春节假期期间，海外内主要科技厂商展开新一轮Al竟赛。阿里巴巴、字节跳动、智谱等国内厂商纷纷上 线最新的大模型Qwen3-Max-Thinking、豆包2.0、GLM-5等。海外厂商同样紧随其后，谷歌发布Gemini3 DeepThink推理增强版本等。场景下沉不断开拓C端用户，筑牢基础用户生态。 国内AI厂商在春节期间开启"红包大战"，腾讯元宝发放10亿元现金红包；千问启动30亿元的"春节请客 计划"，以"帮大家点一杯奶茶"迅速出圈；字节跳动通过春晚送出超过10万份科技好礼等。 国联民生证券表示，持续看好AI投资机会，从四维度验证行业高景气：①科技巨头资本开支增长； ②AI助手Clawdbot验证AI从对话工具向执行者转型，边缘AI产品前景广阔；③算力基础设施供需紧张 持续，推动上游通胀，云厂涨价、H100租赁价格上行，存储板块或延续涨价态势；④AI已进入应用落 地、需求加速与产业链景气共振阶段，全面看多AI产业趋势。 盘面上，春节后第一个交易日（2月24日），重点布局国产AI产业链的科创人工智能ETF（589520）或 由于海外AI行情扰动，意外下跌，场内价格一度下探3.8%，现跌3.14%，场内频现溢价交 ...

21世纪经济报道新质生产力研究院综合报道 早上好，新的一天又开始了。在过去的24小时内，科技行业发生了哪些有意思的事情？来跟21tech一起看看吧。 【巨头风向标】 2月23日，问界汽车官微发布《关于广东省惠州市车辆起火的说明》。其中称，2026年2月21日，一辆问界M9在广东省惠州市附 近起火，事件未造成人员伤亡。经初步调查，车辆三电系统状态正常，且未见车辆其它自身异常，判断事件非车辆自身原因导 致。说明称，将密切配合有关部门完成后续调查并积极协助用户处理后续事宜。请大家不信谣，不传谣。 问界通报"广东惠州车辆起火事件" 2月22日，智谱发布技术报告，全面解读GLM-5的技术细节。据称，GLM-5是一款旨在推动编程范式从"Vibe Coding"（氛围编 程）转向"Agentic Engineering"（智能体工程）的下一代基础模型。GLM-5 在前代模型 GLM-4.5 的智能体、推理与编程 （Agentic, Reasoning and Coding, ARC）能力基础上，采用稀疏注意力（DeepSeek Sparse Attention，DSA）以大幅降低推理成 本，同时保持长上下文能力无损。为了让 ...

Summary of Key Points from Conference Call Records Industry Overview - The records focus on the global AI industry, highlighting significant trends and developments during the Chinese New Year period from February 16 to February 22, 2023. Major tech companies are accelerating their investments in AI computing power infrastructure [1][2]. Key Companies and Developments - **NVIDIA and Meta Partnership**: On February 18, NVIDIA and Meta announced a multi-year strategic partnership, signing a multi-billion dollar agreement. Meta has become NVIDIA's second-largest buyer of chips, with plans to deploy millions of NVIDIA chips, including Blackwell architecture GPUs and next-generation Rubin architecture GPUs [2]. - **NVIDIA's Upcoming Chip Launch**: NVIDIA's CEO Jensen Huang revealed plans to unveil a "world-first" new chip at the GTC conference on March 15, 2023, with multiple groundbreaking products expected [2]. - **OpenAI's Capital Expenditure Plans**: OpenAI updated investors on its long-term capital expenditure plans, aiming for approximately $600 billion in total computing power spending by 2030, alongside a potential financing round exceeding $100 billion, with NVIDIA negotiating a possible $30 billion investment [2]. Market Performance - Notable stock performance in the optical communication sector during the specified week included significant gains: Lumentum (+18.66%), Coherent (+14.25%), AXT (+22.44%), and AAOI (+16.24%) [1]. Technological Advancements - **AI Model Releases**: Several AI models were released during the period, including Google's Gemini 3.1 Pro and Anthropic's Claude Sonnet 4.6, which is noted for its performance at a fraction of the cost of flagship models [2]. - **ByteDance Innovations**: ByteDance introduced several AI models, including Seedance 2.0 for video generation and Doubao 2.0, which features upgrades in language understanding and logic reasoning [2]. Government and Industry Initiatives - The Indian government announced plans to invest $200 billion in building data centers over the coming years to promote AI industry development, with investments from major companies like Google, Microsoft, and Amazon included in this initiative [2]. Additional Insights - The fourth Global AI Impact Summit took place in New Delhi, attended by prominent figures such as French President Macron and CEOs from Google, OpenAI, and Anthropic. Google’s management compared the current AI wave to a "10x faster industrial revolution," with a significant increase in cloud business backlog orders [2]. - ByteDance reported that its AI interaction numbers reached 1.9 billion on New Year's Eve, showcasing the growing engagement with AI-driven content creation [3]. This summary encapsulates the critical developments and trends in the AI industry as discussed in the conference call records, providing insights into company strategies, market performance, and technological advancements.

