Summary of Conference Call Notes Industry Overview - The focus is on the AI industry, particularly the role of cloud service providers and the implications of large models like CloudBot and C-DAS 2.0 on token consumption and software industry dynamics [1][2][5][12]. Key Points and Arguments Demand for Tokens - There is a significant demand for tokens due to high-frequency calls to AI models, with weekly consumption potentially reaching tens of millions [1][4][13]. - The transition from dialogue-based interactions to tool invocation has increased token usage, necessitating more computational power [2][12]. Role of Cloud Service Providers - Cloud providers are crucial in the AI era, offering mirrored services that lower user entry barriers and determining which large models can be accessed [1][5]. - Renting cloud services, such as Tencent Cloud, allows users to utilize complex models without significant changes to their infrastructure [5]. Risks Associated with AI Tools - There are potential security risks when installing plugins or skills, as some may disguise malicious software that can consume server resources [6]. - Users must be cautious to avoid issues similar to those seen in the early internet era, such as virus infections [6]. Impact on Software Industry - AI technology is diminishing the value of traditional software entry points, particularly in the SaaS sector, where Chinese companies lag behind their U.S. counterparts [7][8][9]. - The software industry is expected to face new challenges and opportunities as AI-based platforms gain prominence [7]. Advantages of Chinese Software Companies - Chinese A-share software companies focus on customized development and customer service, making them suitable partners for AI technology [11]. - These companies possess industry-specific knowledge that complements AI's general capabilities, allowing for a synergistic relationship [11]. Future of Cloud Computing and Token Consumption - The importance of cloud providers will increase as models like C-DAS 2.0 require substantial computational resources and token consumption [12][20]. - Major companies like ByteDance and Alibaba anticipate a tenfold increase in token consumption in the coming years, indicating that charging for large model usage will become standard [14]. Recommendations for Investment - Infrastructure-related companies, such as NetSpeed, are recommended due to the growing demand for efficient data transmission in AI applications [15]. - In the AI video production sector, companies like Zhao Chi and Wanxing Technology are highlighted for their innovative tools that enhance production efficiency [18]. - IDC firms should focus on partnerships with major platforms, with recommendations for companies like Dongyangguang and Runze in the ByteChain ecosystem, and Century Internet Data Port in the AliChain ecosystem [19]. Prospects for Domestic Computing Chips - Domestic computing chips like Haiguang and Cambrian are expected to have a positive long-term outlook despite current market pessimism [20]. - The increasing demand for computational power due to rising token consumption presents a buying opportunity for stocks in these companies [20]. Other Important Insights - The transition to AI tools is reshaping the software landscape, with a shift away from reliance on single software applications towards integrated AI solutions [9][10]. - The response time of CloudBot is noted to be longer compared to other models, indicating a need for improvement in processing speed [16].

Group 1 - The core viewpoint of the news highlights the performance of the consumer electronics sector, with the CSI Consumer Electronics Theme Index down by 0.85% as of February 11, 2026, and notable fluctuations among constituent stocks [1] - Key stocks include Shengyi Technology leading with a 2.55% increase, while Dongshan Precision experienced the largest decline at 6.21% [1] - The Consumer Electronics ETF also saw a decrease of 0.89%, with a latest price of 1.22 yuan and a trading volume of 7.855 million yuan during the session [1] Group 2 - The launch of the "Doubao AI Phone" signifies a deep integration of large model capabilities into mobile devices, with major companies like Apple and ZTE accelerating their AI product developments, marking 2026 as a pivotal year for edge AI product iteration [2] - AI functionalities are shifting from optional features to standard capabilities, driving improvements in performance metrics and customization levels for components such as camera modules and optical elements [2] - The CSI Consumer Electronics Theme Index comprises 50 listed companies involved in component production and design, with the top ten weighted stocks accounting for 53.34% of the index [2]

Core Viewpoint - The peak of overseas liquidity shock has passed, and the earnings expectations for the Hong Kong tech sector have been continuously adjusted downwards since September, but have recently stabilized, indicating that current stock prices reflect a phase of market adjustment [1] Group 1: Market Conditions - The relative valuation of the Hong Kong tech sector has reached historical lows, with a regulatory and development environment significantly better than in previous years, indicating that the sector is significantly undervalued [1] - The industry is progressing smoothly, with various large models being promoted actively by companies like Tencent and Alibaba during the Spring Festival, continuously catalyzing the AI industry [1] Group 2: Investment Opportunities - The Hong Kong Tech ETF (513020) tracks the Hong Kong Stock Connect Technology Index (931573), covering core assets in sectors such as internet, innovative pharmaceuticals, and new energy vehicles, reflecting the diversified characteristics of the tech industry and the overall performance of core tech companies in the Hong Kong market [1] - The Hong Kong Stock Connect Technology Index has outperformed the Hang Seng Tech Index by over 140% since its base date at the end of 2014, with a cumulative return of 224.25% compared to the Hang Seng Tech Index's 83.87%, indicating long-term outperformance against similar indices [1]

每经AI快讯，2月11日，据科大讯飞官微消息，今天，科大讯飞正式发布基于全国产算力训练的星火X2 大模型。这次面向关键行业落地的实用化升级，标志着我国基于全国产算力训练的大模型底座在算法和 工程创新方面迈入新阶段。星火X2在实现通用能力全面升级的同时，更加专注于教育、医疗、汽车、 智能体等高专业性、高体验性场景，致力于为这些领域提供领先、可靠的技术方案与支撑。 ...

