Core Viewpoint - The company has demonstrated robust growth in 2024, with a significant reduction in net losses, driven by advancements in AI technology and strategic business focus [1][2][3]. Financial Performance - In 2024, the company achieved revenue of 5.261 billion yuan, a year-on-year increase of 25.1% [1] - Gross profit reached 2.245 billion yuan, maintaining a gross margin of 42.7% [1] - The net loss attributable to shareholders was 269 million yuan, a substantial reduction of 70.4% year-on-year, with a net loss margin of 5.1% [1] - Adjusted net loss for the year was 292 million yuan, narrowing by 29.6% compared to 2023 [1] - R&D expenses amounted to 2.170 billion yuan, representing 41.2% of revenue [1] Business Segments - The "Xianzhi AI Platform" business generated revenue of 3.676 billion yuan, up 46.7% year-on-year [2] - The "SHIFT Intelligent Solutions" business saw revenue decline by 20.3% to 1.022 billion yuan due to a strategic focus shift towards the Xianzhi AI Platform [2] - The "Shishuo AIGS Service" business contributed 563 million yuan, providing efficient development tools and services based on generative AI [2] Industry Expansion - The company maintained its position as the leading machine learning platform in China for six consecutive years, with a customer base expansion of 16% to 161 benchmark users [3] - Average revenue per benchmark user was 19.1 million yuan, with a net revenue growth rate (NDER) of 110% [3] Corporate Restructuring - The company established "Paradigm Group," with its enterprise service business becoming a core subsidiary, and launched a new consumer electronics segment called "Phancy" [4] - Phancy aims to provide AI Agent-based integrated hardware and software solutions, enhancing user interaction and service connectivity [4] Focus on AI Agent - The company is prioritizing the development of AI Agent technology, which can automate various business processes for enterprise clients [5] - In 2024, the company deployed AI Agent solutions across over 10 industries, enhancing operational efficiency [5] Development of Edge AI - The company anticipates significant growth in edge AI, with the launch of the ModelHubAIoT solution expected to facilitate easy deployment of distilled models [6] - Collaborations with partners aim to integrate edge computing capabilities with AI models, enhancing the accessibility of AI technology in consumer electronics [6][7] AI All-in-One Machine Demand - The company has partnered with Huawei to launch the SageOne IA solution, addressing the growing demand for AI model deployment and ensuring data security [8] - The solution enhances performance and flexibility for enterprises, supporting various mainstream large models [8]

Core Viewpoints - The company's core business drives high-quality revenue growth, achieving revenue of RMB 5.261 billion in 2024, a year-on-year increase of 25.1%, with the "Prophet AI Platform" contributing RMB 3.676 billion, a significant growth of 46.7%, accounting for 69.9% of total revenue. Meanwhile, the net profit attributable to the parent company narrowed its loss to RMB -269 million, a substantial reduction of 70.4% [1][2] - The company has a high customer stickiness among industry-leading clients, with the number of benchmark clients increasing by 16% to 161, an average revenue contribution of RMB 19.1 million, and a net retention rate (NDER) of 110%. The company has launched an AI Agent collaborative operation solution in partnership with Zhiyuan Interconnect, integrating the AI Prophet platform with the AICOP platform [1][2] - The company upgraded its group structure and launched the "Phancy" brand to provide C-end AI capabilities in addition to its B-end business. The ModelHub AIoT solution supports the deployment of large models on local devices, offering low-latency and high-privacy edge AI services, collaborating with brands like Lenovo to enhance the deployment of Agents on the edge [1][2] Financial Performance - In 2024, the company reported a total revenue of RMB 5.261 billion, a year-on-year increase of 25.1%, with a gross profit of RMB 2.245 billion and a gross margin of 42.7%. The net profit attributable to the parent company improved to RMB -269 million, a significant reduction of 70.4% [1][2] - The core business revenue and profit both grew rapidly, with the "Prophet AI Platform" achieving revenue of RMB 3.676 billion, a year-on-year increase of 46.7%, and accounting for 69.9% of total revenue. The company's R&D investment reached RMB 2.170 billion, with an R&D expense ratio of 41.2% [1][2] Business Segments - The core business structure remains coordinated across three major segments: the Prophet AI Platform, SHIFT Intelligent Solutions, and Shisuo AIGS Services. In 2024, the Prophet AI Platform generated revenue of RMB 3.676 billion, a year-on-year increase of 46.7%, while SHIFT Intelligent Solutions revenue decreased by 20.3% to RMB 1.022 billion due to resource focus on the AI platform strategy [2][3] - The SHIFT Intelligent Solutions provide intelligent solutions for various industries, further deepening industry applications and accelerating digital transformation. The Shisuo AIGS Services empower software development through generative AI technology, enhancing user experience and product capabilities [3][4] Customer Base and Market Position - The number of benchmark users reached 161 in 2024, a 16% increase year-on-year, with an average revenue contribution of RMB 19.1 million. The customer concentration risk has been effectively mitigated, with the largest single customer revenue share decreasing from 12.7% in 2023 to 10.6% in 2024 [4][5] - The company has helped over 10 industries develop and deploy enterprise-level AI Agents, with practical experience in various verticals such as financial credit risk control and water and electricity equipment operation [4][5] Strategic Developments - The company completed a group structure upgrade and officially rebranded as "Paradigm Group," with its enterprise-level AI business becoming the core B-end brand. The launch of the Phancy brand focuses on "AI Agent + World Model" technology, providing edge AI capabilities to both consumer and industrial sectors [5][6] - The ModelHub AIoT solution supports local deployment of distilled models, achieving low-latency and high-privacy edge AI services, and has collaborated with brands like Acer and Lenovo to upgrade traditional devices into AI terminals [5][6] Profit Forecast - As a leader in enterprise-level AI, the company has seen continuous revenue growth and narrowing net losses, with a clearer profit model. The number of benchmark clients and revenue are steadily increasing, with projections for 2025-2027 revenues of RMB 6.663 billion, RMB 8.347 billion, and RMB 10.413 billion, representing year-on-year growth rates of 26.66%, 25.26%, and 24.77% respectively [5][6]

