启明创投邝子平对话千里科技印奇：商业模式闭环才能持续推动技术进步，AI时代硬件机会巨大

Core Viewpoint - The 2025 World Artificial Intelligence Conference (WAIC) emphasizes the integration of AI technology with various applications, highlighting the potential for significant advancements in the AI+terminal sector over the next three years [2][6]. Group 1: AI+Terminal Trends - The next three years are expected to be pivotal for the AI+terminal sector, particularly in the automotive and mobile phone industries, with significant advancements anticipated in human-machine interaction and application integration [6][8]. - The automotive industry is entering a critical phase, marking the tenth year of smart driving in China, with expectations of substantial changes in technology and product offerings [7][8]. Group 2: Model Evolution - Models are identified as the driving force behind the evolution of the AI industry, with a clear progression in learning paradigms from imitation learning to reinforcement learning and eventually to autonomous learning [9][11]. - The current state of multi-modal models in China is approximately six months behind that of the U.S., although the gap in computational power is widening, indicating a divergence in technological advancement [14][15]. Group 3: Business Model Insights - A sustainable business model must be closed-loop to effectively drive technological progress, emphasizing the importance of integrating technology, product, and commercialization from the outset [16][19]. - Companies focusing on foundational large models must justify significant investments in computational resources, as the long-term profitability of such models is critical for business viability [19]. Group 4: Hardware Opportunities - There are substantial opportunities in hardware beyond just vehicles and smartphones, with a focus on delivering comprehensive AI services through innovative hardware solutions [20][22]. - The future of AI services and operating systems is expected to undergo significant changes, with a shift towards human-machine collaboration and the emergence of new hardware forms driven by AI agents [24][25].