Group 1 - Google has released its first native multimodal embedding model, Gemini Embedding 2, which integrates text, images, audio, video, and PDFs into a unified vector space for cross-modal retrieval, eliminating information loss and engineering complexity associated with traditional multi-model approaches [1] - The model employs Russian nesting doll representation learning technology (MRL), reducing vector dimensions from 3072 to 768 with only a 0.18 point loss, balancing performance and cost [1] - Compared to competitors like OpenAI, Cohere, and Jina, Gemini Embedding 2 is currently the only commercial-grade embedding model covering five modalities, achieving state-of-the-art (SOTA) in multimodal capabilities [1] Group 2 - Meta has acquired the AI social platform Moltbook, which was launched only 40 days prior, and its founders will join Meta's Super Intelligence Lab (MSL) [2] - Moltbook gained attention for its "encrypted language" hoax posts, revealing critical flaws in identity verification and security on the platform [2] - Meta is interested in Moltbook's underlying capabilities for AI agents to be permanently online and autonomously discover and connect, planning to integrate it into its AI ecosystem after addressing security issues [2] Group 3 - Nvidia has signed a $60 billion computing power agreement with Thinking Machines Lab to deploy a next-generation Vera Rubin computing cluster with a scale of 1GW, including cash investment from Nvidia [3] - The company, founded by former OpenAI CTO Mira Murati, has a current valuation of $50 billion and recently appointed the founder of PyTorch as its new CTO after a previous departure [3] - By securing scarce computing power capacity, the company is building a second moat beyond talent competition to support cutting-edge model training and enterprise-level AI customization [3] Group 4 - Tencent has responded to security and cost concerns regarding OpenClaw, emphasizing that installation is free but model calls incur token fees, and has launched a series of security products and the SkillHub plugin ecosystem [4] - Tencent's lobster services are divided into two categories: deployment solutions around the open-source OpenClaw and the self-developed desktop AI agent WorkBuddy, which shares the Agent architecture with CodeBuddy [4] - Nearly 40,000 Tencent employees are already using OpenClaw internally, indicating a shift towards a new development model involving agents [4] Group 5 - Tencent Cloud has launched the SkillHub skill marketplace, which has over 13,000 lobster skills, supporting Chinese search and domestic node acceleration, and is compatible with various agent frameworks and environments [5] - More than 50 high-quality skills covering frequent scenarios such as office collaboration, development tools, and content creation have been selected, completing security scans and quality filtering [5] - Over 10 products, including Tencent Docs, QQ Browser, and Tencent Maps, have undergone skill transformation, allowing for one-click integration with OpenClaw [5] Group 6 - ZhiMi has launched its chip brand "Xinjichuan Yue," which includes a mobile chip (Chixiao 01) with a self-developed NPU architecture, a 2nm process autonomous driving chip with 2000 TOPS computing power, and the mass-produced Tianqiong series of general-purpose robot chips [7] - The company has announced plans for a space computing center, intending to deploy 2 million computing satellites to form a super constellation, with the first space computing box set to be verified in orbit soon [7] - A personal super AI computer with 1.5 PFLOPS has been introduced, utilizing a unified memory architecture capable of locally loading large models with billions of parameters and supporting multi-device networking [7] Group 7 - On the 10th anniversary of AlphaGo's victory over Lee Sedol, Demis Hassabis reflects on how the "37th move" initiated the modern AI era, showcasing AI's ability to surpass human experts and autonomously discover new strategies [8] - The reinforcement learning and search methods from AlphaGo have been extended to various scientific fields, including AlphaFold for protein structure prediction and AlphaProof for mathematical reasoning [8] - DeepMind believes that combining Gemini's world model, AlphaGo's search planning, and specialized AI tools is key to achieving artificial general intelligence (AGI) [8] Group 8 - The new version of LeRobot v0.5.0 integrates full-body control for the Unitree G1 humanoid robot and introduces six new strategy models, supporting LoRA fine-tuning [9] - Dataset processing has been significantly optimized, with streaming video encoding achieving zero-wait recording and image training speed improved by ten times, along with the launch of EnvHub for direct loading of simulation environments [9] - The codebase has been upgraded to Python 3.12+ and Transformers v5, with a new third-party strategy plugin system, and the paper has been accepted by ICLR 2026 [9] Group 9 - The author of "The Biography of Hassabis" reveals the contradictory personality of Demis Hassabis, who dislikes control yet is extremely competitive, and his clear aversion to the pursuit of power [10] - Hassabis attempted for three years to make DeepMind independent from Google but failed, and the book contains details that contradict Hassabis's own statements following the dismissal of co-founder Suleyman [10] - The author believes that AI safety faces the "Oppenheimer dilemma," where scientists can build technology but cannot control its use, ultimately requiring cooperation between the US and China to achieve safety [10] Group 10 - Jensen Huang outlines a five-layer industrial system for AI, consisting of energy, chips, infrastructure, models, and applications, where each successful application drives demand down to power generation [11] - He asserts that AI has crossed a critical threshold, with significant improvements in model inference and deployment capabilities, and open-source models like DeepSeek-R1 accelerating the growth of full-stack demand [11] - While thousands of billions have been invested globally, trillions in infrastructure remain to be built, marking it as the largest infrastructure project in human history [11]