Core Insights - The introduction of the AG-UI protocol completes the necessary framework for AI application ecosystems, following the MCP and A2A protocols [3][24] - The AI application ecosystem is structured around three roles: users, agents, and the external world, with a focus on interoperability among these roles [2][3] - The trend in AI model training is becoming increasingly oligopolistic, with only a few major players capable of developing foundational large models [1] Group 1: Protocols Overview - MCP and A2A protocols serve as foundational infrastructures for AI applications, facilitating communication between agents and the external world, and between agents themselves [2][9] - AG-UI protocol addresses the communication between users and agents, filling the gap left by MCP and A2A [3][24] - AG-UI provides a standard framework for front-end applications to communicate with back-end agents, enhancing user experience [13][24] Group 2: Agent Functionality - Agents act as intermediaries that perform tasks on behalf of users, similar to real-world agents like real estate brokers [8][9] - The efficiency of agents is highlighted by tools like Lovart, which can autonomously generate video content by coordinating various resources [9][10] - The need for standardized protocols like MCP and A2A arises from the necessity for agents to interact with various tools and each other effectively [9][11] Group 3: AG-UI Protocol Features - AG-UI protocol introduces an event-driven model that allows front-end applications to receive real-time updates from agents, improving user interaction [13][16] - It includes five types of events: Lifecycle Events, Text Message Events, Tool Call Events, State Management Events, and Special Events, which facilitate efficient communication [17][20] - The protocol allows for incremental updates, reducing the need for complete data transfers and enhancing performance [17][22] Group 4: User Experience Enhancement - AG-UI enables front-end applications to provide immediate feedback to users based on agent activity, such as displaying loading indicators during processing [16][22] - The protocol supports a seamless user experience by allowing for real-time updates and interactions without waiting for complete responses [16][22] - By standardizing communication between agents and user interfaces, AG-UI aims to improve the overall efficiency and effectiveness of AI applications [24]

Core Insights - The article discusses the risks associated with the Multi-Agent Collaboration Protocol (MCP), particularly the potential for tool poisoning attacks that could manipulate AI agents to perform unauthorized actions [1][8][9] - The emergence of AI agents is highlighted as a transformative trend, with predictions indicating that by 2028, at least 15% of daily work decisions will be made autonomously by AI agents [2][4] - The commercial viability of AI agents is emphasized, with a focus on their ability to meet consumer needs and create a self-sustaining economic cycle [3][10] Group 1: Agent Ecosystem and Trends - The development of AI agents is expected to either replace traditional applications or enhance them with intelligent, proactive capabilities [2][4] - The introduction of DeepSeek has accelerated the adoption of AI agents, with a notable increase in inquiries and revenue generation in the industry [3][10] - The transition from single assistants to collaborative networks of agents is anticipated, leading to the formation of an "Agent Economy" [4][9] Group 2: Security Risks and Challenges - Security challenges are identified as critical for the stable operation of agent systems, with vulnerabilities in the MCP protocol posing significant risks [7][9] - Tool poisoning attacks (TPA) are highlighted as a major concern, where attackers can embed malicious instructions within the MCP code, leading to unauthorized actions by AI agents [8][9] - The lack of adequate security mechanisms during the design phase of protocols like MCP and A2A has resulted in hidden vulnerabilities that could be exploited [9][12] Group 3: Safety Measures and Industry Response - The industry is urged to implement proactive security measures across the entire value chain to mitigate risks associated with AI agents [11][12] - The responsibility for security varies depending on the application context, with general SaaS products having different security obligations compared to industry-specific applications [11][12] - Collaboration between AI model developers and security firms is essential to address both internal and external security challenges in the deployment of AI agents [12][13]

