The hearing will be one of the most closely watched in Washington this summer.

Sources: Financial Times, AI Weekly, China Media Project, Business Standard, CNBC, Human Rights Watch, Jamestown Foundation