China launches ‘model–data resonance’ AI+manufacturing campaign

The “模数共振” action is China’s most explicit attempt yet to industrialise AI, turning large models and domain‑specific agents into standard tooling across core manufacturing sectors. Rather than focusing on one flagship frontier model, Beijing is pushing for a dense ecosystem of industry models, specialized agents and curated datasets spanning steel, petrochemicals, machine tools, autos, electronics and more. That’s less about glamorous chatbots and more about embedding AI into the real economy’s control loops.

From an AGI‑race perspective, this program matters because it links national compute, data and industrial policy into a single flywheel. If factories continuously generate high‑quality operational data that flows back into sector‑specific models, China could accumulate an enduring advantage in applied AI even if its frontier general‑purpose models lag slightly behind US labs. It’s also a template for how states can steer model development toward sectors they care about—climate, logistics, defense—by coordinating data access and deployment incentives. For Western labs and governments, the question is whether similar model–data programs emerge around semiconductors, energy and critical infrastructure.