半导体行业月报：豆包AI生态加速发展，关注国内AI算力产业链

Investment Rating - The report maintains an "Outperform" rating for the semiconductor industry [1]. Core Insights - The domestic AI computing power industry chain is accelerating, driven by the rapid growth of the Doubao AI model, which has seen a daily token call volume increase from 120 billion in May 2024 to over 4 trillion by December 2024, representing a growth of over 33 times in just seven months [6][29]. - The semiconductor industry is currently in an upward cycle, with AI being a significant driving force for growth [6][29]. Summary by Sections 1. December 2024 Semiconductor Market Performance - The domestic semiconductor industry saw a decline of 0.29% in December 2024, underperforming the 0.47% increase in the CSI 300 index [11]. - Year-to-date, the semiconductor industry has risen by 24.63% [11]. 2. Global Semiconductor Monthly Sales Growth - Global semiconductor sales in November 2024 increased by 20.7% year-on-year, marking the 13th consecutive month of growth [22]. - The sales figure for November 2024 was approximately $578 billion, with a month-on-month increase of 1.6% [22]. - The forecast for the global semiconductor market in 2024 has been raised to a 19% year-on-year growth, with expectations of a 12.5% growth in 2025 [25][26]. 3. AI and Consumer Electronics Demand - Consumer demand is gradually recovering, with global smartphone shipments in Q3 2024 increasing by 5% year-on-year [6]. - The AI smartphone and AI PC penetration rates are expected to rise rapidly, contributing to the recovery of global PC shipments [6][29]. 4. Inventory and Capacity Utilization - Global semiconductor equipment sales in Q3 2024 grew by 19% year-on-year, with Chinese semiconductor equipment sales increasing by 17% [6]. - The inventory levels of some domestic chip manufacturers continued to decline in Q3 2024, indicating ongoing improvements [6]. 5. Investment Recommendations - The report suggests focusing on investment opportunities within the AI computing power supply chain, including AI computing chips, copper connections, PCBs, server power supplies, and liquid cooling systems [6][29].