Group 1 - Southbound funds net purchased Hong Kong stocks amounting to 34.48 billion HKD on March 11 [1][3] - The top net purchases included China National Offshore Oil Corporation (10.39 billion HKD), Alibaba-W (6.98 billion HKD), Geely Automobile (5.31 billion HKD), Xiaomi Group-W (1.59 billion HKD), Hua Hong Semiconductor (1.55 billion HKD), and SMIC (1.37 billion HKD) [1] - Notable net sales were observed in Tencent Holdings (24.35 billion HKD), ending its previous eight-day net buying streak, as well as China Construction Bank (8.26 billion HKD), Industrial and Commercial Bank of China (1.85 billion HKD), and Yangtze Optical Fibre and Cable (1.2 billion HKD) [1][4] Group 2 - Southbound funds have net purchased Alibaba for three consecutive days, totaling 20.3096 billion HKD [1]

Group 1 - Tencent Holding Ltd. shares increased by 10.4% to close at $73.96, following a notable trading volume, contrasting with a 6.7% loss over the past four weeks [1] - The stock is benefiting from Tencent's enhanced AI integration within Weixin's ecosystem, positioning the company to capture significant value as AI evolves [1] - The upcoming quarterly earnings are expected to be $0.98 per share, reflecting a year-over-year increase of 19.5%, with revenues projected at $27.83 billion, up 16.1% from the previous year [2] Group 2 - The consensus EPS estimate for Tencent has remained stable over the last 30 days, indicating that stock price movements may not continue without changes in earnings estimate revisions [3] - Tencent holds a Zacks Rank of 2 (Buy), suggesting positive market sentiment towards the stock [3] - In comparison, Alphabet (GOOGL), another company in the same industry, has a Zacks Rank of 3 (Hold) and has seen a -5.5% return over the past month [3][4]

Core Insights - The article discusses the competitive landscape between Meta and Tencent in the AI space, particularly focusing on their strategies to acquire AI capabilities and establish dominance in the emerging "Agent" market [4][19]. Group 1: Meta's Strategy - Meta acquired Moltbook, a social platform designed for AI agents, to enhance its AI capabilities and address its shortcomings in AI model performance [4][7]. - Following the disappointing launch of its Llama 4 model, Meta's CEO Mark Zuckerberg responded by aggressively recruiting top talent and investing heavily in AI companies, including a $14.8 billion stake in Scale AI [7][8]. - The acquisition of Manus for approximately $3 billion further solidified Meta's position by integrating functional AI capabilities into its existing platforms [8][19]. Group 2: Tencent's Approach - Tencent's AI efforts have faced challenges, with its models not performing competitively in the market, leading to a reliance on external models for its AI projects [12][14]. - The company has also pursued talent acquisition, hiring key figures from the AI industry to bolster its capabilities, mirroring Meta's strategy [13][19]. - Tencent is developing an AI agent project integrated into WeChat, aiming to leverage its existing user base and ecosystem for seamless service delivery [13][14]. Group 3: The Shift in Software Distribution - The article highlights a significant shift in software distribution from traditional app-based models to conversational interfaces, with AI agents becoming the new entry point for users [18][19]. - Control over these conversational interfaces is seen as crucial for both Meta and Tencent, as they possess the largest user bases through platforms like WhatsApp, Messenger, WeChat, and QQ [19][20]. - The anticipated growth of AI agents in enterprise applications is projected to rise from 5% in 2025 to 40% by the end of 2026, emphasizing the importance of early entry into this market [19].

