8 February 2026 | 9:10PM HKT Equity Research Allen Chang +852-2978-2930 | allen.k.chang@gs.com Goldman Sachs (Asia) L.L.C. Verena Jeng Kingsoft Office (688111.SS) AI features upgrade continues; Down to Neutral on fair valuation 688111.SS 12m Price Target: Rmb365.00 Price: Rmb303.62 Upside: 20.2% We remain positive on Kingsoft Office (KO) as the local office software leader, riding on growing paying ratio of ToC (Consumer) users and ToB (Business) client expansion. With rising Gen-AI trend, we see investor c ...

Core Insights - The company initially believed that AI had reached its limits but quickly shifted its perspective after the release of ChatGPT 3.5, recognizing the transformative potential of AI in user interaction [1][2] - The CEO emphasized that the current AI wave could have a more significant impact than previous technological shifts, and the company decided to fully commit to AI [1][2] Strategic Decisions - After deciding to go "ALL in AI," the company faced critical choices regarding how to engage with the AI wave, including whether to develop large models or focus on application [4][6] - The company chose not to create large models but instead to enhance AI applications by collaborating with existing large model providers, positioning itself as an application developer rather than a model creator [6][8] Product Development - The company has focused on refining its WPS AI product, which has evolved through multiple iterations, with the latest version being WPS AI 3.0, aimed at providing a seamless AI-assisted office experience [5][17] - The transition from WPS AI 1.0 to 2.0 involved making AI more user-friendly and integrating it into the office software, allowing users to interact with AI in a more intuitive manner [12][17] Market Positioning - The company has shifted its focus towards the B2B market, launching WPS 365 as a comprehensive AI office platform for organizations, which includes various collaboration tools [18][19] - The company aims to help businesses build their "enterprise brain" by integrating AI capabilities that enhance data management and decision-making processes [22][27] International Expansion - The company is prioritizing international expansion as a core strategy, planning to launch WPS 365 internationally and integrate AI capabilities into its global offerings [35][37] - Despite facing challenges in international markets, the company is committed to leveraging AI to enhance its competitive position globally [38][39] Future Outlook - The company is focused on improving data quality as a key factor for successful AI applications, transitioning from a model-centric to a data-centric approach [27][32] - The ongoing evolution of WPS 365 and its AI capabilities indicates a long-term strategy to capture growth opportunities in both domestic and international markets [39][40]

2023年6月20日，金山办公股价创下历史最高点530.50元。 某券商于2023年6月21日发布的研报《金山办公(688111)：AI+办公商业模式逐步成型WPS AI产品力升 级》表示，维持金山办公"买入"评级。 中国经济网北京1月30日讯金山办公(688111.SH)今日股价下跌，截至收盘报336.84元，跌幅3.21%。 ...

2023年底，ChatGPT以一种炫目的方式闯入大众视野，这一年，被称为AI"元年"，各种大模型的概念火得一塌糊涂。但比起之前有点雷声大雨点小的"元 宇宙"，AI这次带来的突破，可是实打实、持续不断的。 首先，是多模态融合。现在的AI不仅能读懂文字，还能看懂图片、听懂声音。比如，DeepSeek的R2模型就凭本事大大提升医疗报告的解读准确率，并接 入了北京协和医院的影像系统辅助诊断。 时间来到2024年，大家谈论AI的热点悄悄变了。不再只是惊叹于它有多"炫酷"，而是更关心它到底能"干啥"。这一年，AI的发展就像是从一个充满好奇的 少年，开始踏踏实实地干活了。 从"比谁聪明"到"看谁能干"： 2024年的务实转身 技术上，基础模型仍在飞快进步。工程师们不再只追求模型参数的"大"，更在意它能力的"全"。AI的"能力树"上，除了处理文字，理解图像、生成视频甚 至进行复杂推理的"新枝桠"都长了出来。 更关键的是，整个行业的竞争焦点升级了。竞争变成了技术、实际应用和产业生态的"全方位比拼"。国内的智能算力增长特别快，目标很明确：就是要推 动AI从实验室的"性能冠军"，变成各行各业的"得力助手"。 2025年：AI开始" ...

