中信建投证券股份有限公司 关于云从科技集团股份有限公司2024年度现场检查报告 中信建投证券股份有限公司（以下简称"中信建投"或"保荐人"）为云从 科技集团股份有限公司（以下简称"云从科技"或"公司"）首次公开发行股票 并在科创板上市的保荐人及持续督导机构，根据《证券发行上市保荐业务管理办 法》《上海证券交易所科创板股票上市规则》《上海证券交易所科创板上市公司自 律监管指引第 1 号——规范运作》等有关规定，于 2025 年 4 月 28 日-2025 年 4 月 30 日对公司进行了现场检查。现将本次现场检查情况报告如下： 一、本次现场检查的基本情况 中信建投针对云从科技的实际情况制订了云从科技现场检查工作计划。为顺 利实施本次现场工作，提高现场工作效率，切实履行持续督导工作，中信建投证 券提前将现场检查事宜通知云从科技，并要求公司提前准备现场检查工作所需的 相关文件和资料。 2025 年 4 月 28 日-2025 年 4 月 30 日，中信建投保荐代表人根据事先制订的 现场检查工作计划，采取与相关负责人员沟通及询问、查看公司生产经营场所、 查阅公司主要业务合同、往来款明细账、部分原始凭证、募集资金台账及 ...

云从科技自主研发的"从容多模态大模型"在Open Compass评测中以65.5分位列全球前三，超越谷歌 Gemini1.5Pro等模型，并在跨模态跟踪、3D人脸识别等细分领域10次刷新世界纪录。基于这一技术优 势，公司与华为昇腾联合推出的智用一体机解决方案，已在天津港（600717）智慧物流调度、国网山东 能源管理等多个标杆项目中落地，助力企业运营效率提升超20%。(36氪) ...

Core Viewpoint - CloudWalk Technology, one of the "AI Four Little Dragons," reported a significant decline in performance for 2024, with revenue dropping to 398 million yuan, a year-on-year decrease of 36.69%, marking the lowest revenue point in seven years [1][2] Financial Performance - The company recorded a net profit of -722 million yuan, with losses expanding compared to the previous year, marking eight consecutive years of losses [1][2] - Revenue projections made during the company's IPO in 2022 were significantly overestimated, with actual revenues from 2022 to 2024 being 526 million yuan, 628 million yuan, and 398 million yuan respectively, resulting in a negative compound annual growth rate [2][3] - The company's gross margin fell by 16.39 percentage points to 35.81% in 2024, while its peers like SenseTime maintained a gross margin of 42.90% [3] Workforce and R&D - CloudWalk underwent significant layoffs in 2024, reducing its workforce from 801 to 453 employees, a decrease of 43%, with R&D personnel decreasing by 51% [5][6] - The company’s R&D expenses fell by 18.27% to 472 million yuan, yet the R&D expense ratio increased to 118.72% due to declining revenue [5][6] - The departure of key technical personnel, including a core technology staff member, raises concerns about the company's ability to maintain its technological edge [5][6] Strategic Challenges - The company is facing challenges in keeping pace with technological advancements, particularly in the development of large models, which has hindered its competitive position against major players like Baidu and Alibaba [3][4] - CloudWalk's cash flow from operating activities has been negative for three consecutive years, totaling nearly 1.2 billion yuan in cash outflows [3][4] Industry Context - The struggles faced by CloudWalk are not isolated, as other members of the "AI Four Little Dragons" are also experiencing similar issues, including prolonged losses and workforce reductions [6]

Core Insights - CloudWalk Technology is deeply embedding AI agents into key sectors such as finance, transportation, and manufacturing, driving an efficiency revolution through technology implementation [1][2] Group 1: Financial Sector - CloudWalk has established a digital security barrier for six major state-owned banks and over 100 financial institutions, enhancing identity verification accuracy to 99.99% and increasing attack resistance by 80% compared to traditional methods [1] - The AI anti-fraud model implemented by a large bank has tripled the efficiency of identifying high-risk transactions, preventing losses exceeding 100 million yuan annually [1] Group 2: Transportation Sector - The AI video analysis platform from CloudWalk covers nine major airports and hundreds of hub airports in China, reducing flight delay rates by 15% through applications like intelligent security checks and flight scheduling [1] - The launch of PortGPT at Tianjin Port has improved container dispatch efficiency by 40% by optimizing cargo loading paths and ship docking plans, receiving coverage from CCTV's "Economic Half Hour" [1] Group 3: Manufacturing and Energy Sectors - In the manufacturing sector, CloudWalk's intelligent customer service system has achieved a 95% accuracy rate in Q&A and improved customer handling efficiency by 24% [2] - The Shandong power grid has successfully reduced transmission and distribution losses through CloudWalk's smart energy management solution, which utilizes AI for predictive maintenance and load optimization [2] Group 4: Broader Impact and Vision - CloudWalk's long-term vision focuses on defining smart living and enhancing human potential, as evidenced by its AI identification system achieving 95% accuracy in ecological monitoring and its smart city management system reducing incident detection time from hours to minutes [2] - The company's competitive edge lies in addressing industry pain points with deep innovations rather than superficial technological showcases [2] - The transition of AI from laboratories to practical applications in various industries signifies not only commercial value but also serves as a fundamental driver for the intelligent transformation of China's industries [2]

