Core Viewpoint - The company reported stable revenue growth in Q1 2025, with a narrowing of losses compared to the previous year [1][4]. Financial Performance - Q1 2025 revenue reached 364 million yuan, a year-on-year increase of 1.84% [1]. - The net profit attributable to shareholders was -102 million yuan, a 32.14% reduction in losses compared to Q1 2024 [1]. - The non-recurring net profit was -108 million yuan, with a 30.41% reduction in losses year-on-year [1]. - Gross margin stood at 55.73%, a decrease of 4.76 percentage points compared to the previous year [1]. Cost Management - The company has implemented cost control measures starting in 2024, leading to a reduction in total operating expenses to 357 million yuan, down 4.8% year-on-year [2]. - Selling expenses were 173 million yuan, a decrease of 4.27% [2]. - R&D expenses were 133 million yuan, down 5.39% [2]. - Management expenses were 51 million yuan, a reduction of 5.95% [2]. Strategic Initiatives - The company is focusing on the "AI + Security" sector, launching the Fengyunwei AI security capability platform, which integrates various models and knowledge bases to enhance security operations and responses [2]. - The AI-Scan product was introduced for assessing risks associated with large models, covering key technologies for security throughout the AI lifecycle [2]. Market Outlook - The company is expected to benefit from the acceleration of Agent deployment driven by open-source models like DeepSeek and Qwen3 [3]. - Revenue forecasts for 2025-2027 have been adjusted to 2.583 billion, 2.792 billion, and 3.148 billion yuan respectively, with net profit forecasts revised to 7 million, 64 million, and 116 million yuan [4].

2025 年 3 月全球 AI 重要趋势。 文 丨 贺乾明 2025 年 3 月的 AI 月报，我们开始尝试一种新形式：和知乎一起举办 "AI 脑暴" 线下活动，围绕每月 一个热门 AI 主题，邀请相关学界研究者、业界从业者（研发或技术人员）、投资人等一起做闭门圆 桌讨论。 3 月 30 日，第一期 AI 脑暴举行，主题是 Agent，我们邀请 6 位嘉宾参加，他们来自高校、互联网公 司和非营利 AI 研究机构。本期月报中，我们摘录了部分 AI 脑暴中的讨论。 本期月报，你会看到： 为什么开发 Agent 简单，但做好很难 AI 脑暴活动中 "壳重要还是模型重要" 部分讨论要点 以下是我们第 5 期 AI 月报，欢迎大家在留言区补充我们没有提到的重要趋势。 开发 Agent 简单，让它好用难 3 月初，通用 Agent 产品 Manus 上线第二天，就被复刻出两个开源版本——OpenManus、OWL。 OpenManus 甚至只用 4 个人，花 3 个小时。 这种 "速成" 似乎暗示：做 Agent 没那么难。但从实际体验和系统复杂度来看，让 Agent "真正好用" 仍存在挑战。 大模型公司 Anthro ...