Core Insights - The capital market's enthusiasm for AI has shifted from grand narratives to financial performance, marking a significant transition in 2026 [1] - The AI industry has evolved into distinct segments: computing infrastructure, model building, end-user applications, and regulatory reshaping [2] - The A-share market experienced a structural bull market in 2025, with leading companies like Zhongji Xuchuang and Industrial Fulian achieving significant growth [2][3] Group 1: Market Dynamics - In 2025, the AI sector saw a notable split between A-shares and Hong Kong stocks, with companies like Zhongji Xuchuang and Industrial Fulian leading in valuation restructuring [1] - The successful IPOs of companies like Moer Thread and Muxi Co. on the Sci-Tech Innovation Board indicate a shift of domestic GPUs from experimental stages to large-scale production [3] - The Hong Kong market acted as a bellwether for AI applications, with companies like Cloud Wisdom successfully raising funds and positioning themselves in the AI+ healthcare sector [3] Group 2: Regulatory Environment - The beginning of 2026 brought regulatory scrutiny to industries like food delivery and transportation, impacting companies that rely heavily on algorithm-driven profits [2][7] - The case of Ctrip Group highlights the risks associated with algorithmic monopolies, as the company faced an antitrust investigation that affected its stock price [7] Group 3: Technological Advancements - The focus in 2026 is shifting towards "intelligent agents" and edge hardware, with expectations for a significant drop in development costs for AI agents [8] - The anticipated end of Windows 10 support and a decrease in edge computing costs are expected to drive the adoption of AI PCs and smartphones [8] Group 4: Investment Opportunities - Companies specializing in advanced packaging and new storage solutions, such as Zhaoyi Innovation and Changdian Technology, are expected to benefit from the evolving AI infrastructure [9] - The market is optimistic about the potential of domestic computing capabilities, viewing 2026 as a pivotal year for AI infrastructure development in China [9] Group 5: Future Outlook - The AI industry is expected to transition from speculative narratives to tangible productivity enhancements, with companies that can effectively leverage intelligent agents poised for success [10] - The market is likely to favor businesses that operate within regulatory boundaries and demonstrate real value creation through AI [10]

这次 AGI-Next 前沿峰会，更像一次行业"分岔口"的确认。 在清华大学基础模型北京市重点实验室与智谱联合发起的圆桌上，唐杰、杨植麟、林俊旸、姚顺雨与杨强院士罕见同台。讨论的焦点，不再是谁更接近 AGI，而是中国大模型行业正在不可逆地走向分化。 这种分化，并非简单的 ToB 与 ToC 之争，而是对智能价值如何兑现的不同选择：一条路，把模型嵌入生产流程，押注确定性的生产力；另一条路，进入 真实生活场景，押注体验、规模与长期心智。 几乎在同一时间，智谱与 MiniMax 相继上市，前者锚定 ToB，后者押注 ToC。当峰会共识与上市路径相互印证，中国大模型行业也正式走到了一次真正 的分岔口。 同一张牌桌，不同走法 在AGI-Next峰会的圆桌讨论中，"分化"成为贯穿始终的核心议题。 横跨中美行业的姚顺雨、深耕阿里千问的林俊旸、掌舵智谱的唐杰，以及见证多轮AI周期的杨强，从不同视角解读了ToB与ToC的分化本质，形成了诸多 共识。 作为曾任职于OpenAI、如今扎根国内AI行业的"跨语境者"，姚顺雨对分化的感知最为直接。他在峰会中明确提出："ToC和ToB已经明显发生了分化，这 是最核心的感受。" 在他看来， ...

Core Insights - Kingsoft Office (688111.SH) is a leading provider of office software products and services in China, listed on the Shanghai Stock Exchange in 2019, representing a key player in "hard technology" [1] - The company focuses on digital transformation in education, offering products like the WPS Smart Education Platform and WPS365 Education Edition, serving over 20 million teachers and students across more than 1,000 universities [1][2] Group 1 - The WPS365 Smart Education Platform aims to address challenges in digital transformation within universities, such as system barriers, data fragmentation, and inefficient service responses, by implementing a comprehensive solution combining unified collaboration, lightweight management, and AI services [2] - The platform has been successfully implemented in several universities, including Nanjing University of Posts and Telecommunications and Shandong University, demonstrating significant improvements in operational efficiency and management processes [2][3] Group 2 - The platform enhances efficiency by reducing repetitive tasks for students and faculty, allowing for the generation of full process plans in 5 minutes and social practice briefs in 10 minutes, with an overall efficiency increase of over 80% in tasks like proposal collection and document transfer [3] - AI tools within the platform have significantly improved the efficiency of tasks such as literature processing and resume creation, leading to a substantial reduction in learning and administrative costs [3] - The platform effectively addresses three major pain points in educational digitalization: information fragmentation, passive service, and operational gaps, exemplified by the "Campus Brain" at Nanjing University, which serves 30,000 students and faculty with rapid response capabilities [3]

