从算力到端侧：AI硬件景气上行 端侧AI成2026年主线

Group 1 - The demand for AI computing power is continuously rising, driven by the needs of AI inference and related hardware supply-demand imbalances [1] - In the semiconductor upstream foundry and packaging sectors, foundries are raising prices for certain mature process foundry services due to increased AI-related power demand and production cuts by major manufacturers [2] - The storage industry is projected to reach a value of $551.6 billion by 2026, with a year-on-year growth of 134%, driven by AI [2] Group 2 - Domestic computing power hardware is making significant technological breakthroughs, with companies like Cambrian, Haiguang Information, and Moore Threads emerging in the computing chip sector [3] - In the packaging field, domestic manufacturers are advancing in advanced packaging technologies, with Changdian Technology achieving important progress in optoelectronic packaging [3] - AI is expected to accelerate its integration into various hardware products and industrial scenarios by 2026, creating investment opportunities across the industry chain [4] Group 3 - Investment targets in AI computing-related hardware include foundry manufacturers like SMIC and Hua Hong Semiconductor, and semiconductor equipment companies such as Zhongwei Company and Northern Huachuang [5] - In the end-user storage sector, companies like Zhaoyi Innovation and Baiwei Storage are highlighted, along with terminal manufacturers such as Hikvision, Lenovo Group, and Xiaomi Group [6]