Next stop for China's domestically produced AI: Chips and models running in both directions under the high walls of the ecosystem.

Recently, with companies such as Smart Universe, MiniMax, TianSuan Smart Core, and BiRen Technology landing on the Hong Kong Stock Exchange and the Science and Innovation Board, the Chinese AI industry has officially entered a new stage of commercial validation and large-scale application. However, in the industrial reality, under the high walls of the ecosystem built by NVIDIA, the "bottleneck" dilemma faced by domestic chips still exists. Some listed GPU companies have experienced significant corrections after a sharp rise in stock prices, reflecting to some extent the market's scrutiny of their commercialization paths and long-term growth logic. Since domestic chips are difficult to catch up with NVIDIA in absolute computing power in the short term, they seek to surpass from the perspective of system efficiency and scenario fit. In the recent releases of chip companies and large-scale enterprises, emphasis has been placed on "domestic adaptation," that is, through joint optimization to improve computing power utilization efficiency, accelerating the application of large models in various industry scenarios. It is generally believed in the industry that breakthroughs in individual technologies are not enough to win this competition, and ecosystem synergy, especially the "bidirectional running" of models and chips, is becoming the key to whether domestic AI can truly be independent.