洪依真 A0230519060003 hongyz@swsresearch.com 刘洋 A0230513050006 liuyang2@swsresearch.com 2025 年 05 月 07 日 云从科技 (688327) ——阶段性承压，大模型多标杆项目打造 报告原因：有业绩公布需要点评 增持（下调） | 市场数据： | 2025 年 05 月 07 日 | | --- | --- | | 收盘价（元） | 13.95 | | 一年内最高/最低（元） | 23.84/7.25 | | 市净率 | 13.9 | | 股息率%（分红/股价） | - | | 流通 A 股市值（百万元） | 10,513 | | 上证指数/深证成指 | 3,342.67/10,104.13 | | 注："股息率"以最近一年已公布分红计算 | | | 基础数据: | 2025 年 03 月 31 日 | | --- | --- | | 每股净资产（元） | 1.00 | | 资产负债率% | 59.05 | | 总股本/流通 A 股（百万） | 1,039/754 | | 流通 B 股/H 股（百万） | -/- | 一年内 ...

Core Insights - The multi-modal large model industry focuses on deep learning models capable of processing, understanding, and generating various types of data, including text, images, audio, and video, enabling complex and intelligent tasks [1] - The industry has a wide application potential across various sectors such as natural language processing, image recognition, speech recognition, intelligent driving, and medical imaging diagnosis [1] Industry Overview - The multi-modal large model industry chain is complex, encompassing hardware facilities, software development, and various model types, including CLIP, BLIP, and LLaMA, among others [1] - The industry is divided into three layers: the foundational layer (hardware and basic software), the model layer (various types of multi-modal large models), and the application layer (industry-specific applications) [1] Cost Structure - The training costs for mainstream domestic large models range from tens of millions to hundreds of millions of dollars, with major companies like Baidu, Alibaba, and Tencent investing over $200 million [3][5] - Startups like Kimi and DeepSeek have managed to reduce training costs to between $30 million and $60 million through technological optimizations [3] - Cloud hosting costs are significantly influenced by model scale, with major companies leveraging their own cloud platforms to reduce costs [3] Development History - The global large model industry has evolved through several phases: early exploration (1956-2005), rapid growth (2006-2019), the rise of large models (2020-2022), and the current phase of widespread application starting in 2023 [6] Computational Demand - The demand for computational power in AI is increasing, with larger models requiring exponentially more computational resources; for instance, the GPT-3 model requires 3640 PF-days of computation and at least 10,000 GPUs [9] - As model parameters increase, the computational investment needed grows significantly, influenced by model architecture, optimization efficiency, and hardware capabilities [9]

Group 1 - The People's Bank of China announced a 0.5 percentage point reduction in the reserve requirement ratio, expected to provide approximately 1 trillion yuan in long-term liquidity to the market, along with a 0.1 percentage point decrease in policy interest rates [1] - The AI sector is experiencing a significant transformation, moving from quantitative to qualitative changes, with advancements in general large models demonstrating near-human capabilities in various cognitive tasks [1] - The AI technology is reshaping social production methods and human existence, indicating a profound impact on various industries [1] Group 2 - The release of multiple AI models by Alibaba and the financial results from major US tech companies highlight the competitive landscape in the AI sector [2] - The upcoming 2025 Lenovo Tech World and other significant industry events indicate a growing focus on AI and related technologies [2] - The emergence of new job roles, such as prompt engineers, reflects the changing employment landscape driven by AI advancements [4] Group 3 - The diversification of AI applications is evident, with digital human technology marking a shift towards multi-dimensional penetration in various fields, including education and healthcare [5] - The market for digital humans is projected to grow significantly, with estimates indicating a market size exceeding 640 billion yuan by 2025 [5] - The integration of AI into public services and commercial sectors demonstrates the expanding boundaries of technology applications [5] Group 4 - The competition in the AI industry is shifting towards breakthroughs in underlying technologies and cost control, with advancements in embodied intelligence and multi-modal models [7] - The technology sector is expected to regain momentum as concerns over previous performance and tariff disruptions dissipate, with a focus on long-term industry trends [8] - The upcoming months are critical for the tech sector, with numerous industry conferences and events expected to catalyze new growth opportunities [8] Group 5 - The TMT sector is showing signs of recovery, with a notable increase in net profit growth rates, particularly in the AI industry [9] - Institutional investors have significant room for increasing allocations in the TMT sector, particularly in computer and media segments [9] - The AI ETF, which tracks the innovation board's AI index, includes major companies across the AI value chain, indicating a strategic investment opportunity [9][10]

