Group 1: AI Computing Power - The global infrastructure wave is driving significant growth in AI computing power, with major cloud service providers (CSPs) expected to increase capital expenditures (CapEx) to over $600 billion in 2026, representing a 40% year-on-year increase [5][24][25] - NVIDIA anticipates that the total shipment of Blackwell and Rubin GPUs will reach 20 million units by the end of 2026, generating approximately $500 billion in sales [31][33] - The demand for AI servers is expected to maintain high growth, with shipments projected to rise from 1.6 million units in 2024 to 2.4 million units in 2026 [25][28] Group 2: AI Storage - The storage industry is entering a "super cycle" due to a rebound in demand driven by AI applications, with DRAM and NAND Flash prices expected to rise significantly [52][64] - Major storage manufacturers have successfully reversed the supply-demand imbalance through strict production control, leading to a clear upward trend in prices for both DRAM and NAND Flash [52][64] - The demand for high-capacity SSDs is rapidly increasing, particularly in AI training, which is accelerating the replacement of traditional HDDs with QLC SSDs [7][52] Group 3: End-Side AI - AI is reshaping the hardware landscape for smart terminals, with AI smartphone penetration expected to rise from approximately 18% in 2024 to 45% in 2026 [7][19] - The success of AI glasses, such as Ray-Ban Meta, indicates a growing market for AI-integrated wearable technology, with significant sales growth anticipated in 2026 [7][20] - The development of humanoid robots is gaining momentum, with traditional consumer electronics manufacturers entering the robotics supply chain, driven by advancements in AI models [7][21] Group 4: Investment Recommendations - Recommended companies in the AI computing power sector include Industrial Fulian, Huadian Technology, and Shenghong Technology for overseas chains, and Cambrian, Chipone, and SMIC for domestic chains [7][8] - In the AI storage sector, companies like Demingli, Jiangbolong, and Baiwei Storage are highlighted, along with niche players such as Zhaoyi Innovation and Beijing Junzheng [7][8] - For end-side AI, recommended companies include Rockchip, Lexin Technology, and Lens Technology, focusing on SoC and consumer electronics [7][8]
电子行业2026年度策略报告:云端共振,算存齐飞-20251202