Anthropic表示，外国AI实验室通过约24，000个虚假账户进行了约1600万次交互，以提取Claude的能 力，并警告称，非法蒸馏的模型可能缺乏安全保障，构成国家安全风险。 责任编辑：张俊 SF065 Anthropic表示，外国AI实验室通过约24，000个虚假账户进行了约1600万次交互，以提取Claude的能 力，并警告称，非法蒸馏的模型可能缺乏安全保障，构成国家安全风险。 责任编辑：张俊 SF065 ...

Core Insights - Anthropic is collaborating with Intapp to launch AI agents targeted at professional firms [1] - The company has developed various plugins in partnership with FactSet and other collaborators [1] - Anthropic is integrating AI agents with tools designed for investment banking and human resources [1]

SAN FRANCISCO, Feb 24 (Reuters) - Artificial intelligence lab Anthropic on Tuesday unveiled 10 new ways for business customers to plug in its technology to key ‌areas of their work, weeks after other releases sparked an aggressive selloff ‌of traditional software company shares. The San Francisco-based startup said its plug-ins could now help with investment banking tasks ​like reviewing deals, wealth-management tasks such as portfolio analysis and human resource-related tasks such as making new-hire mate ...

市场有风险，投资需谨慎。本文不构成个人投资建议，也未考虑到个别用户特殊的投资目标、财务状况或需要。用户应考虑本文中的任何 意见、观点或结论是否符合其特定状况。据此投资，责任自负。 Anthropic将与Intapp面向专业公司推出AI agents。Anthropic与FactSet等合作伙伴开发出多种插件， Anthropic将AI agent与面向投行、HR的工具关联。 风险提示及免责条款 ...

证券研究报告 2026.02.11 人工智能十年展望（二十七）：越过"遗忘" 的边界，模型记忆的三层架构与产业机遇 SAC 执证编号：S0080518070011 SFC CE Ref：BOP246 于钟海 分析员 韩蕊 分析员 王之昊 分析员 SAC 执证编号：S0080523070010 SFC CE Ref：BXD683 rui.han@cicc.com.cn SAC 执证编号：S0080522050001 SFC CE Ref：BSS168 zhihao3.wang@cicc.com.cn 纵轴：相对值（%） 88 100 112 124 136 148 2025-02 2025-05 2025-08 2025-10 2026-01 沪深300 中金软件及服务 投资建议 大模型的演进史，本质上是一部与"遗忘"抗争的历史。当我们惊 叹于模型的推理能力时，往往忽视了一个重要短板：在缺乏记忆留 存的架构下，模型每一次对历史信息的处理，本质上都是一次昂贵 的"重复计算"。这种以高昂算力对抗遗忘的粗放模式，正面临着显 存墙与上下文窗口的物理极限。我们认为，2026 年及之后的AI Infra主战场将增加"模型记 ...

Summary of Key Points from the Conference Call Industry Overview - The focus is on the **AI** industry, specifically the themes of **AI Adopters** and **AI Infrastructure**. Concerns regarding **AI disruption** are affecting various service industries, including financial advisory and brokerage [2][4]. Core Insights - **AI Disruption Concerns**: Market fears about AI disruption have expanded beyond software to include a broader range of service industries. This has led investors to reassess the earnings durability of service-oriented business models, impacting their performance [2]. - **Service Exposure**: The concept of 'service exposure' has emerged as a critical factor influencing thematic performance. The analysis categorizes 25 GICS Industry Groups into Service and Non-Service categories, revealing that portfolios with higher service exposure have underperformed [3]. - **Performance Metrics**: The **AI Adopters** sub-theme has a service exposure of **53%**, significantly higher than other themes. This high exposure introduces uncertainty regarding competitive dynamics and pricing sustainability, which is reflected in current market pricing [4]. - **AI Infrastructure**: In contrast, the **AI Infrastructure** theme has the lowest service exposure at **14%** and has shown the strongest year-to-date performance. This theme is less vulnerable to service disruption risks and benefits from ongoing capital expenditure and demand for compute, semiconductors, and enabling hardware [5]. Quantitative Analysis - Six quantitative lenses are used to evaluate the attractiveness of themes: 1. Information ratio 2. Earnings revision breadth 3. Bottom-up estimates 4. Valuation 5. Mutual fund positioning 6. Factor exposures - **AI Infrastructure** stands out with strong risk-adjusted performance, favorable earnings revisions, and positive fund positioning among outperforming mutual funds [6]. Additional Insights - The year-to-date performance of thematic portfolios is inversely related to their service exposure, indicating that lower service exposure correlates with better performance [3][8]. - The report emphasizes the importance of being selective within the **AI Adopters** theme, recommending a focus on companies that possess pricing power to navigate the uncertainties introduced by AI adoption [4]. Conclusion - The analysis indicates a clear distinction between the performance and risk profiles of **AI Adopters** and **AI Infrastructure**. Investors are advised to consider service exposure as a key factor in their investment decisions within the AI sector [2][5].

Ronald Keung, CFA 23 February 2026 | 6:46AM HKT Equity Research +852-2978-0856 | ronald.keung@gs.com Goldman Sachs (Asia) L.L.C. MiniMax Group (0100.HK) Eric Sheridan Global full-modal AI company reaching hypergrowth stage; initiate at Neutral on valuation 0100.HK 12m Price Target: HK$1,018.00 Price: HK$970.00 Upside: 4.9% We initiate on MiniMax with DCF-based 12-m TP of HK$1,018 per share, at a sector-relative Neutral rating, on balanced risk-reward based on our bull/bear framework. MiniMax is one of the t ...

