Core Insights - The company reported that its own GPU computing resource utilization rate is between 85% and 90% [1] Group 1: Company Operations - The company introduced its computing resource construction and operation cooperation models, which include external purchase and co-construction models [1] - The company has tested and validated the integration of various types of domestic chip computing resources through a service integration platform, providing related technical services [1] Group 2: Strategic Partnerships - The company disclosed its strategic cooperation with Alibaba Cloud, which includes co-building a domestic computing center, jointly operating a public cloud area, and promoting industry large models together [1]

Group 1: Investor Relations Activities - The company conducted a specific investor survey on July 29, 2025, at the company meeting room, attended by Zhongyou Securities and the board secretary, Mr. Shi Jianwei [3][4]. Group 2: Computing Resource Acquisition Models - The company acquires computing resources through two main models: external purchase agreements with upstream suppliers and a co-construction model where the company builds its own computing clusters [4]. - The co-construction model allows the company to rent and use data center facilities to provide proprietary computing resources to downstream clients [4]. Group 3: Domestic Computing Resource Utilization - The company’s domestic computing resource usage is categorized into several application scenarios, with the highest demand for "basic large model training" and "mainstream large model inference" [5]. - The company has integrated various domestic chip computing resources into its platform, supported by a team of engineers for installation and optimization [5]. Group 4: Computing Resource Utilization Rate - The utilization rate of the company's proprietary GPU computing resources is currently between 85% and 90% [6]. Group 5: Strategic Cooperation with Alibaba Cloud - The strategic cooperation with Alibaba Cloud focuses on three main areas: building domestic computing centers, co-developing public cloud services, and promoting industry large models [7][8]. - The partnership aims to create a closed-loop system of "co-construction, sharing, and promotion" to enhance resource efficiency and commercial value [8]. Group 6: Shareholder Reduction Plans - The major shareholder, Qingkong Yinxing Nantong Venture Capital Fund, has completed its planned share reduction, and future decisions on further reductions will depend on various factors including market conditions [9]. Group 7: Gross Margin Decline in Intelligent Computing Cloud Business - The gross margin of the intelligent computing cloud business has declined due to factors such as product demand differences, resource allocation, and market competition [10]. - The revenue share from general cloud and industry cloud services has decreased, while the share from intelligent computing cloud services has increased, affecting overall gross margin [10].

中国国际金融股份有限公司 关于北京并行科技股份有限公司 募投项目结项的核查意见 中国国际金融股份有限公司（以下简称"保荐机构"）作为北京并行科技股份有限 公司（以下简称"并行科技"或"公司"）向不特定合格投资者公开发行股票并在北京 证券交易所上市的保荐机构，对并行科技履行持续督导义务。根据《证券发行上市保荐 业务管理办法》《北京证券交易所股票上市规则》《北京证券交易所上市公司持续监管 办法（试行）》《北京证券交易所证券发行上市保荐业务管理细则》等相关法律、法规 和规范性文件的规定，对并行科技募投项目结项的事项进行了认真、审慎核查，具体情 况如下： 二、募集资金管理情况 万股、33.40 万股及 6.00 万股，合计 690.40 万股股份，预计募集资金不超过 20,712.00 万元。根据立信会计师事务所（特殊普通合伙）出具的信会师报字[2021]第 ZB11551 号《验资报告》确认，截至 2021 年 12 月 10 日止，公司已收到中小企业发展基金（江 苏南通有限合伙）、吕大龙、宁波卓辉创曜股权投资合伙企业（有限合伙）、北京兴 健投资发展中心（有限合伙）、徐放及北京君利联合创业投资合伙企业（有限合伙） ...

