这些不仅是市场关注的问题，也是政策研究部门关注的问题。 2026年两会前夕，在国家信息中心的支持下，《财经》将2026年初的一次算力政策讨论与意见征集，梳理出了七个关键问题。这次国家信息中心组织的讨 论主要聚焦"算力调度"对产业发展的价值，参与者以电信运营商、科研院所、咨询机构的专家为主。 国家信息中心是国家发展改革委直属事业单位，长期进行全国一体化大数据中心、"东数西算"工程、全国一体化算力网等政策研究。多年来，国家信息中 心在不断组织大型科技公司、创业公司、咨询机构、科研院所等进行算力领域的专题讨论和意见征集。 在历经多年的长期研讨中，各方对"算力调度"的理解不断深化，并逐渐达成了一些共识： 算力调度，是充分利用现有算力，减少闲置资源，实现合理配置的重要方法。算力调度的复杂度，要远超水电的调度。水电是同质的物理资源，但算力是 高度异构和非标准化的。全国算力调度"一张网"的建设，不能只靠行政指令推进，而是通过政府搭台、标准引领、市场运作的综合模式，让全国一体化算 力网真正高效运转起来。 中国算力产业布局中，产业政策一直发挥着重要作用。其中有两大关键节点——2022年2月，"东数西算"工程启动。2023年12 ...

1、 李强主持召开国常会 研究促进有效投资政策措施 重要程度：★★★★★ 以下为晚报正文： 国务院总理李强2月6日主持召开国务院常务会议，听取2025年国务院部门办理全国人大代表建议和全国政协提案工作情况汇报，研究促进有效投资政策措 施，部署修订《环境空气质量标准》，讨论《中华人民共和国招标投标法(修订草案)》。 会议指出，促进有效投资对于稳定经济增长、增强发展后劲具有重要作用。要创新完善政策措施，加力提效用好中央预算内投资、超长期特别国债、地方政 府专项债券等资金和新型政策性金融工具。要结合制定实施"十五五"规划，着眼于长远发展需要和构筑未来竞争优势，在基础设施、城市更新、公共服务、 新兴产业和未来产业等重点领域，深入谋划推动一批重大项目、重大工程。 此外，李强主持召开国务院第十次全体会议，讨论拟提请十四届全国人大四次会议审议的政府工作报告稿和"十五五"规划纲要草案稿。 李强指出，宏观政策要靠前发力，财政资金尽可能提前安排，加强资金下达和项目建设的协同配合，使政策尽快落地见效。各项重点工作要抓紧推进，条件 成熟的及早组织实施。坚持政策支持和改革创新并举，更好激发市场活力，挖掘内需新增长点。要密切跟踪形势变化 ...

Group 1: Company Overview - The company has established a solid foundation in the telecommunications industry, focusing on empowering 30 operators and various enterprises to facilitate their digital transformation [1][2]. - The company aims for comprehensive breakthroughs and value growth through its development in the telecommunications sector [2]. Group 2: Competitive Strategy - The company's competitive strategy varies by industry; it adopts a direct sales model in the operator sector while emphasizing ecosystem partnerships in urban and digital economy sectors [2][3]. - In the face of competition, the company prioritizes collaboration to expand market opportunities rather than solely focusing on direct sales [3]. Group 3: International Expansion - The company has initiated international business expansion, starting from Shenzhen and targeting operators in Hong Kong as a launch point for further growth [3]. - The focus in international markets is on standardized products in AI and computing power [3]. Group 4: AI and Computing Power - The company integrates demand in the computing power sector, participating in national development rather than just focusing on supply-side solutions [3]. - AI applications are utilized to enhance operational efficiency and reduce labor costs through automation in code verification and digitalization of processes [3]. Group 5: Revenue and Orders - The company’s revenue is significantly dependent on operator orders, with expectations for growth in the second and third lines of business in the coming years [3]. - Specific order-related information will be available in the company's annual report [3].

