伏击“物理短板”，鹤禧投资的科技方法论

Group 1 - The core logic of technology investment is based on "first principles," emphasizing that the performance of AI systems is determined by the shortest supply chain segments rather than the most powerful components [2] - The investment strategy focuses on identifying bottleneck segments in the supply chain, particularly in storage, as demand for storage is expected to grow significantly due to the increasing size of AI models and their requirements [4][12] - The investment approach is characterized by a shift from macro capacity considerations to micro-level KPIs and contract details, highlighting the importance of operational efficiency and management practices in the supply chain [5][6] Group 2 - The storage sector is identified as a critical bottleneck in the AI supply chain, with demand growth outpacing that of GPUs, driven by the evolution of multi-modal models and increased data throughput [8][11] - The supply constraints in the storage market are attributed to high capital expenditure requirements and technological barriers, making it difficult for new capacity to meet the surging demand [14][16] - The expected structural changes in the storage market indicate that by 2026, over half of storage procurement will come from AI data centers, with a significant shift towards customized and long-term contracts [15][16] Group 3 - The investment philosophy emphasizes the importance of understanding industry dynamics and avoiding speculative behaviors based on price movements, advocating for a focus on fundamental industry truths [7][12] - The narrative economy is highlighted as a significant factor influencing asset pricing, where trends and narratives can drive valuations faster than actual earnings growth [27][28] - The competitive landscape in the AI sector is evolving, with companies like Google and Nvidia driving demand for optical devices and storage, indicating a shift in value distribution within the supply chain [23][29]