AI的下一个突破口在哪

Core Insights - The report indicates that AI is transitioning from the "tool era" to the "partner era" by 2025, with clearer development trends emerging for 2026 [1] Group 1: Infrastructure and Computing Power - The report emphasizes that computing power is the primary engine of the intelligent industry, with unprecedented global demand for AI computing power driving the construction of large-scale data centers [5] - Cloud computing vendors are shifting investments from traditional general computing resources to dedicated computing infrastructure that meets AI demands, evolving their partnerships with AI companies into deep strategic collaborations [5] - The demand for AI is reshaping chip innovation, with a significant market opening up for new opportunities in chip technology, challenging the dominance of GPUs and promoting the adoption of NPUs and custom chips [6] Group 2: Model Innovation - Pre-training determines the hierarchy of large models, while architectural innovation influences pre-training levels, with hybrid expert models becoming mainstream due to their cost-effective design [7] - By 2025, large models will enter a "reasoning time," with continuous breakthroughs in multi-modal deep reasoning and adaptive reasoning frameworks [7] Group 3: Embodied Intelligence and Applications - There is a growing trend in the research and development of physical AI and embodied intelligence, with significant orders being placed for humanoid robots, indicating a shift towards broader application exploration [8] - The report highlights that AI is reshaping traffic entry points, transitioning from "people finding services" to "services finding people," which is becoming the next generation of interaction paradigms [10] - The emergence of lightweight models and edge computing technologies is facilitating the proliferation of AI capabilities in devices such as smartphones, cars, and IoT devices, addressing key issues of data privacy and cost efficiency [10]