启明创投发布2026 AI十大展望：基础模型、具身智能与AI基建将迎来关键突破

Odaily Planet Daily News Qiming Venture Partners released the "2026 Qiming Venture Partners AI Top 10 Outlook," focusing on four major directions: foundation models, embodied intelligence, AI infrastructure, and AI applications. It predicts the development trends of the artificial intelligence industry over the next 12 to 24 months. In terms of AI infrastructure, the demand for AI computing power is shifting from the training phase to the inference phase, and areas such as storage, advanced processes, and advanced packaging will continue to face structural shortages. Over the next two years, computing power asset reserves will become a crucial strategic resource for AI companies, and competition in AI infrastructure will also upgrade from competition over individual chips to system-level competition encompassing chips, interconnectivity, cooling, and power supply. In the next 12 to 24 months, the commercialization of AI applications will focus more on vertical industries and high-paying users. Applications that improve enterprise efficiency will be the first to take off, and as token costs decrease and interaction models innovate, phenomenon-level AI consumer applications are also expected to gradually emerge.

Furthermore, Qiming Venture Partners anticipates that AI-Native organizations will move from concept to practice, with some companies expected to achieve several times the per capita productivity improvement of traditional organizations.