你说荒诞不，在API调用动辄几十万、上百万token的时代， API选型居然变成了一件靠经验反复试错的事儿 。 这就导致想要接个API做开发，还得先被迫兼职下采购员。东市买骏马，西市买鞍鞯，必须把市面上的供应商挨个测一遍。 （写到这儿的时候，我的表情就是那个大家可以想象的痛苦面具闭眼表情包.jpg） 衡宇 发自 凹非寺 量子位 | 公众号 QbitAI 忍不了了，这个槽我真的不吐不快！ 比面对大模型黑盒更让人抓瞎的事情，就是要去选既靠谱、性价比又高的API服务 。 这几乎是每一个涉足AI应用开发的团队都会经历的至暗时刻，抹泪.gif。 同一个模型架构在不同的供应商手里，不仅价格上有出入，延迟、稳定性、吞吐量等用户关心的指标，波动幅度简直堪比霸天虎过山车。 不er，就没有一个工具能把这些API的底裤扒得干净，让咱开发者省点心吗？ 带着如此沉痛的心情跟周围人打听了一圈，你还真别说，有人告诉我有家 清华系的AI Infra公司——清程极智 ，真就做了这个事儿。 产品叫AI Ping，之前没做过什么宣发，基本一直就靠口碑口口相传。 用一句话来概括功能，可以说它就像 大模型API领域的大众点评 。 用7×24小时持续运 ...

针对这一痛点，佳杰云星依托自身技术积累，深度参与东莞人工智能大模型中心的建设与运营，以自主研发的算力调度与运营平台产品体系，提供了针对性 解决方案。该方案围绕大模型中心运营平台及算力调度平台搭建展开，通过四大关键举措破解行业难题。 精细化运营：同时，平台配备完善的运营管理模块，涵盖运营分析、费用核算、算力调度分配等核心功能，实现全流程精细化管控。 该平台自2024年启动实施以来，已取得显著成效，2025年正式上线后全面释放价值。 作为算力调度领域的创新实践，佳杰云星通过技术赋能与模式创新，不仅解决了城市级算力平台建设的核心痛点，更探索出算力资源市场化配置的有效路 径。 北京佳杰云星数据科技有限公司深耕信息技术软件领域，聚焦智能体开发、算力调度和管理、多云管理三大核心业务，凭借持续加码的研发投入与技术创 新，已成长为数字基础设施建设领域的中坚力量。 东莞作为制造业重镇与数字经济发展前沿城市，亟需构建城市级人工智能基础公共服务平台，为区域产业升级注入动能。但在平台建设过程中，东莞市数字 经济发展集团有限公司面临两大挑战： 挑战一，AI算力资源分布不均，众多企业存在迫切的算力需求缺口，亟需高效调度机制实现供需平衡； ...

Summary of the Conference Call for Hengwei Technology Company Overview - Hengwei Technology has acquired Shuheng Technology to enhance its AI application capabilities, addressing the revenue gap in AI applications compared to computational power in China [2][3] - Shuheng Technology focuses on marketing scenarios, providing marketing services through an AI platform, leveraging the founding team's extensive experience in big data and traffic investment [2] Core Business and Development Direction - Hengwei Technology's main business is divided into two segments: network visualization and intelligent system platforms, with a focus on AI infrastructure and computational products [3] - The company plans to seek acquisition targets in the AI application sector by late 2024 to early 2025, aiming to bridge the revenue gap in AI applications [3] Acquisition Rationale - Shuheng Technology was chosen for its ability to convert deep understanding of marketing scenarios into tangible revenue and profit [4] - The company operates on a results-driven business model, charging based on agreed KPIs, ensuring steady profit growth [4][15] Technological Development - Shuheng Technology has developed the SGPT model since 2020, emphasizing small parameter models to solve practical problems, with the local deployment of SGPT 1.0 completed in 2023 [2][6] - The company has built a complete technology stack from computational power to application layers, adapting various GPU types and creating a flexible computational scheduling platform [7] Marketing Services and AI Platform - The Zhixin AI platform integrates advanced AI technologies to enhance marketing efficiency, offering a comprehensive solution that includes a business process center and general Q&A functions [9] - The AI agents can interpret client needs and guide planners in creating precise marketing strategies, improving communication efficiency between sales and planning teams [10] Market Analysis and Data Utilization - Zhixin's market analysis combines official sources and online search capabilities, ensuring comprehensive and reliable data collection [11] - The company has accumulated significant digital assets, which provide a solid foundation for its marketing solutions [12] Client Structure and Industry Focus - Shuheng Technology is focusing on fast-moving consumer goods, beverages, and travel industries, with plans to expand into automotive, education, and telecom sectors [13] - The company aims to develop new AI products for various industries, expecting significant growth in 2026 [13] Capacity and Human Resource Management - The company is experiencing a capacity explosion starting in 2025, with AI empowering teams to improve efficiency without explosive personnel growth [14] Profitability and Future Trends - Despite ongoing audits, net profit is expected to remain stable due to technology reuse and efficiency improvements, with a strategy focused on rapid expansion into new industries [20] Synergies with Hengwei Technology - Collaboration between Hengwei and Shuheng includes integrating AI models with hardware products and jointly developing computational scheduling platforms, enhancing market penetration [21]