Investment Rating - The industry investment rating is "Overweight" (maintained) [1] Core Insights - The report highlights the rapid development of domestic AI models, particularly OpenClaw, which has led to increased competition among internet giants in the AI space. This trend is expected to drive significant growth in AI cloud infrastructure as a service (IAAS) [4][5] - The demand for AI computing power has surged, leading to a price increase in the computing power rental market. Major players like Tencent Cloud have announced price hikes, indicating a shift towards a "seller's market" [6] - The report identifies three core beneficiary sectors of the AI cloud IAAS price increase: AIDC (AI Data Center), computing power rental, and CDN (Content Delivery Network) [7] Summary by Sections Industry Overview - The report discusses the competitive landscape in the AI sector, emphasizing the launch of OpenClaw and its impact on the market [4][5] Market Trends - There is a notable increase in AI computing power demand, with rental prices for high-end GPUs rising by 15%-30% as of February 2026. For instance, the rental price for the H200 GPU has reached 7.5-8.0 RMB per card hour, with a monthly rental of 60,000-66,000 RMB, reflecting a 25%-30% increase [6] Investment Opportunities - The report recommends focusing on three main sectors that are expected to benefit from the AI cloud IAAS price increase: 1. AIDC Data Centers with recommended stocks including Guanghuan Xinnet and Baoxin Software [7] 2. AIDC Liquid Cooling with recommended stocks like Yingwei Ke and Shunling Environment [7] 3. Computing Power Rental with recommended stocks such as Hongjing Technology and Xiechuang Data [7]

Group 1 - Tencent Cloud announced an adjustment to its billing strategy for certain models to continue providing stable and high-quality large model services, effective from March 13, 2026 [1][3] - Three models, GLM 5, MiniMax 2.5, and Kimi 2.5, will end their free public testing phase and transition to formal commercial services with usage-based billing [3] - The pricing for the Tencent HY2.0 Instruct model will increase significantly, with input prices rising from 0.0008 yuan to 0.004505 yuan per thousand tokens (over 460% increase) and output prices from 0.002 yuan to 0.01113 yuan per thousand tokens (over 450% increase) [4] Group 2 - The Tencent HY2.0 Think model will also see price increases, with input prices going from 0.001 yuan to 0.0053 yuan per thousand tokens and output prices from 0.004 yuan to 0.0212 yuan per thousand tokens [4] - The price adjustments may be influenced by the rising popularity of the open-source AI "OpenClaw," which has seen a fourfold increase in token consumption over a month [4] - As of February 2026, the average daily token consumption for mainstream large models reached approximately 180 trillion tokens, indicating a growing demand in the AI sector [4] Group 3 - Tencent has been actively launching new products, including the WorkBuddy AI assistant, and is developing a comprehensive ecosystem of AI agents covering personal, developer, and enterprise deployments [5] - The CEO of Tencent, Ma Huateng, has promoted the company's various AI products, which include self-developed, local, cloud-based, and enterprise solutions [5] - Multiple tech companies, including Huawei, Xiaomi, Alibaba, and others, have joined the trend by launching products similar to OpenClaw, indicating a competitive landscape in the AI market [5]