作者 | AICon 全球人工智能开发与应用大会 策划 | 李忠良 编辑 | 宇琪 随着技术的快速发展，大模型在各行业的应用潜力日益凸显，但如何将大模型能力高效转化为实际业 务价值，仍是企业面临的核心挑战。 近日 InfoQ《极客有约》X AICon 直播栏目特别邀请了 华为云 AI 应用首席架构师郑岩 担任主持人， 和 蚂蚁集团高级技术专家杨浩、明略科技高级技术总监吴昊宇 一起，在 AICon 全球人工智能开发 与应用大会 2025 上海站 即将召开之际，共同探讨大模型如何驱动业务提效。 部分精彩观点如下： 在 5 月 23-24 日将于上海举办的 AICon 全球人工智能开发与应用大会 上，我们特别设置了 【大模型 助力业务提效实践】 专题。该专题将围绕模型选型与优化、应用场景落地及效果评估等关键环节，分 享行业领先企业的实战经验。 查看大会日程解锁更多精彩内容： https://aicon.infoq.cn/2025/shanghai/schedule 以下内容基于直播速记整理，经 InfoQ 删减。 场景探索 郑岩：在探索大模型应用场景时，企业常会遇到"看起来很美但落地难"的需求，各位在实际项目中是 ...

Summary of Key Points from the Conference Call Industry Overview - The focus of the humanoid robot industry has shifted towards the application of AR capabilities and large model capabilities to meet user demands, with expectations for deep integration of hardware and models within 3-5 years in everyday scenarios [1][3] - The development of humanoid robots can be categorized into three stages: initial focus on core components, establishment of hardware architecture, and eventual deep integration of hardware and models for widespread application [3] Core Insights and Arguments - AI Agents play a crucial role as the "brain" of embodied robots, responsible for task decision-making and reasoning, enhancing task execution efficiency through tailored applications for different scenarios [1][8] - The mainstream framework for embodied robot brains is structured in five layers: physical layer, training layer, data layer, model layer (including LLM, VLM, and VLA), and application layer [1][9] - The introduction of 3D spatial perception capabilities is essential for improving spatial modeling and perception, which is vital for achieving general AGI [1][19] - The industrial sector predominantly employs a hierarchical embodied large model architecture to avoid retraining software due to hardware upgrades, contrasting with the academic sector's end-to-end approach [1][17] Technological Developments - Google's RT series models have significantly advanced VLA model development, although they have not been open-sourced, while Stanford and Berkeley's open-source models have accelerated industry growth [1][10][12] - Philips' Helix architecture, released in February 2025, differs fundamentally from the VOLATI model by employing a layered system that allows for cost-effective hardware upgrades [1][14][15] - The VELAN model is currently simple, utilizing text, visual, and action encoding for training, similar to Tesla's autonomous driving approach [1][16] Challenges and Future Directions - Current VLA models face challenges such as insufficient data volume, low task generalization ability, and significant performance impacts from lighting changes [1][18] - The importance of establishing industry standards for humanoid robots is emphasized, as it will influence market development and safety certifications [1][24] - Future trends in intelligent large models will focus on data collection and training to enhance the generalization capabilities of VRA models, with potential for unified foundational VOI models [1][27] Additional Insights - The competition among terminal manufacturers will hinge on optimizing foundational large models and unique advantages in scene data training for better hardware integration [1][2][27] - The VRM model's core in interaction capabilities includes voice recognition, output, and expression management, which are crucial for enhancing robot interaction [1][26] - Data collection in humanoid robotics is evolving, with a focus on human sensory perception data to improve design richness and reduce the Sim-to-Real gap [1][23]

每周一景： 云南玉龙雪山 点击右上角菜单，收听朗读版 【 非银】公募新规利好非银，监管引导险资入市稳市 一、本周观点： 【非银一周观点】： 5 月 7 日中国证监会发布《公募基金高质量发展行动方案》，其中明确建立与基金业绩表现挂钩的浮动管 理费收取机制，强化业绩比较基准的约束作用，考核从重同业排名转向重业绩比较基准。 而当前非银板块 的配置比例从 1.67% 降至 1.52% ，大幅欠配 5.34Pct ；其中券商板块欠配 3.29Pct ，保险板块欠配 1.60Pct ，重视业绩比较基准下非银板块作为大幅欠配板块更为受益。 5 月 7 日，金融监管总局局长李云泽在国新办新闻发布会上表示，将充分发挥保险资金作为耐心资本和长 期资本的作用，加大入市稳市力度， 包括： 1 ）进一步扩大保险资金长期投资的试点范围，近期拟再批 复 600 亿元； 2 ）调整偿付能力监管规则，将股票投资的风险因子进一步调降 10% ； 3 ）推动长周期 的考核机制，促进"长钱长投"。我们认为上述举措有助于稳定资本市场，推动保险公司增加优质权益资产 配置，改善低利率环境下固收类资产收益下行压力，实现资产负债匹配优化。 金融科技方面，大模 ...