Core Insights - The article highlights the successful implementation of a dual-rate mixed network trial with 400G and 800G speeds in key computing power regions of China, particularly in the Guizhou-Guangzhou corridor, enhancing efficient interconnection and collaborative development of computing resources [1] - The "East Data West Computing" initiative has positioned Guizhou as a national hub for integrated computing networks, transitioning from a "data warehouse" to a "computing factory" [1][4] - Guizhou's digital economy is projected to exceed 100 billion yuan in software and information technology services revenue by 2024, with continuous growth rates among the highest in the country [4] Group 1: Computing Power Infrastructure - Guizhou has established a direct connection to 24 cities through a low-loss optical cable network, significantly reducing data transmission time by 33% and lowering costs by over 30% for businesses in the Guangdong-Hong Kong-Macau Greater Bay Area [2] - The computing power capacity in Guizhou has reached 150 billion billion calculations per second, with over 90% of intelligent computing being domestically sourced and regionally concentrated [4] Group 2: Efficient Computing Power Scheduling - The "Xirang" computing power scheduling platform has been launched in Guizhou, enabling nationwide coordination and scheduling of computing resources, akin to utilities like water and electricity [3] - Collaborations with local meteorological agencies have led to improved weather prediction accuracy and efficiency through the application of lightweight meteorological diffusion models [3] Group 3: Future Development and Industry Focus - Guizhou aims to further develop its computing power, data, applications, and industries, with a focus on intelligent computing, high-quality data aggregation, and artificial intelligence [5] - The establishment of an artificial intelligence laboratory in Guizhou is expected to attract more enterprises along the computing power industry chain, contributing to high-quality digital economic development [3][5]

