综合算力大数据：河北江苏广东居前三

Group 1 - The core viewpoint of the articles highlights the rapid growth of China's artificial intelligence (AI) computing power market, which is expected to reach $25.9 billion by 2025, reflecting a year-on-year increase of 36.2% [3][1] - The Ministry of Industry and Information Technology and other departments have issued guidelines to strengthen AI computing power supply and promote the construction of a national integrated computing power network [1][22] - The "Blue Book" on comprehensive computing power index indicates that while progress has been made in China's computing power ecosystem, challenges such as uneven infrastructure and insufficient application depth remain [1][2] Group 2 - As of mid-2025, China has released 1,509 large models, accounting for approximately 40.2% of the global total of 3,755 [2] - The number of standard racks in data centers in China has increased from 3.15 million in 2019 to over 9 million by 2024, with a total of 10.85 million racks in use by mid-2025 [2][9] - The top provinces for standard racks in use are Hebei, Guangdong, and Jiangsu, each exceeding 600,000 racks [2] Group 3 - The global AI server market is projected to grow from $125.1 billion in 2024 to $158.7 billion in 2025, with generative AI servers expected to increase their market share from 29.6% in 2025 to 37.7% by 2028 [3] - The comprehensive computing power index ranks Hebei, Jiangsu, and Guangdong as the top three provinces, while cities like Langfang, Zhangjiakou, and Dalian lead in computing power sub-index [14][21] Group 4 - The national strategy to accelerate the construction of an integrated computing power network is leading to a more extensive distribution of computing resources beyond core regions [21] - The average Power Usage Effectiveness (PUE) of computing centers in China is 1.46, indicating progress in energy efficiency, although there is still significant room for improvement [21] - The implementation of the "Opinions" encourages the use of cloud computing services to quickly build intelligent infrastructure capabilities, reducing technology investment costs [22]