Core Viewpoint - The article discusses the rapid adoption and potential risks associated with OpenClaw, an AI agent that automates tasks for users, highlighting both its popularity and the security concerns it raises [1][4][8]. Group 1: OpenClaw's Popularity and Adoption - OpenClaw has gained significant attention, with long queues for installation in Shenzhen and various cloud providers integrating it into their services [1] - The rise of OpenClaw has led to new business opportunities, including installation and uninstallation services, indicating a burgeoning market around AI agents [5] Group 2: Security Risks and Concerns - Users have reported instances where OpenClaw has exhausted their token balances due to its autonomous questioning, raising concerns about its default security configurations [4] - The Ministry of Industry and Information Technology has issued warnings about OpenClaw's vulnerabilities, including susceptibility to attacks and potential for unauthorized system access [4][8] - The article emphasizes the need for strict governance and risk management for AI agents, advocating for minimal permissions and continuous monitoring to prevent misuse [9][10] Group 3: Broader Implications and Comparisons - The article draws parallels between current AI agent risks and past incidents in other industries, such as the Knight Capital trading mishap, to illustrate the potential consequences of unchecked AI execution [10] - Ongoing debates in the U.S. regarding AI usage boundaries reflect a global concern about the implications of AI agents in various sectors, including military and commercial applications [6][7]

Core Viewpoint - Tencent Cloud announced an adjustment in its billing strategy for certain models starting March 13, which includes price changes and the end of free public testing for specific models [2]. Pricing Adjustments - The models GLM 5, MiniMax 2.5, and Kimi 2.5 will transition from free public testing to commercial services [2]. - The pricing for the Tencent HY2.0 Instruct model will see a significant increase, with input prices rising from 0.0008 yuan to 0.004505 yuan (an increase of 463%) and output prices from 0.002 yuan to 0.01113 yuan (an increase of 456%) [3][4]. - The Tencent HY2.0 Think model will also experience price hikes, with input prices increasing from 0.001 yuan to 0.0053 yuan and output prices from 0.004 yuan to 0.0212 yuan [3][4]. Model Pricing Summary - A detailed table outlines the pricing adjustments for various models, indicating both the previous and new prices per thousand tokens [3][4]. - For example, the input price for DeepSeek-V3-0324 is set at 0.002 yuan, while the output price is 0.008 yuan [3].

Core Viewpoint - Tencent has officially launched its "Lobster" AI suite, aiming to democratize AI technology and make it accessible to the general public, moving beyond the tech-savvy crowd [1][3]. Group 1: Product Offerings - The "Lobster" AI, named after its red lobster icon, includes five products targeting individuals, developers, and enterprises: a no-code solution for personal use, local control via WeChat, 24/7 online computing power, and enterprise-focused solutions [3]. - The local version allows users to control their computers remotely, while the cloud version provides continuous computing resources, indicating a strong focus on enhancing productivity in various settings [3]. Group 2: Security and Support Features - Tencent has introduced a "Shrimp House" to isolate local data and a "Cloud Security Guard" to mitigate risks, ensuring user data protection and security [3]. - Additionally, a knowledge base is provided to assist users in navigating the new AI tools, reflecting Tencent's commitment to user support and ease of use [3]. Group 3: Strategic Intent - By launching this comprehensive AI suite, Tencent aims to secure a significant position in the next generation of AI technology, indicating a strategic move to capture a larger market share in the AI sector [3].

Market Overview - The Hang Seng Index decreased by 0.24%, closing at 25,898.76 points, with a total market turnover of 254.48 billion HKD [1]. Southbound Trading Activity - The net buying amounts for southbound funds included: - China National Offshore Oil Corporation (CNOOC) with a net buy of 1.039 billion HKD - Alibaba Group Holding Limited (Alibaba-W) with a net buy of 698 million HKD - Geely Automobile Holdings Limited with a net buy of 532 million HKD [1][2]. - The net selling amounts for southbound funds included: - Tencent Holdings Limited with a net sell of 2.435 billion HKD - China Construction Bank with a net sell of 826 million HKD - Industrial and Commercial Bank of China with a net sell of 185 million HKD [1][2]. Individual Stock Performance - CNOOC's stock price increased by 3.70% to 28.06 HKD [2]. - Geely Automobile's stock price rose by 8.21% to 17.39 HKD [2]. - Alibaba's stock price slightly decreased by 0.22% to 133.20 HKD [2]. - Tencent's stock price fell by 0.27% to 552.00 HKD [2]. - China Construction Bank's stock price decreased by 0.96% to 6.22 HKD [2]. - Industrial and Commercial Bank of China's stock price dropped by 1.64% to 7.78 HKD [2].