Core Insights - The report highlights the emergence of AI infrastructure software (AI Infra) as a critical enabler for the deployment of generative AI applications, marking a golden development period for infrastructure software [1] - Unlike the model training phase dominated by tech giants, the inference and application deployment stages present new commercial opportunities for independent software vendors [1] - Key products in this space include computing scheduling software and data-related software, with computing scheduling capabilities directly impacting the profitability of model inference services [1][2] Computing Scheduling - AI Infra is designed to efficiently manage and optimize AI workloads, focusing on large-scale training and inference tasks [2] - Cost control is crucial in the context of a price war among domestic models, with Deepseek V3 pricing significantly lower than overseas counterparts [5] - Major companies like Huawei and Alibaba have developed advanced computing scheduling platforms that enhance resource utilization and reduce GPU requirements significantly [5][6] - For instance, Huawei's Flex:ai improves utilization by 30%, while Alibaba's Aegaeon reduces GPU usage by 82% through token-level dynamic scheduling [5][6] Profitability Analysis - The report indicates that optimizing computing scheduling can serve as a hidden lever for improving gross margins, with a potential increase from 52% to 80% in gross margin by enhancing single-card throughput [6] - The sensitivity analysis shows that a 10% improvement in throughput can lead to a gross margin increase of 2-7 percentage points [6] Vector Databases - The rise of RAG (Retrieval-Augmented Generation) technology has made vector databases a necessity for enterprises, with Gartner predicting a 68% adoption rate by 2025 [10] - Vector databases are essential for supporting high-speed retrieval of massive datasets, which is critical for RAG applications [10] - The demand for vector databases is expected to surge, driven by a tenfold increase in token consumption from API integrations with large models [11] Database Landscape - The data architecture is shifting from "analysis-first" to "real-time operations + analysis collaboration," emphasizing the need for low-latency processing [12][15] - MongoDB is positioned well in the market due to its low entry barriers and adaptability to unstructured data, with significant revenue growth projected [16] - Snowflake and Databricks are expanding their offerings to include full-stack tools, with both companies reporting substantial revenue growth and customer retention rates [17] Storage Architecture - The transition to real-time AI inference is reshaping storage architecture, with a focus on reducing IO latency [18] - NVIDIA's SCADA solution demonstrates significant improvements in IO scheduling efficiency, highlighting the importance of storage performance in AI applications [18][19]

Core Viewpoint - The "Electric Power Heterogeneous Intelligent Scheduling Technology" developed by Nanjing Nari Ruijun Technology Co., Ltd., a subsidiary of State Grid NARI Group, has achieved international leading standards, effectively addressing the power industry's computing resource supply-demand contradiction [1][2]. Group 1: Technology Development - The technology enables efficient collaboration of heterogeneous computing resources from different brands and models, overcoming the "computing island" phenomenon where high-end resources are over-utilized while mid-to-low-end resources remain idle [1]. - The team has innovated a series of technologies to achieve "interconnectivity" of heterogeneous computing resources, including a unified interface for management and a "network + computing" collaborative mechanism [1]. Group 2: Application and Performance - The "Ruiteng Intelligent Computing Scheduling Platform" has demonstrated outstanding performance, with an average work order response time of only 7.241 seconds and an increase in concurrent processing capacity from 40 to 800, effectively doubling the efficiency of grassroots business processing [2]. - The platform is currently operational in 11 provincial power companies, including those in Jiangsu and Shandong, and is gradually being promoted nationwide, with plans to expand into military, telecommunications, and public security sectors [2]. Group 3: Achievements and Future Plans - The project has secured 19 patent authorizations, published 19 high-level papers, and led the formulation of 3 national standards, with core technologies being industry-first innovations [2]. - The team aims to continuously optimize the technology system to create a self-controllable intelligent computing foundation platform, empowering more industries in their digital transformation and contributing to high-quality development of the digital economy [2].

Core Insights - The Future Network Experimental Facility, China's first major national technology infrastructure in the information and communication sector, has officially commenced operations [1] Group 1: Facility Overview - The facility is located in Nanjing, Jiangsu, and was completed in August 2024 [1] - It covers 40 cities nationwide, featuring 88 backbone network nodes and 133 edge network nodes, with a total optical transmission length exceeding 55,000 kilometers [1] - The facility can support 4,096 heterogeneous services for parallel testing and is capable of interconnecting with existing domestic and international networks [1] Group 2: Performance Metrics - The facility enables efficient, high-speed, low-latency, and low-jitter data transmission, with a packet loss rate of only one in a million [1] Group 3: Service and Collaboration - To date, the facility has served major national research institutions such as the National Astronomical Observatory and the Institute of High Energy Physics, as well as telecom operators like China Telecom, China Mobile, China Unicom, and China Broadcasting Network [1] - It has collaborated with universities including Peking University, Nanjing University, Zhejiang University, and the Chinese University of Hong Kong, along with leading companies like Huawei, H3C, and Baidu, completing 120 significant innovation experiments [1] - The experiments cover critical dimensions such as core chips, network operating systems, routing control, security and trust, large-scale networking, and new AI services [1]