ITmedia Reports SUGENA’s Mapping of 15 Chinese Physical AI Companies and Five Major Hubs

On September 17, 2026, ITmedia published an article mapping China’s physical AI sector through 15 leading companies and five major hubs. The article also noted that SUGENA published an online seminar, “The Second China Physical AI Company Map 2026,” on August 27.

The mapping offers a way to understand the geography of physical AI and the connections between stages of the process, which can be difficult to see from individual product announcements alone. While reports have covered UBTECH opening factories and beginning mass production, a regional map makes it easier to see which metropolitan areas concentrate development and manufacturing.

This report is notable not for covering individual robot products or one-off announcements, but for providing an overview of China’s physical AI industry through its companies and regional clusters. The focus shifts from “which company released what” to “which companies are located in which regions and what stages of the process they handle.” The central issue, therefore, is not competition over performance, but how to map the locations connecting development, manufacturing, exhibitions and distribution. Among this newspaper’s related coverage, UBTECH opens robot factory in Liuzhou and UBTECH begins operations at humanoid mass-production factory showed that individual companies’ factory investments were taking shape. A regional-map approach such as this one, by contrast, supplements that coverage by offering a way to view individual investments not as isolated events, but as parts of supply networks within industrial clusters. Together with the expansion of exhibitions of embodied intelligence at the Bund Conference, what emerges is not only an increase in the number of products, but the formation of pathways linking exhibitions to factories and factories to mass production. The implications for industry structure include, first, that competitive strength may depend not only on the technical capabilities of individual companies, but also on which metropolitan areas concentrate research, prototyping, mass production and exhibitions. Second, amid continued reports of new mass-production factories and production launches, it is difficult to compare company announcements without a picture of supporting functions such as parts supply, assembly and testing. Finally, as SUGENA’s mapping suggests, tracking the flow of data and robots by region may become as important as product specifications themselves. However, the central point of the original source is the fact that the sector has been mapped; the source text alone does not establish the relative merits of individual companies or a ranking of regions.

For Japanese companies, it will be necessary to assess Chinese competitors on the basis of both production sites and regional clusters, not just the performance of individual robots. In particular, it will be difficult to grasp the full picture by comparing products alone without identifying where companies can differentiate themselves—in the robots themselves, components or factory operations.