1. 车企董事长薪酬曝光！李想6.4亿元，吉利李书福仅37.6万元 2.阿里通义应用视觉团队负责人薄列峰离职，入职某互联网大厂，任多模态部门负责人 3. 宣传三元锂，装车变成磷酸铁锂？问界回应M8出厂合格证印错 4.曝苹果智能国行版将在iOS 18.6上线，百度、阿里巴巴提供支持 5.乐道员工炫耀"卖爆了"被沈斐批评：绝不应该轻易立flag 6.京东骑手招募规则悄然改变：「目前」由公司承担所有五险一金 7.美国盯上"对华特供芯片"，拟立法跟踪英伟达芯片定位 8.苹果据称考虑调整iPhone发布节奏：标准版、Pro版机型或错开发布 再来看民营车企，最受关注的当属比亚迪。财报显示，董事长、执行董事、总裁王传福税前报酬总额 765.5万元，在公司高管中排名第9，薪酬最高的是执行副总裁李柯，达1689万元，执行副总裁罗红斌、 高级副总裁何志奇税前工资同样均超千万元。吉利汽车集团CEO桂生悦、吉利控股集团CEO李东辉薪资总 额分别为3201万元、1967万元，而作为掌舵者的李书福，其全年薪酬仅37.6万元。长城财报显示，长城 汽车董事长魏建军全年薪酬568.56 万元，与2023年持平，长城汽车副总经理、副董事长赵国庆 ...

曾被誉为"AI四小龙"中的云从科技，披露2024年年报。这是近7年来，其营收最低的一次年度披露，同 时已连续第8年亏损，累计亏损额超45.79亿元。 这份财报显示，2024年全年，云从科技营收约3.98亿元，同比下降36.69%；归母净利润约-6.96亿元，亏 损同比扩大8.12%。 文 丨 新浪科技 周文猛 此外，云从科技在一年内整体员工数量锐减超四成，研发人员减员比例高达51%，此外，公司仅有的三 位核心技术人员之一——云从科技副总裁、AI平台中心负责人张岭，也在年内离任，目前公司仅余2位 核心技术人员。 面对业绩萎靡与大减员的双重风波，云从科技董事长、总经理周曦的年薪也有微调：年薪为191.49万 元，比去年少了7.46万元，降幅为3.75%。 | | | | | | | 里包:刀 巾种:人民申 | | --- | --- | --- | --- | --- | --- | --- | | | | 2023-31 | | 本期比 上年同 | 202211 | | | 主要会计数索 | 2024年 | 调整后 | 试题前 | 期增速 | 调整后 | 调整前 | | | | | | (%) | | | | ...

Investment Rating - The report maintains a "Buy" rating for the company [1][4] Core Views - The company is exploring the implementation of AI models across various sectors, including smart governance, smart travel, smart finance, and more [4] - The company has launched several AI models, including the "Congrong Intelligent Review Model" and a self-developed financial model, and has partnered with Huawei to release integrated AI training and inference machines [4] - The company aims to promote the prosperity of the Hongmeng ecosystem and has secured projects with Guangzhou Bank and Chongqing Rural Commercial Bank [4] - The expected EPS for 2025-2027 is projected to be -0.45, -0.38, and -0.19 yuan respectively [4] Financial Summary - In 2024, the company's revenue is expected to be 398 million yuan, a decrease of 36.69% year-on-year, with a net profit of -696 million yuan, a decrease of 8.12% year-on-year [6][8] - Revenue is projected to grow to 440 million yuan in 2025, with a growth rate of 10.64%, and further to 560 million yuan in 2026, with a growth rate of 27.27% [6][8] - The company’s net profit is expected to improve from -468 million yuan in 2025 to -192 million yuan in 2027 [6][8]