Summary of Key Points from the Document Industry Overview - The document discusses various breakthrough technologies and innovations, particularly focusing on advancements in AI, energy, and materials science, as well as the emerging space industry in Southeast Asia. Core Insights and Arguments 1. **AI and Energy Demand**: The increasing energy requirements of AI data centers are highlighted, raising questions about the sustainability of current energy sources and the potential for breakthroughs in nuclear power to meet these demands [1][2][3]. 2. **Breakthrough Technologies**: The document lists ten technologies expected to significantly impact the world, including quantum computing, carbon capture, gene editing, and nuclear fusion, emphasizing their potential to address global challenges [11][12][13]. 3. **Radiative Cooling Technologies**: Innovations in radiative cooling, which can reduce air conditioning energy needs by 15% to 20%, are discussed. These technologies utilize materials that reflect sunlight and dissipate heat without additional energy [44][50][51]. 4. **Southeast Asia's Space Industry**: The Thai Space Expo showcases the growing interest and investment in the space sector within Southeast Asia, with potential developments such as a spaceport in Thailand and startups exploring satellite data applications [99][100][102]. Additional Important Content 1. **Technological Skepticism**: The document reflects on the past decade's focus on business model disruptions rather than true technological breakthroughs, suggesting a shift back towards meaningful innovations [9][10]. 2. **Environmental Concerns**: The limitations and environmental risks associated with some cooling technologies, particularly those relying on non-biodegradable materials, are acknowledged, indicating the need for sustainable solutions [51]. 3. **Cultural Impact of Emotions**: The emergence of new emotional vocabulary, termed "neo-emotions," is discussed, highlighting how online interactions shape emotional expressions and societal connections [71][76][79]. This summary encapsulates the key themes and insights from the document, focusing on the implications of technological advancements and their potential impact on various industries and societal issues.

February 19, 2026 10:37 PM GMT Knowledge Atlas Technology JSC Ltd | Asia Pacific A Global AI Foundation Model Leader with Favorable Mix Shift With leading AI foundation model capabilities and a favorable mix shift, we see rapid revenue growth ahead for Z.ai despite limited US exposure. As a pure play, we believe it also commands a scarcity valuation premium. Key Takeaways A global AI foundation model leader with deep academic roots in Tsinghua University: Z.ai (also known as Zhipu and Knowledge Atlas) consi ...

Summary of MiniMax Group Conference Call Company Overview - **Company**: MiniMax Group - **Industry**: AI Foundation Model Development - **Rating**: Initiated coverage with a Buy rating and a price target of HK$1,000 Core Insights - **Market Position**: MiniMax is positioned to benefit from the global AI development trend, focusing on proprietary and frontier AI models for both consumer applications and enterprise solutions [2][10] - **Revenue Growth Forecast**: Projected revenue growth of US$73 million in 2025, US$209 million in 2026, and US$809 million in 2027, indicating a 232% CAGR [4][55] - **Commercialization Strategy**: Emphasis on scaling commercialization through a comprehensive product offering and a global go-to-market strategy, with a focus on enterprise-facing (B2B) and consumer-facing (B2C) services [10][32] Technological Innovation and R&D - **Competitive Advantages**: MiniMax's strengths include technological leadership in model intelligence, efficient R&D processes, and rapid commercialization [3][10] - **Model Performance**: MiniMax's models have achieved top rankings in various modalities, including first place in open-source reasoning models and significant positions in image-to-video and text-to-speech categories [14][25] - **R&D Investment**: The company maintains a high level of R&D investment to support continuous model iterations, which may impact near-term profitability but is expected to enhance long-term growth [10][41] Financial Metrics - **Valuation**: Price target of HK$1,000 based on a 125x P/ARR multiple applied to projected ARR of US$318 million for December 2026 [5] - **Market Capitalization**: Approximately HK$266 billion (US$34 billion) with a share price of HK$847 as of February 16, 2026 [6][7] - **Gross Margin Expectations**: Anticipated expansion from 22% in 2025 to 43% by 2027, driven by improved efficiency and scale benefits [4][55] Risks and Challenges - **Market Risks**: Acknowledgment of macroeconomic, regulatory, and geopolitical uncertainties, as well as competitive pressures in the rapidly evolving AI landscape [3][10] - **Investor Sentiment**: Market skepticism regarding MiniMax's ability to sustain its technological leadership and commercial viability due to its relatively small scale and short operating history [36][37] Revenue Breakdown and Growth Drivers - **Revenue Composition**: Expected increase in revenue contribution from the Open Platform business from 32% in 2025 to 49% in 2027, with overall revenue growth driven by user expansion and improved monetization [4][55] - **User Metrics**: By September 30, 2025, MiniMax's AI-native products had served over 212 million individual users and more than 100,000 enterprise customers globally [44][43] Conclusion - **Investment Outlook**: MiniMax is well-positioned to capture growth in the AI foundation model sector, with significant upside potential despite current premium valuations. The company's focus on technological innovation, efficient R&D, and a robust commercialization strategy supports a positive long-term outlook [10][17]