AI Warfare, Now Written Into Law: China Just Made Its Civilian Tech a Military Reserve

China's first defence-mobilisation law revision in 16 years quietly turns AI, drones, and data into wartime assets — and formalises the civil-military fusion watchchina.ai has documented all summer

The Quiet Law With Loud Implications

For sixty-five briefings, watchchina.ai has documented the blurring line between China's civilian AI boom and its military ambitions — the drone swarms of Package #12, the DeepSeek-Norinco links, the Pentagon's own autonomous-weapons push in Package #37. This week, Beijing did something that makes that blur official, permanent, and legally binding, and it did so with characteristically little fanfare.

China has made the first amendment in 16 years to its national defence mobilisation law to include a fresh focus on new technologies. The revised law, passed by the top legislature on August 28 and taking effect October 1, states that the state promotes the application of advanced technologies in national defence mobilisation and develops national defence mobilisation forces in emerging fields. Though the text does not name them, a Beijing-based military law expert specified exactly what those advanced technologies are: cyberspace, AI, unmanned systems including drones, the electromagnetic spectrum, biotechnology, and quantum technologies. Crucially, the revised law also includes data and software as part of the military's reserve capacity, alongside conventional equipment. China has, in a single legislative stroke, defined its civilian AI industry as a strategic military reserve — to be activated at the state's command.

What "Civil-Military Fusion" Now Means in Practice

watchchina.ai has referenced China's civil-military fusion doctrine repeatedly, but this law converts doctrine into enforceable mechanism, and the specifics matter. The amendment incorporates civilian hi-tech enterprises and professionals into China's mobilisation potential — meaning that in peacetime, these capabilities serve economic and social development; in wartime, they can be rapidly reassigned to provide operational support. The revised law even provides financial penalties for entities or individuals that fail to comply with designated orders or deliberately delay implementation. This is not aspirational language. It is a compliance regime with teeth, binding private companies to the state's wartime needs.

Consider what that means for every entity watchchina.ai has profiled this summer. DeepSeek, Alibaba's Qwen, ByteDance's models, Unitree's robots, the 140-trillion-token infrastructure, the DUV lithography lines — all of it now sits, formally, within a legal framework that allows the state to convert economic and social capabilities into national defence strength on demand. The military law expert was explicit about the trigger: the law applies when the country's sovereignty, unity, territorial integrity, security and development interests are under direct threat — with development interests a notable new addition that broadens the threshold considerably. And the same expert named the reference conflict directly: the Russia-Ukraine war validated the critical role of unmanned, intelligent and cyber technologies in modern warfare, reinforcing the necessity of this legislation. China is legislating for a war it is studying carefully.

The Honest Reading — Symmetry and Its Limits

As across all sixty-five briefings, watchchina.ai insists on measuring both superpowers by the same standard, and here the symmetry is real and important. The United States does much the same thing through different means: the Defense Production Act allows Washington to compel private industry in emergencies, Palantir and Anduril fuse Silicon Valley with the Pentagon, and — as watchchina.ai documented in Package #37 — the Pentagon pressured Anthropic to permit autonomous weapons and branded it a security risk for refusing. Every major power is racing to militarise civilian AI; China is not uniquely aggressive in wanting to, and anyone framing this law as proof of singular Chinese belligerence is ignoring Washington's parallel trajectory. The convergence watchchina.ai has tracked in commercial AI has a military twin.

But the differences are also real, and watchchina.ai will not flatten them to force a false equivalence. China's system offers no independent courts to challenge mobilisation, no public dissent of the kind that saw forty US researchers file briefs supporting Anthropic, no meaningful check on a compliance regime that penalises companies for hesitating. When the American state reaches for private tech, it faces friction — lawsuits, leaks, congressional hearings, employee revolts. When the Chinese state reaches, this law ensures it closes its hand. That is the consequential distinction: not that one power militarises civilian AI and the other doesn't — both do — but that one does so within a system of contestable checks and the other within a system engineered to eliminate them. The deeper significance for watchchina.ai's readers is timing. This law takes effect October 1, just weeks after the September 24 Trump-Xi summit where AI cooperation is on the agenda. Beijing arrives at that table having just legally fused its entire civilian AI ecosystem to its war machine — a reminder that beneath the diplomatic language of "AI safety dialogue," both powers are hardening, not softening, and the civilian models the world runs on are, increasingly, strategic assets of states preparing for the possibility of conflict. The future of AI is being written in law now. This week, China wrote the military chapter.

Sources: South China Morning Post, UPI, Taipei Times, AzerNews, Xinhua (via SCMP), CGTN