Core Insights - Qianxun Intelligent has successfully completed two rounds of financing totaling nearly 2 billion, attracting top-tier investors from various sectors, including leading venture capital firms and state-owned enterprises, indicating strong confidence in the company's technology and growth prospects [1][2][4] Investment Landscape - The financing round reflects a comprehensive ecosystem of support, with top-tier capital betting on technological foresight, industrial capital injecting resources, and state-owned capital providing long-term stability [2][4] - The investment landscape is characterized by a rapid influx of capital into the embodied intelligence sector, with projections indicating that the total financing in this field in China will reach 73.543 billion by 2025, with over 740 investment events [1][2] Technological Advancements - Qianxun Intelligent's Spirit v1.5 model has become the first Chinese open-source model to surpass the performance of Pi0.5, showcasing strong zero-shot generalization capabilities and efficient execution across various tasks [4][5] - The company has significantly reduced data collection costs by 90% through self-developed equipment, enabling the large-scale deployment of real-world data [5][7] - The company emphasizes the importance of diverse data, asserting that "dirty data" is key to scaling VLA models, which has led to a more valuable data philosophy compared to traditional clean data approaches [9] Commercialization and Industry Impact - Qianxun Intelligent has achieved significant breakthroughs in commercialization, with the world's first humanoid embodied intelligence production line operational at CATL's base, demonstrating superior adaptability and efficiency in production processes [9][11] - Collaborations with major companies like JD.com are expanding the application of embodied intelligence in retail, exploring the potential of large-scale deployment in retail networks [11] Team and Expertise - The company boasts a highly skilled team, often referred to as an "AI dream team," with members from top universities and expertise in core areas such as multimodal large models and robotics, positioning them competitively against global leaders [11] - Investors recognize the team's unique combination of academic excellence and practical industry experience, which is crucial for driving innovation in embodied intelligence [11] Future Outlook - Following this financing round, Qianxun Intelligent plans to increase investments in foundational models and real data systems, aiming to deepen industry collaboration and contribute to the development of embodied intelligence as a core infrastructure for the future [11]

Group 1 - The core viewpoint of the article highlights the strong performance of major AI model companies, with Zhiyu (02513) rising by 9.82% to 615 HKD and MINIMAX-WP (00100) increasing by 7.73% to 905.5 HKD [1] - OpenRouter's latest weekly data shows that the total token volume of the top ten models is approximately 8.7 trillion, with Chinese models dominating at 5.3 trillion, accounting for 61% [1] - MiniMax M2.5 ranks first with 2.45 trillion tokens, followed by Kimi K2.5 with 1.21 trillion tokens, while Zhiyu GLM 5 and DeepSeek V3.2 rank third and fifth respectively [1] Group 2 - Guolian Minsheng Securities believes that the trend in the AI industry has not changed this year, transitioning from technological research and development to large-scale commercial implementation [1] - In the long term, the continuous iteration of reasoning and multimodal large models, integration of major ecosystems, and increasing penetration across various industries are expected to drive rapid growth in token usage [1] - The large-scale commercial application and industrial implementation of large models are anticipated to be promising [1]

Core Insights - Anthropic, an AI startup, is offering current and former employees the opportunity to sell company shares, valuing the company at approximately $350 billion [1] - The company has prepared $5 billion to $6 billion for this share sale, depending on the number of eligible employees who choose to sell [1] - Anthropic's latest funding round earlier this month resulted in a post-money valuation of $380 billion [1] Group 1 - The share sale will be conducted by external investors, not by Anthropic itself, and is open to employees who have worked at the company for at least 12 months [1] - Secondary share sales are becoming increasingly common for startups, allowing employees to benefit from company valuation growth even before an acquisition or IPO [1] - This strategy is crucial in the competitive AI talent market, especially as many large startups opt to extend their privatization period [1] Group 2 - Competitors like Stripe Inc. and SpaceX also allow employees to sell shares, while Anthropic's main competitor, OpenAI, regularly conducts share sales, including a $6.6 billion secondary share transaction last year at a $500 billion valuation [2] - Recent actions by Anthropic, OpenAI, and SpaceX indicate preparations for potential future IPOs [2] - Anthropic has released a new tool that significantly reduces the maintenance and analysis costs of COBOL systems, which are critical to IBM's enterprise services [2] - Dario Amodei, head of Anthropic, predicts that AI is on the brink of "exponential growth," with revenue increasing at a rate of tenfold [2]

