算力的"回弹效应"加速显现，释放算力服务新需求。通用人工智能技术的创新突破、广泛应用，将大幅提升算力使用效率，引发算力领域的"杰文斯悖论"， 带来"两个新变化"。一是算力总量的指数增长，未来3年，我国智能算力规模增长超2.5倍，年均复合增速近40%。二是算力结构的显著变化，推理算力需求 将超过训练算力需求，未来3年，推理算力年复合增速将达到训练算力的近4倍，到2028年，推理算力规模将超过训练算力规模。这一进程将给算网发展注入 新的动能，推动产业进入新一轮增长周期。 连接的"加速效应"不断拓展，催生信息消费新形态。高速、移动、安全、泛在的网络，推动"人机物"多元主体的全面连接，加速数算智的系统融入，引 发"比特×瓦特"的融合聚变，实现能力和价值的显著倍增，催生信息消费"新三样"等新业态新模式。一是AI智能终端，云边端算力、数据和模型在网络的带 动下加速整合贯通，显著提升AI智能体的性能，深度嵌入各类终端设备，打造能够提供专家级、个性化服务的"生活助理"，未来AI智能体数量将超过人的数 量。二是智能网联汽车，依托高性能算力和低时延网络，车路云协同不断深化，加速端到端智能驾驶的落地，拓展人类数智生活的"第四空间" ...

Core Insights - China Mobile's Chairman Yang Jie announced that the company's dispatchable computing resources account for one-sixth of the national total, with plans to further enhance cloud intelligence based on the computing network [1] - The company aims to focus on "four key directions" to build a comprehensive "AI full-stack service" that includes computing power, tools, models, and applications [1] Group 1: Infrastructure Development - China Mobile plans to construct ultra-large-scale "computing factories" to promote the exponential growth of tens of thousands and thousands of intelligent computing centers [1] - The company is also conducting research on a hundred thousand card intelligent computing center, aiming to establish a "centralized + distributed" inference computing resource system [1] Group 2: Intelligent Computing System - The goal is to create an intelligent computing system characterized by "centralized aggregation, ubiquitous edge, training and inference integration" [1]

4月10日，以"由云向智，共绘算网新生态"为主题的2025中国移动（600941）云智算大会在苏州举行。 会上，中国移动董事长杨杰提出，中国移动的可调度算力资源已经占全国的六分之一，未来还将在算力 网络的基础上进一步强化云智算。 《每日经济新闻》记者在现场注意到，杨杰还提到了信息消费的一些新变化。在他看来，未来AI智能 体的数量将超过人的数量；到2030年，智能网联汽车在新车中占比将超过80%；到2030年，80%的人类 将每天与智能机器人打交道，未来机器人数量也将超过人的数量。 在杨杰看来，随着以数智化为主要特征的新一轮科技革命和产业变革深入发展，数据成为新生产要素， 算力成为新基础设施和基础能源，人工智能成为新生产工具，推动经济社会从"互联网 +""5G+"向"AI+"加速转变，为算网基础设施演进带来新机遇、提出新要求。 "网络枢纽"方面，杨杰提出，未来中国移动还将加快天地一体、智能融合、通感一体等6G特性的现网 应用，强化800G和1.2T超高速下一代网络的试点落地，拓展低空网络、航空网络、卫星网络，支撑"人 机物"的全域连接。 此外，杨杰还提出要打造自助式工具平台（AI PaaS），以"人人都能够训练 ...

央视网消息：2025中关村论坛年会正在举行，作为年会的系列活动，"人工智能主题日"活动从3月27日持续至31日，共有2场核心论坛、5 场专场论坛和多场特色活动。其中未来人工智能先锋论坛是核心论坛，现场发布的"公共算力平台生态网络"备受嘉宾关注。 未来人工智能先锋论坛作为人工智能主题日的一场核心论坛，汇聚了图灵奖得主、政产学研等大量海内外学者。现场还围绕人工智能共发 布了7项成果，其中一项是北京人工智能公共算力平台生态网络正式启动了，这个网络将整合国内各地的算力资源，实现算力的优化配置，所 以备受嘉宾的关注和讨论。 记者从未来人工智能先锋论坛上了解到，目前，北京人工智能公共算力平台已有多家算力供应主体入驻，汇聚了京津冀、内蒙古、新疆、 宁夏等区域超8万P的绿色算力，构建了国内超大规模的跨域协同的智能算力网络。 2025中关村论坛年会·科技助残论坛：为汇聚助残科技创新力量搭建平台 2025中关村论坛年会正在举行，本届中关村论坛首设了残疾人领域论坛，聚焦科技助残，为汇聚更多助残科技创新力量搭建平台。 2025中关村论坛年会科技助残平行论坛主题为"科技有爱，共创美好世界"，多位国内外知名专家学者做主旨演讲、参与圆桌对 ...

Core Viewpoint - Anbotong Technology Co., Ltd. has established the "Luban" AI Research Institute to integrate security, AI, and computing power technologies, aiming to create a secure connection between AI and the world [1][2]. Group 1: Establishment and Purpose of the AI Research Institute - The "Luban" AI Research Institute was inaugurated in Shanghai, focusing on the integration of security and AI technologies in response to the rapid penetration of generative AI into industrial transformation [1]. - The establishment of the institute is timely, as it addresses new challenges and opportunities in network security brought about by advancements in generative AI and computing networks [1]. Group 2: Strategic Vision and Growth - Anbotong's chairman, Zhong Zhu, articulated the company's strategic shift from being a "visual network security innovator" to an "AI era security computing ecology builder," emphasizing the importance of embedding security into computing infrastructure [2]. - Since its listing, Anbotong has achieved a compound annual growth rate of 26% in revenue, highlighting its significant accomplishments in the network security sector [2]. Group 3: AI Delivery Architecture and Innovations - The "Luban" AI Research Institute introduced the "ESAiD" AI delivery architecture, which includes a three-tier AI delivery system focused on intelligent development, computing resource scheduling, and security protection [3]. - Anbotong launched a super silent liquid-cooled intelligent computing workstation, designed for various scenarios, featuring high-performance CPUs and AI acceleration cards [3]. Group 4: Future Directions and Collaborations - The research institute aims to gather top AI security talent, develop autonomous security models, and build an intelligent security ecosystem to support national digital security strategies [4]. - The event highlighted the importance of supply chain security, as emphasized by Jiangyuan Technology's vice president, underlining the need for self-controlled chips in the context of national security [3].