Zhi Tong Cai Jing·2026-01-29 08:00Group 1 - The commercialization of AI applications has fully commenced, marking a new wave of large model innovation in China and the US [1] - Gemini 3 has unlocked the upper limits of large models, transitioning AI applications from "usable" to "practical," with Google establishing a mature "computing model-application" internal cycle that serves as a blueprint for global AI industry development [1] - Chinese AI applications are experiencing rapid growth in overseas revenue and monthly active users, with notable applications like 可灵, PLAUD, and HeyGen achieving nearly $100 million in ARR, covering various sectors such as video generation and article summarization [1] Group 2 - Large model manufacturers are upgrading their model capabilities and penetrating C-end scenarios, while terminal hardware manufacturers are developing native AI operating systems to reshape the core operational logic of AI smartphones [2] - The super application ecosystem is maximizing user value by leveraging hundreds of millions of MAU traffic, high-value user data, and vertical scene know-how, which may become core barriers for C-end manufacturers [2] Group 3 - The global AI computing power leaders are transitioning from training to inference, with domestic large model inference efficiency improvements and rapid growth in core application user numbers acting as catalysts for this shift [3] - The construction of super nodes and software ecosystems will lay a crucial foundation for the long-term rise of domestic computing power [3] Group 4 - Breakthroughs in the launch and recovery sectors of domestic low-orbit satellite networking are significant, with fields like aerospace computing expected to benefit in the long term [4] - Quantum technology is anticipated to become an important engine for the development of new productive forces [4]