China's Ministry of Industry and Information Technology has launched an industrial data foundation initiative that aims, by the end of 2026, to cultivate industry data collaboratives, build trusted interconnection platforms for key sectors, aggregate industry data, draft industrial data standards and create high-quality, standardized, tradable industry datasets. The article argues that industrial AI adoption is shifting from model capability contests toward data resources and closed-loop deployment scenarios, because equipment, process, quality-inspection, energy, supply-chain and after-sales data remain scattered across separate systems and cannot be used directly to train industry models. Citing Tianyancha, it says about 460,000 industrial data service companies exist in China, with roughly 68,000 new registrations since the start of 2026, and names industrial software, data governance, edge computing, industrial knowledge graphs and industry-model providers as beneficiaries, while general-purpose model vendors lacking industry data and on-site delivery capability come under pressure.