Core Insights - The report emphasizes that the competition for computing power in the AI era is fundamentally a competition for electricity, highlighting the critical role of electricity supply in supporting AI-driven growth in the U.S. [1] Group 1: AI's Impact on U.S. Electricity Demand - The overall electricity consumption growth in the U.S. appears slow, but electricity demand from data centers, driven by AI, is rapidly increasing [1] - Data centers currently account for 4.4% of total U.S. electricity consumption, with projections indicating this could rise to between 6.7% and 12% by 2028-2030, necessitating an additional power generation capacity of up to 100 gigawatts [1][14] - In 2023, U.S. data center energy consumption reached 176 terawatt-hours, a significant increase from 76 terawatt-hours in 2018 [14] Group 2: Current Electricity Supply Challenges - The U.S. electricity supply system is under significant strain due to limited new capacity in natural gas, insufficient interconnection capacity in the grid, and the concentration of data centers in a few states like Virginia, leading to localized grid pressure [1][3] - Natural gas is currently the primary source of electricity generation in the U.S., with a generation capacity of 207.2 GW, accounting for 43% of the total generation mix [20][23] Group 3: Solutions to Electricity Supply Issues - Short-term solutions include optimizing existing energy sources and rapid deployment of nuclear power, which is seen as an ideal stable and clean power source for data centers [2] - Gas-fired power generation is also a practical choice due to its short construction cycle and flexibility in meeting the rapidly growing electricity demand [2] - Long-term innovations such as small modular reactors (SMRs) and controlled nuclear fusion are being explored, with SMRs already in early project stages [2][3] Group 4: Future Energy Landscape - The energy transformation driven by AI demand is reshaping the competitive landscape of the electricity industry, focusing on stability, cleanliness, and innovative energy solutions [3] - The race for reliable and low-carbon energy sources is intensifying, with advancements in traditional energy optimization, nuclear technology upgrades, and exploration of fusion energy [3]

Core Viewpoint - The competition for computing power, particularly in AI model training, is intensifying, with the availability of graphics cards being a crucial factor in determining success in this arena [1][12]. Group 1: H20 Graphics Card Overview - The H20 graphics card, launched by NVIDIA at the end of 2023, is designed for AI model inference training and has a performance level of 10% to 15% compared to the previous flagship H100 [3]. - Despite the high price of 110,000 yuan (approximately 15.5 million USD), the H20 is currently the best-performing graphics card available in China, leading many tech companies to purchase significant quantities [4]. Group 2: Market Demand and Sales Dynamics - Major companies like ByteDance, Alibaba, and Tencent ordered at least 16 billion USD worth of H20 graphics cards in the first quarter of last year, with ByteDance reportedly stockpiling 100,000 GPU modules [5]. - The sales of H20 were initially banned in China on April 9, causing significant concern for NVIDIA, which prompted efforts to lift the ban [7]. Group 3: Competitive Landscape and Strategic Implications - NVIDIA's CEO is motivated to restore sales of the H20 due to the relatively low performance of the card compared to the latest GB200, which is estimated to be over 50 times more powerful [8]. - If NVIDIA fails to supply the H20, domestic competitors like Huawei's Ascend 910C, which is approaching the performance parameters of the H100, may fill the market gap [9]. - The resumption of H20 sales indicates that the race for computing power in the AI era is far from over, emphasizing the importance of graphics card availability in training superior AI models [10][13].