Core观点 - The AI spending boom presents significant investment opportunities, particularly in infrastructure buildout, with predictions of AI infrastructure spending exceeding $4 trillion by the end of the decade [2]. 分组1: CoreWeave - CoreWeave provides essential capacity for AI workloads, operating over 250,000 GPUs across 32 data centers, specifically designed for AI, resulting in enhanced speed and efficiency [4]. - The company's revenue has tripled in recent quarters, indicating strong demand, and it is positioned to continue delivering top GPUs due to its partnership with Nvidia, which holds about 7% of CoreWeave [5]. - CoreWeave shares have increased over 200% since their March IPO, suggesting strong potential for future growth as the demand for AI training and inferencing rises [6]. 分组2: Broadcom - Broadcom specializes in networking, offering high-performance switches and routers, as well as custom AI accelerators (XPUs) that are gaining traction among cloud service providers [8]. - The company's AI revenue has grown significantly, with a 63% increase to $5.2 billion in the recent quarter, and it anticipates $6.2 billion in the upcoming quarter, driven by the expanding AI infrastructure [10]. - Broadcom's niche in providing specific chips complements Nvidia's offerings, allowing customers to utilize both companies' products effectively [9]. 分组3: Taiwan Semiconductor Manufacturing - Taiwan Semiconductor Manufacturing (TSMC) is crucial for the production of chips designed by companies like Nvidia, benefiting from the overall demand for AI innovations [11]. - TSMC's growth is expected to accelerate due to the increasing AI infrastructure spending, as it plays a key role in manufacturing the chips necessary for AI platforms [13]. - Currently trading at 28x forward earnings estimates, TSMC is viewed as a strong long-term investment opportunity as AI infrastructure spending continues to rise [15].