GenAI系列报告之38：智谱AutoGLM拉开国内端侧AI Agent序幕，AI应用进入新阶段

Investment Rating - The report rates the industry as "Overweight" indicating a positive outlook for the sector compared to the overall market performance [2]. Core Insights - The launch of Zhipu AI's AutoGLM marks the beginning of a new era for domestic AI agents, allowing users to perform various tasks on mobile applications through simulated interactions [2][3]. - AutoGLM's functionality includes completing shopping tasks, sending messages, and summarizing content, showcasing its practical utility in daily user interactions [4][5]. - The competitive landscape for mobile AI agents is expected to diversify, with key players including hardware manufacturers like Apple and Huawei, internet giants like Google and Tencent, and AI model developers like Zhipu and OpenAI [6][7]. - The gap between domestic and international AI model capabilities is narrowing, with domestic models beginning to meet the requirements for potential blockbuster AI applications [7]. Summary by Sections Product Launch - Zhipu AI introduced AutoGLM, which can simulate user actions on mobile interfaces, allowing for tasks such as ordering food and messaging through voice or text commands [3][4]. Competitive Landscape - The future of mobile AI agents is anticipated to be vibrant, with various manufacturers leveraging their unique advantages. Key competitors include: 1. Hardware and OS manufacturers (e.g., Apple, Huawei, Xiaomi) with strong user bases but weaker AI model capabilities. 2. Internet giants (e.g., Google, META, Tencent) with extensive app ecosystems and strong AI model capabilities. 3. AI model developers (e.g., Zhipu, OpenAI) focusing on enhancing inter-app communication [6][7]. Market Trends - The report highlights a shift in focus from training AI models to optimizing inference capabilities, suggesting that domestic AI applications are poised for significant growth [7]. - The report identifies potential investment targets in the media and computer sectors, including companies like Giant Network, Kainet, and iFLYTEK, which are well-positioned to benefit from advancements in AI applications [7][8].