Group 1 - The core point of the article is the recent changes in the shareholder structure and registered capital of Beijing Huixi Intelligent Information Technology Co., Ltd, indicating potential growth and investment interest in the company [1] - The registered capital of the company increased from approximately 26.09 million RMB to about 30.91 million RMB, reflecting a significant investment [1] - New shareholders include Zhiyuan Innovation (Shanghai) Technology Co., Ltd, Beijing Qishui Huayu Intelligent Technology Partnership (Limited Partnership), and Wuxi Bozhi Future Industry Investment Fund Partnership (Limited Partnership) [1] Group 2 - Beijing Huixi Intelligent Information Technology Co., Ltd was established in April 2022 and is involved in various sectors including integrated circuit design and artificial intelligence software development [1] - The company is co-owned by Wei Wei, Xu Ningyi, and Jingwei Hengrun (688326), indicating a diverse ownership structure [1] - Huixi Intelligent is positioned as a provider of embodied intelligent computing power, with products spanning the fields of embodied intelligence and autonomous driving [1]

Core Viewpoint - The artificial intelligence (AI) industry is experiencing a significant surge, driven by the release of new AI models and a shift in focus towards productivity and efficiency in AI applications [1][2]. Group 1: Market Performance - As of February 24, 2026, the CSI Artificial Intelligence Theme Index (930713) rose by 1.29%, with notable gains from stocks such as Runze Technology (up 20.00%), Beijing Junzheng (up 3.52%), and Xiechuang Data (up 2.88%) [1]. - The AI Artificial Intelligence ETF (512930) increased by 1.10%, with the latest price reported at 2.3 yuan [1]. Group 2: Industry Developments - During the Spring Festival, both domestic and international AI developments accelerated, with major companies releasing updated AI models, including ByteDance's Seedance 2.0, Alibaba's Qwen 3.5, and Google's Gemini 3.1 Pro [1]. - The focus of the current AI model updates has shifted from a "parameter competition" to a "productivity competition," emphasizing the ability of models to solve complex real-world tasks [1]. Group 3: Index Composition - The CSI Artificial Intelligence Theme Index comprises 50 listed companies that provide foundational resources, technology, and application support for AI, reflecting the overall performance of AI-related securities [1]. - As of January 30, 2026, the top ten weighted stocks in the index include Zhongji Xuchuang, Xinyi Sheng, and Hanwujing, accounting for a total of 57.27% of the index [2].

Group 1 - The core viewpoint is that the Spring Festival marks a turning point for the widespread adoption of AI among consumers, with applications evolving from Chat to Agent [1][2] - The competition among internet giants for AI traffic entry is intensifying, and the integration of multiple business lines along with AI empowerment may drive performance and valuation [1][2] - Recommended companies include Alibaba-W and Baidu Group-SW, which are expected to benefit from the continuous improvement of open-source model capabilities and the rapid growth of domestic AI chips [2][5] Group 2 - OpenAI's revenue is projected to exceed its previous target, with a potential valuation of $850 billion following a $100 billion funding round [3] - Google's Gemini3.1Pro has significantly improved its performance in abstract reasoning tests, surpassing competitors like GPT-5.2 and Claude Opus4.6 [3] - Domestic AI companies like Zhiyu and MiniMax have seen substantial stock price increases, reflecting market recognition of their new large models [4] Group 3 - The investment recommendation emphasizes the commercialization of AI and the expansion of application scenarios, with a focus on companies benefiting from IT spending in state-owned enterprises [5] - In the automotive and autonomous driving sectors, the approval of domestic L3-level pilot licenses is expected to accelerate the commercialization of advanced driving technologies [5]

Core Viewpoint - The article discusses the controversy surrounding Anthropic's claims about other AI companies allegedly using its model for distillation, which has sparked significant debate and skepticism within the AI community [1][14]. Group 1: Controversy and Reactions - Anthropic accused DeepSeek, Moonshot AI, and MiniMax of using 24,000 accounts to interact with its model Claude for 16 million times, leading to a public outcry [1][15]. - The public reaction has been largely critical of Anthropic, with many questioning the legitimacy of its claims and expressing support for the accused companies [2][11]. - Some commentators have drawn parallels to reverse engineering, arguing that paying customers should have the right to use products as they see fit [2][16]. Group 2: Model Distillation Explanation - Model distillation is described as a common method in AI training, aimed at creating smaller, more efficient models by transferring knowledge from larger models [13][31]. - The process involves a large model acting as a teacher, simplifying its knowledge for a smaller model, which can then operate more efficiently on less powerful devices [13][31]. - This technique is essential for making AI applications more accessible and usable on everyday devices, thus lowering operational costs and improving performance [29][31]. Group 3: Industry Implications - The ongoing debate highlights the competitive nature of the AI industry, where companies invest heavily to protect their intellectual property while also facing pressure to democratize technology [14][32]. - The article suggests that defining the boundaries of data usage and balancing copyright protection with innovation will be critical challenges for the industry moving forward [14][32].