Workflow
北京数字经济算力中心
icon
Search documents
从旧工厂到新地标,北京数字经济算力中心“绿意浓”
Cai Jing Wang· 2025-06-30 06:23
李辉表示,算力中心创新采用工业厂房低碳改造模式,应用AHU间接蒸发冷系统、多联热相变冷却系 统、液冷系统、余热回收等高效节能的绿色技术,外立面采用光伏一体化(BIPV)立面设计,并运用 AI赋能的运营管理平台和数字孪生系统,打造技术赋能与可持续发展融合的新型数字基础设施典范, 将整体能耗降低40%以上。 "我们应该是国内第一家把三种不同的先进节能技术融合到一个建筑里的算力中心,非常有挑战性,未 来我们还将进一步打造零碳产业园区。"李辉说。 "PUE(能源使用效率)值是衡量算力中心低碳化水平的核心指标,当前中心的PUE值仅为1.146,这一 数值远低于行业平均水平,处于领先地位。"近日,在"京彩不设限·经济热力站"活动上,北电数智智算 中心部李辉详细介绍了该算力中心的建筑减排和能源循环利用情况。 眼下,AI技术的快速发展导致算力需求呈指数级增长,尤其是在计算密集型任务时,算力基础设施的 用电量随之激增。当全球数字经济与"双碳"目标深度交织,算力"底色"的绿色化已成必然。作为北京 市"三个100"重点项目,北京数字经济算力中心通过创新探索,为全国绿色智算中心的建设和发展提 供"北京经验"。 在北京朝阳区酒仙桥的核 ...
北京数字经济算力中心开放预约,商业核心区优质算力赋能更快产业创新
Huan Qiu Wang Zi Xun· 2025-06-23 09:19
Core Insights - The rapid development of artificial intelligence (AI) technology has made computing power a key area for competition globally, with Beijing's Digital Economy Computing Power Center officially opening for reservations, providing strong momentum for the capital's AI industry [1][3] Group 1: Infrastructure and Capabilities - The Beijing Digital Economy Computing Power Center has deployed 1000P mixed computing power, making it the only commercial AI base in Beijing's business core to exceed 1000P, capable of meeting large-scale and complex AI computing demands [3] - The center is strategically located in Jiu Xian Qiao, Beijing, allowing it to serve as a high ground for AI innovation experiments, facilitating rapid integration of surrounding research and industrial resources for innovation [3] - The center features a super-large cluster configuration with tens of thousands of CPU/GPU cores and EB-level storage, improving computing efficiency by 30% and ensuring stable operations with high reliability architecture [4] Group 2: Technological Advancements - The center has achieved a high level of data security and business continuity with a network availability of 99.9999%, and it has implemented a green energy strategy that reduces energy consumption by 30% [4] - It integrates cutting-edge technologies such as 5G, AI, blockchain, and edge computing, accelerating the commercialization of these technologies [4] Group 3: Service Model and Ecosystem - The center offers a flexible, low-cost token-based billing model for AI computing services, lowering the entry barrier for small and medium enterprises to access powerful computing resources [5] - It aims to reconstruct the value chain from data to decision-making, enhancing data value extraction and application capabilities [5] Group 4: Economic Impact - The Beijing Digital Economy Computing Power Center not only supports the development of Beijing's AI industry but also attracts high-quality enterprises and innovative resources, promoting deep integration of AI technology with the real economy [7] - This initiative is expected to help Beijing gain a competitive edge in the global AI industry and contribute to the high-quality development of China's digital economy [7]
脑机接口、AI底座,多项“首创”硬科技亮相中关村
Group 1: DPU and AI Infrastructure - The Data Processing Unit (DPU) is recognized as the "third main chip" following CPU and GPU, with the latest third-generation DPU chip K2-Pro showcased at the Zhongguancun Forum [3] - The K2-Pro chip is designed to address challenges in large-scale data centers, achieving network throughput of 25G and 100G, filling a gap in the high-performance network card market [3] - The rapid development of AI, 5G, and the metaverse necessitates innovative computing solutions, with heterogeneous computing (DPU + CPU + GPU) becoming mainstream for next-generation intelligent computing centers [3] Group 2: AI Industry Development - The "Artificial Intelligence +" exhibition area highlighted the importance of autonomous innovation in computing infrastructure for AI industry growth, featuring products from GPU company Moore Threads [4] - The "Kua'e" system-level computing solution is tailored for large model training and inference, meeting the core demands of the large model era [4] - The Beijing Digital Economy Computing Power Center, a key project, integrates thousands of P-level domestic computing resources and provides various services for AI industry development [6] Group 3: Medical Innovations - The first non-invasive multi-channel brain oxygen monitoring device developed by Zhongke Boyue can monitor eight brain regions and is expected to enter clinical use in 2026 [7] - The Lattice NEXT blood flow-directed mesh stent, developed by Aike Medical Devices, expands the treatment range for intracranial aneurysms and has been recognized as an innovative medical device [9]