科大讯飞今日正式发布基于全国产算力训练的星火X2大模型。据介绍，此次升级有两大亮点：通用能 力全面升级，星火X2整体能力对标国际顶尖模型水平，130+多语言综合能力继续提升；依托星火X2的 算法突破、行业高质量数据集以及专业思维链的强化学习，星火行业大模型持续升级，产品应用方案取 得显著进步。 (本文来自第一财经) ...

Core Viewpoint - The company has made progress in developing a large model application focused on cardiovascular diseases, which includes risk assessment capabilities and has undergone successful pilot testing [2] Group 1 - The large model application is primarily aimed at cardiovascular disease [2] - The application has been launched and has completed relevant pilot tests with positive results [2] - The company is unable to predict the future impact of this business on its performance [2]

Core Viewpoint - The article discusses the emergence of China's AI models, particularly the new Qwen3.5 model from Alibaba and the upcoming DeepSeek V4 model, indicating a significant growth phase for AI applications in February 2024 [3][15]. Group 1: AI Model Developments - Alibaba's new AI model Qwen3.5 has been integrated into the global AI open-source community project on HuggingFace, marking a notable advancement in AI technology [3]. - DeepSeek is set to launch its next-generation flagship AI model, DeepSeek V4, contributing to the competitive landscape of AI models in China [3][15]. Group 2: Investment Opportunities in Hong Kong Stocks - The Hong Kong Internet ETF is heavily weighted towards major tech and internet companies, with Tencent and Alibaba accounting for nearly 30% of the index [6][18]. - The top ten stocks in the index include significant players in platform technology and AI applications, collectively representing over 78% of the index [6][18]. Group 3: Performance Metrics - The Hong Kong Internet ETF has shown a remarkable performance over the past three years, with a total return of 156.48%, outperforming similar indices like the Hang Seng Tech Index [10][22]. - Historical performance data indicates that the Hong Kong Internet Index has experienced significant volatility, with maximum gains and annualized volatility metrics highlighting its aggressive growth potential [10][22].

Summary of Key Points Industry Overview - The report discusses the usage of large model tokens in the AI industry, specifically for the week of February 2 to February 9, 2026, with a total usage of 9.81 trillion tokens, reflecting a week-over-week increase of 18.9% [1] Market Share Analysis - **Google**: - Token usage amounted to 2.12 trillion, with a week-over-week increase of 11.0%, capturing a market share of 22.1% [1] - **xAI**: - Token usage was 0.796 trillion, showing a decrease of 15.3%, resulting in a market share of 8.3% [1] - **Anthropic**: - Token usage reached 1.44 trillion, with a week-over-week increase of 7.5%, holding a market share of 15.0% [1] - **OpenAI**: - Token usage was 1.24 trillion, reflecting an increase of 8.8%, with a market share of 13.0% [1] - **DeepSeek**: - Token usage was 0.900 trillion, showing a significant increase of 17.6%, capturing a market share of 9.4% [1] Additional Insights - The overall growth in token usage indicates a rising trend in the adoption of large models within the AI sector, suggesting potential investment opportunities in companies leading this growth [1]

Core Viewpoint - The stock of Zhiyu (02513) has surged over 70% this week, with a current price of 340 HKD and a trading volume of 2.92 billion HKD, driven by the anticipation of the release of its new model GLM-5 [1] Group 1: Company Developments - Zhiyu is advancing a confidential project related to the newly launched anonymous model "Pony Alpha" by OpenRouter, which has garnered significant attention [1] - The company's Chief Scientist, Tang Jie, confirmed that Zhiyu will soon release the GLM-5 model, indicating a strong pipeline of innovation [1] Group 2: Market Position and Growth Potential - According to Guangfa Securities, Zhiyu is a leading provider of large model services in China, with a core focus on its self-developed GLM base model [1] - The company has established a comprehensive model matrix that includes multimodal, agent, and coding capabilities, offering API services, localized deployment, and industry solutions to enterprise clients [1] - Guangfa Securities projects that Zhiyu's revenue will continue to grow rapidly from 2025 to 2027, with scale effects becoming increasingly evident, suggesting strong long-term profitability [1]

2026年的科技圈，正在经历一场昂贵且有趣的竞赛。 一边是微信里头闹出的动静。腾讯自家的大模型应用元宝，想趁着春节冲一波用户，老办法，砸钱10个 亿，发红包。结果场面一度十分尴尬，红包链接没刷屏多久，就被微信安全机制精准拦截。这种大水冲 了龙王庙的乌龙，被外界调侃为自家兄弟打了自家的脸。 这让人纳闷：腾讯搞AI，怎么还是撒钱拉人那一套？这种动作显得有些心有余而力不足。 另一场竞赛关于大洋彼岸的苹果。 即便新iPhone已经薄得像片刀，工艺达到了工业巅峰，可一年前被库克视为"个人智能重头戏"的Apple Intelligence，依然没能点燃市场的兴奋点。 那个传说中能接管一切的Siri，至今没见着真身。 媒体的标题整齐划一：苹果的AI，又慢了。在数次推迟后，苹果确认在iOS底层引入了谷歌的Gemini模 型。 对于一个拥有100万用户的初创AI产品来说，引入大模型是一场令人兴奋的，甚至是允许失败的实验。 即使模型产生幻觉，回答出现了逻辑漏洞，用户往往会报以宽容，甚至将其当作有趣的梗在社交平台传 播。 这是初创公司的特权，也是它们能够做到以周为单位迭代的天然优势。 但微信、苹果不一样。 这两个动作在当时的舆论场 ...