证券代码：839493 证券简称：并行科技 公告编号：2025-106 北京并行科技股份有限公司 投资者关系活动记录表 本公司及董事会全体成员保证公告内容的真实、准确和完整，没有虚假记载、 误导性陈述或者重大遗漏，并对其内容的真实性、准确性和完整性承担个别及连 带法律责任。 一、 投资者关系活动类别 □业绩说明会 □媒体采访 □现场参观 □新闻发布会 □分析师会议 □路演活动 □其他 与上游算力供应商（各类超算/智算中心、电信运营商及其他厂商）签订算力资 源采购协议，通过对接调试将其算力资源纳入公司超算云平台，并根据下游客户 的需求量提供服务的形式；共建模式是指公司采购算力服务器、存储及网络设施 等，租赁和使用数据中心相关设施形成计算集群，并为下游客户提供公司自有算 力的形式。 √特定对象调研 二、 投资者关系活动情况 活动时间：2025 年 7 月 29 日 活动地点：公司会议室 参会单位及人员：中邮证券 上市公司接待人员：董事会秘书师健伟先生 三、 投资者关系活动主要内容 公司就投资者普遍关注的问题进行沟通交流，主要问题及回复情况如下： 问题 1、请介绍公司算力资源建设运营合作模式 回答：公司主要通过共建 ...

证券代码：839493 证券简称：并行科技 公告编号：2025-105 北京并行科技股份有限公司 关于募集资金投资项目结项并将 节余募集资金永久补充流动资金的公告 本公司及董事会全体成员保证公告内容的真实、准确和完整，没有虚假 记载、误导性陈述或者重大遗漏，并对其内容的真实性、准确性和完整性承 担个别及连带法律责任。 北京并行科技股份有限公司（以下简称"公司"）2021年第二次定向发行股 票募集资金投资项目及 2023 年向不特定合格投资者公开发行股票募集资金投资 项目之"超算云算力网络平台建设项目"（以下统称"募投项目"）已达到预定 可使用状态，公司将该募投项目予以结项，具体情况公告如下： 一、募集资金基本情况 公司自新三板挂牌及北京证券交易所上市以来，共进行过七次股票发行， 前五次股票发行募集资金已分别于 2019 年 11 月 4 日、2019 年 11 月 6 日、2020 年 8 月 7 日、2021 年 8 月 31 日和 2022 年 3 月 7 日使用完毕，募集资金账户分别 于 2019 年 11 月 4 日、2019 年 11 月 6 日、2020 年 11 月 16 日、2021 年 8 ...

新京报贝壳财经讯（记者黄鑫宇）今年春节期间DeepSeek推出的DeepSeek-V3及推理版本 R1火爆出 圈，其以仅600万美元的训练成本达到与OpenAI媲美的性能，成本优势显著。在此带动下，A股市场算 力板块大幅飙升。被称为"北交所算力服务第一股"的并行科技（即北京并行科技股份有限公司； 839493.BJ），在2月中更是创下10个交易日里涨幅接近4倍的纪录。 半年报显示，2025年5月27日，并行科技与阿里云在北京正式签署框架合作协议，双方将围绕国产算力 中心共建、公共云算力生态打造、大模型行业应用落地三大方向展开深度合作，联合推动国产化算力基 础设施创新与人工智能技术在科研、教育及产业场景的规模化落地，共同探索算力普惠与AI协同的新 生态。 公开资料显示，并行科技成立于2007年，于2023年11月登陆北交所，是全A股市场中首家提供专业算力 服务的上市公司。今年一季度，并行科技实现营业收入1.98亿元，同比增长51.68%；实现归母净利润约 为359.40万元，同比增长43.77%。 另一方面，半年报显示，截至2025年6月末，并行科技的资产负债率（合并）为78.37%，较上年同期的 69.56% ...

Core Viewpoint - Securities firms have provided "buy" ratings for nearly 10 A-share listed companies on July 29, focusing on industries such as food and beverage, consumer electronics, fertilizers, non-ferrous metals, chemical products, software development, and pharmaceuticals [1]. Group 1: Company Ratings - Dongpeng Beverage received significant attention from securities firms, with 14 reports in the past month, including three on July 29, making it the top stock recommended by brokers on that day [4]. - Yingshi Innovation was the second most recommended stock, receiving five reports in the past month, with a notable report on July 29 highlighting its product and brand barriers [4]. - Yaqi International was the third most recommended stock, with three reports in the past month, and a new coverage report predicting substantial revenue growth [5]. Group 2: Financial Projections - Yaqi International's projected revenues for 2025-2027 are 5.835 billion, 7.980 billion, and 9.870 billion yuan, with net profits of 1.732 billion, 2.450 billion, and 3.213 billion yuan respectively [5]. - Dongpeng Beverage's growth is driven by a strong demand for energy drinks and tea, with expectations for platform development and international expansion [4]. Group 3: Industry Focus - The industries attracting broker attention include food and beverage, fertilizers, non-ferrous metals, chemical products, software development, and pharmaceuticals, indicating a diverse investment interest [1].

