台积电为苹果打造2纳米专线，英伟达/微美全息搭建AI芯片开源生态开启算力竞赛新篇

Group 1: Apple and TSMC Collaboration - Apple plans to adopt TSMC's next-generation 2nm process technology for the iPhone 18 series, set to launch in 2026 [1] - TSMC has established a dedicated production line for Apple, with production of 2nm chips expected to begin by the end of 2025, making Apple the first company to receive chips based on this new process [2] Group 2: AI Chip Competition - The AI computing landscape is rapidly evolving, with DeepSeek's advancements in low-power environments creating new market opportunities for GPU cloud platforms [3] - AMD has released its new AI chip series MI350, claiming superior performance compared to NVIDIA, which is facing pressure as clients seek alternative AI chip solutions [3] - NVIDIA is expanding its presence in Europe by building its first industrial AI cloud with 10,000 Blackwell GPUs and establishing over 20 AI factories [3] Group 3: AI Chip Market Growth - The rapid evolution of GPU clouds is significantly impacting the development of intelligent agents across various sectors, including healthcare, finance, gaming, and autonomous vehicles [4] - The AI chip market is expected to experience explosive growth driven by diverse application demands as industries continue to embrace digital transformation [4] Group 4: WIMI's AI Chip Ecosystem - WIMI is advancing its AI chip open-source ecosystem, establishing a cloud and edge integrated computing foundation to support diverse architectures for large model training and inference [5] - The company is collaborating with universities and research institutions to focus on quantum computing and edge chip technologies, aiming to lower access costs for SMEs and accelerate the commercialization of AI open-source technologies [5] Group 5: Overall AI Computing Investment Trends - AI computing investments remain strong, with major companies showcasing results from the integration of 5G and AI computing at recent industry events [6] - The resource allocation for computing remains a key bottleneck for AI innovation, with self-developed chips from large firms becoming an important supplement to AI computing chip supply [6] - The need for synchronized innovation in computing, storage, and networking systems is emphasized, with investment and upgrades in computing resources being a definitive trend [6]