The 9th Digital China Summit opened on April 29, 2026, in Fuzhou, Fujian Province, where Yu Ying, Deputy Director of the National Data Administration, released the Digital China Development Report (2025). According to the report, China’s Digital China Development Index for 2025 reached 170.1, up 12.99% year on year. It also showed that China is now the world’s largest holder of AI patents, accounting for 60% of the total, that the number of high-quality datasets has expanded to more than 110,000 with a total volume exceeding 908 PB, and that the core AI industry has grown to more than 1.2 trillion yuan. In other words, this announcement was not merely a report on the progress of digitalization, but a demonstration to both domestic and international audiences of the depth of China’s national industrial base centered on AI.
The most striking figure in this report is, of course, the phrase “60% of AI patents.” This indicates that China has built up a very substantial quantitative expansion of research and development in the AI field. In fact, Stanford University’s 2025 AI Index also identifies China as the global leader in AI patents, and Reuters, citing WIPO-based data on generative AI patent trends, reported that China far outpaced the United States in the number of patent filings from 2014 to 2023. The exact definition of the “60%” figure presented by the Chinese side may not be completely identical to external statistics, but the broader point—that China is in a position to strategically emphasize its quantitative advantage in AI intellectual property—is consistent with international data.
That said, the important point here is that a large number of patents does not automatically mean leadership in cutting-edge models. In the same Stanford AI Index, the United States produced the largest number of “notable AI models” in 2024, with 40, while China produced 15. At the same time, Chinese players are also assessed as rapidly narrowing the gap with the United States on major benchmarks. Therefore, it is more accurate to read this news not as “China has already become the world leader in all aspects of AI,” but rather as “China has built depth in patents, academic papers, and social implementation, while also closing the gap with the United States in the race for cutting-edge models.”
What is more essential in this report, however, is not patents but the figures relating to data and operational infrastructure. The scale—more than 110,000 high-quality datasets, over 908 PB, and a cumulative 85 PB of data annotation—shows that China views AI competition not only as a contest in “model development,” but as a contest in building the data supply system that supports training, inference, and industrial deployment. Moreover, the report explicitly states that AI applications are becoming more advanced in the form of agent-based systems, and that token consumption is rising sharply as a result. This means that China itself has a very clear understanding that the axis of future competition is shifting away from mere research achievements toward massive volumes of data, continuously expanding inference demand, and the computing resources needed to process it.
Furthermore, what China has demonstrated is not just the “size of the AI industry.” The report states that the added value of the core industries of the digital economy exceeds 10.5% of GDP, that the number of internet healthcare users has reached 411 million, and that there are 306 national-level green computing facilities. These figures show that AI and data utilization have already penetrated deeply into public services, healthcare, education, and industrial operations. What emerges from this is that China is positioning AI not as an isolated advanced technology, but as a foundational technology for restructuring government, industry, and social services. The summit should therefore be seen as a venue for showcasing the breadth of that implementation phase.
What this news is presenting to the international community is the question: where is the real battleground of AI competition? Until now, attention has tended to focus on the race for high-performance semiconductors and the capabilities of foundation models. China’s latest message, however, is that true competitiveness lies in comprehensive strength: the ability to integrate intellectual property, data, the annotation industry, computing resources, and public-sector implementation. In fact, when WIPO and Stanford data are considered together, China is extremely strong in patents and research, while the United States still retains an advantage in cutting-edge models, though that gap is narrowing. In other words, AI supremacy is no longer just a contest over who can build the smartest model, but increasingly a contest over who can operate AI as the largest social system. This is a well-grounded assessment.
There are also clear implications for Japan in observing this trend. It is not enough to think about AI policy solely in terms of semiconductor support or R&D subsidies. Unless Japan takes an integrated approach—covering the development of high-quality data, institutional design for data sharing, the industrial base responsible for annotation and operations, and implementation in real-world settings such as healthcare and government—it will not be able to build up real competitiveness. China’s latest announcement is not only a news story boasting large numbers, but also a signal that AI competition has already entered a phase of “total competition.” What Japan should pay close attention to is not China’s headline-grabbing figures themselves, but the national-scale infrastructure development advancing behind them.