Core Viewpoint - The report highlights the significant revenue growth of Parallel Technology, driven by its intelligent computing cloud services, with a projected revenue increase of 69% year-on-year for the first half of 2025, supported by strategic collaborations and advancements in AI integration [1][4]. Financial Performance - In H1 2025, the company achieved revenue of 458 million yuan (yoy +69%), with a net profit of 5.08 million yuan (yoy +20%) and a net cash flow from operating activities of 39.26 million yuan (yoy +323%) [1]. - For Q2 2025, revenue reached 259 million yuan (yoy +86% / qoq +31%), with a net profit of 1.48 million yuan [1][2]. Business Segments - The company reported revenue from its four main business segments: computing power services (441.72 million yuan, yoy +71%), supercomputing cloud system integration (10.23 million yuan, yoy +121%), supercomputing software and technical services (5.69 million yuan), and other services (0.03 million yuan) [2]. - The intelligent computing cloud service segment alone generated 264.63 million yuan, reflecting a year-on-year growth of 175% [2]. Strategic Collaborations - Parallel Technology signed a framework cooperation agreement with Alibaba Cloud to enhance domestic computing power infrastructure and promote AI technology applications across various sectors [3]. - The company's MaaS platform has integrated the Zhiyu GLM-Z1 series inference models, aiming to provide efficient and cost-effective solutions for clients [3]. Industry Trends - The intelligent computing power scale in China is projected to reach 725.3 EFLOPS in 2024, with a year-on-year growth of 74.1%, driven by the national "East Data West Computing" initiative [4]. - The report anticipates that by 2025, the general computing power scale will reach 85.8 EFLOPS, while intelligent computing power is expected to grow to 1,037.3 EFLOPS, marking a 43% increase [4]. Profit Forecast - The company forecasts net profits of 24 million yuan, 36 million yuan, and 48 million yuan for 2025, 2026, and 2027 respectively, with corresponding EPS of 0.40, 0.61, and 0.80 yuan per share [5].

Investment Rating - The investment rating for the company is "Accumulate" (maintained) [5] Core Views - The company's revenue in H1 2025 reached 458 million yuan, representing a year-on-year increase of 69%. The growth was driven by the high increase in intelligent computing cloud services, which saw a 175% year-on-year growth [6][9] - The company has signed a framework cooperation agreement with Alibaba Cloud to enhance AI technology accessibility through the integration of the GLM-Z1 series inference models into its MaaS platform [6][7] - The "Parallel Computing Network" is expected to support the national "East Data West Computing" strategy, with the intelligent computing scale projected to reach 725.3 EFLOPS in 2024, a year-on-year increase of 74.1% [7] Financial Performance Summary - In H1 2025, the company achieved a net profit of 5.08 million yuan, a year-on-year increase of 20%, and a net cash flow from operating activities of 39.26 million yuan, up 323% year-on-year [6] - Revenue projections for 2025 are estimated at 863 million yuan, with a year-on-year growth rate of 31.86% [8] - The company is expected to achieve net profits of 24 million yuan in 2025, with corresponding EPS of 0.40 yuan per share [9]

Core Viewpoint - The 2025 World Artificial Intelligence Conference (WAIC2025) highlighted the importance of user-centric computing power services and the mismatch in the current market for computing power [1] Group 1: Industry Insights - The chairman of Parallel Technology, Chen Jian, emphasized that having computing power does not guarantee sales, indicating a significant mismatch in the market [1] - Parallel Technology differentiates between computing power services and computing power leasing, with the former focusing on business needs rather than the specifics of computing power [1] - The long-term goal of Parallel Technology is to provide on-demand computing power, ensuring users do not need to consider the location of the computing resources when they require them [1]