Group 1 - The core viewpoint of the article highlights the emergence of DeepSeek as a significant player in the AI landscape, breaking the long-standing dominance of major companies like ByteDance and Tencent in consumer products [2][4] - DeepSeek's application, DeepSeek, achieved over 100 million downloads within a month without any marketing expenses, marking a notable shift in the competitive landscape [2] - The article discusses the growing interest and investment in AI Agents, with Manus achieving 7.5 million USD in funding and a fivefold increase in valuation shortly after its launch [5][4] Group 2 - AI Agents are identified as a key competitive focus for companies in 2025, despite current limitations such as slow response times and hallucinations [5][4] - The article emphasizes the need for new narratives in AI development, suggesting that the evolution of AI Agents will enhance decision-making and content generation capabilities [4][5] - The potential for AI Agents to transform various industries, including healthcare, finance, and education, is discussed, with specific examples of applications already in use [17][12] Group 3 - The article outlines the two main categories of companies developing AI Agents: those focused on technology frameworks and those targeting specific application scenarios [13] - It highlights the importance of foundational model advancements for the effectiveness of AI Agents, indicating that stronger models will lead to better automation and generalization capabilities [11][14] - The future of work is expected to involve collaboration between humans and AI Agents, with a shift in job roles and the emergence of new professions such as AI trainers [15][16]

本文来自微信公众号：R.S.罗夏，作者：罗夏Leo，题图来自：AI生成 那天回国的航班，在伊斯坦布尔机场足足停留了近24小时。 不是暴风雨，不是罢工，而是一个听起来就很模糊的理由："operation issue"。这趟航班据说经常延误，而这次我们刚好是那"常规异常"的 一部分。 机场没给酒店，只能在椅子上蜷缩一晚。休息厅的椅子无法躺平，根本靠不住腰。最难熬的不是疲惫，而是那种——你明明买了服务，对 方却毫无歉意地"例行失责"，你却毫无应对工具的感觉。回国第一天的工作也被耽误。 于是我打开电脑，决定不再忍耐。 一、五分钟Brief之后，GPT开始替我"打这场仗" 我先问："这种情况我可以投诉索赔吗？" GPT回答得干脆利落，引用的是欧盟的261/2004号条例： "对于从欧盟出发的航班，如因航空公司责任导致延误超过3小时，乘客有权获得250至600欧元不等的补偿。9小时延误显然已满 足条件。" 然后，它快速给出了几步执行策略： 1. 拟好一封带有法条引用的英文邮件； 2. 语气保持冷静但强硬，强调维权立场； 3. 提供多个投诉路径，包括航空公司客服邮箱、欧洲管理局平台、土耳其航空民航总局投诉信箱，甚至是相关高 ...

Core Insights - The report from Northeast Securities highlights that banks are expected to become pioneers in AI implementation in China due to ample IT budget, market-oriented systems, and high integration of internal data [1][3] - DeepSeek-R1's inference cost is only 1/30 of comparable products, marking a new phase of "AI popularization" in the industry [1] - The year 2025 is anticipated to be the starting point for AI Agents, with significant competition among major companies in this area [2] Group 1: AI Technology and Applications - DeepSeek has launched several well-known open-source models since its establishment in July 2023, with the DeepSeek-R1 model achieving performance comparable to OpenAI's o1 series at a significantly lower cost [1] - Major banks have actively integrated AI technology into various applications such as investment research, customer service, credit approval, and more, enhancing the intelligence of financial services [3] - IDC predicts that the banking sector will account for over 20% of global AI solution spending from 2024 to 2028 [3] Group 2: Specific Companies and Their AI Initiatives - Yuxin Technology has fully integrated DeepSeek models into its product system, focusing on applications in credit, data, and marketing channels [4] - Jingbeifang has launched an AI large model service platform and several intelligent assistants, achieving breakthroughs in smart fraud prevention and investment advisory across multiple industries [4] - Gaoweida has deeply integrated DeepSeek with its credit business, enhancing credit efficiency and financial report analysis through AI applications [4] - Tianyang Technology has released intelligent testing analysis systems and compliance models, providing banks with intelligent data analysis solutions [4] - Shenzhou Information has upgraded its financial knowledge Q&A and coding assistants, improving development efficiency by 20% and automating 30% of code generation [5]