China’s National Data Administration released a draft plan on June 9 to expand the supply, circulation, and commercialization of high-quality training datasets, explicitly admitting the country lacks enough data for its AI ambitions. The plan spans nine core industrial sectors, including healthcare, finance, transportation, and autonomous driving, and calls for multimodal datasets covering text, code, images, audio, and video. The target deadline is 2028, aiming for a comprehensive ecosystem of multimodal datasets ready for commercial use. Alongside this, the Chinese government is investing approximately $295 billion, roughly 2 trillion yuan, into AI-focused data centers over the next five years. The plan also addresses data governance, management standards, and rights frameworks around training data. This initiative is part of China’s broader “AI Plus” directive, issued in August 2025, to embed AI into the national economy, serving as its infrastructure component. US restrictions on semiconductor exports have pushed Beijing to prioritize domestic, sector-specific datasets to insulate AI development from external supply shocks. Global estimates suggest publicly available human-generated data could shrink dramatically by 2032, with research groups warning that AI training is consuming internet data faster than humans produce it, potentially creating bottlenecks within the next few years. The data center investment compounds this dynamic, as China builds both data and computational infrastructure to create a vertically integrated AI development stack.