Core Insights - The report emphasizes that the AI application market is not yet over, with a shift from broad-based gains to a focus on stocks with strong fundamentals expected to occur [2][4]. - Three key factors are identified as catalysts for the continuation of the AI application market: macroeconomic conditions, industry trends, and performance validation [4][5]. Macroeconomic Background - A relatively abundant liquidity environment in the A-share market supports high-growth software applications, with expectations for continued loose monetary policy in 2026 [4]. - The Chinese yuan has been appreciating, and increased geopolitical tensions abroad have heightened global interest in Chinese assets, creating favorable conditions for the stock market [4]. Industry Trends - The underlying technology framework for AI applications is rapidly maturing, with significant improvements in computing efficiency and cost-effectiveness supporting commercialization [5]. - AI application business models are transitioning from concept validation to revenue generation, with some vertical models entering the performance validation phase [5]. Performance Validation - The third-quarter performance of AI applications shows a notable recovery, with revenue growth accelerating from 0.74% in the mid-2025 report to 1.55% in the third quarter, marking the highest level since 2023 [5][6]. - Specific AI vertical companies have demonstrated impressive performance, such as 360's net profit growth increasing from 17.43% in the mid-year report to 78.88% in the third quarter [6]. Focus on Strong Fundamentals - The software application market is expected to shift from broad gains to a focus on stocks with strong fundamentals, similar to trends observed in the 2013-2015 mobile internet boom [9]. - Companies with strong fundamentals in AI applications are concentrated in sectors like "AI + entertainment," "AI + office," "AI + gaming," and "AI + marketing," indicating potential investment opportunities in these niches [10].

Core Insights - The article discusses the rapid integration of AI into daily life and work, highlighting how AI applications are evolving to become essential partners in various tasks, from scheduling to data analysis and meal planning. It emphasizes that 2025 will mark a significant shift where AI will be deeply embedded in everyday activities, reshaping time perception, decision-making, and lifestyle [1]. Group 1: AI Applications and Developments - Doubao has become a "national-level AI application" with over 100 million daily active users and a daily token usage exceeding 50 trillion, making it the largest AI assistant in the Chinese market. Its model has evolved from 1.5 to 1.8, achieving top scores in 38 out of 60 public evaluation benchmarks [2][5]. - The video generation model of Doubao has improved significantly, achieving sound and image synchronization, which enhances creator efficiency by 65% [3]. - Tencent's Yuanbao has integrated AI into social interactions seamlessly, allowing users to interact with AI without needing to download or register a separate application. It can summarize complex documents and transcribe meetings, making AI a natural part of daily communication [6][9]. - Tongyi Qianwen has released models Qwen2.5 and Qwen3, with Qwen3 utilizing approximately 36 trillion tokens for pre-training, significantly enhancing its capabilities in document processing and language understanding [10][13]. - WPS AI has introduced features that allow users to interact with AI directly within the WPS Office suite, enabling document creation and editing through natural language commands, thus lowering the barrier for intelligent office applications [14][17]. Group 2: Market Trends and User Engagement - Manus, an AI agent, experienced a dramatic rise and fall in user engagement, with its monthly visits dropping from 23.76 million to 16.16 million within two months due to performance issues and high operational costs [19][20]. - Zhiyu Qingyan has gained popularity with 7.02 million monthly active users, but it faced scrutiny for violating user privacy by collecting excessive personal information, raising concerns about data security [23][25].

Core Insights - The first batch of large model companies in China, including Zhiyu AI and MiniMax, are set to go public, marking a significant milestone in the AI startup landscape by the end of 2025 [1][5] - These companies are taking different paths in their business models, with Zhiyu AI focusing on technology infrastructure and MiniMax targeting consumer applications [1][2] Company Strategies - Zhiyu AI, founded in 2019, has a strong academic background and focuses on a full-stack self-developed approach, primarily generating revenue through Model as a Service (MaaS), which contributed 84.8% of its revenue in the first half of 2025 [6][10] - MiniMax, established in early 2022, adopts a "technology as product" strategy, with a significant portion of its revenue coming from AI-native products, which accounted for 71.4% of total revenue in 2024 [11][12] Financial Performance - Zhiyu AI's valuation reached 24.4 billion yuan, making it a leader among domestic large model companies, with a market share of 6.6% in the general large model development sector [6][10] - MiniMax reported revenues of $3.46 million in 2023, $30.52 million in 2024, and $53.44 million in the first three quarters of 2025, with overseas revenue becoming a significant contributor [11][12] Market Trends - The global large model market is expected to grow from $10.7 billion in 2024 to $20.65 billion by 2029, with a compound annual growth rate (CAGR) of 80.7% [16] - The Chinese large language model market is projected to reach 53 billion yuan in 2024, with enterprise clients contributing 47 billion yuan [15] Competitive Landscape - Zhiyu AI's focus on B-end clients aligns with the market's payment structure, while MiniMax's C-end strategy allows it to avoid direct competition with giants like OpenAI and Google [15][18] - Both companies are exploring potential expansions into each other's domains, with Zhiyu AI looking to enhance consumer electronics and MiniMax expressing intentions to develop B-end services [21] Challenges and Risks - Zhiyu AI faces risks related to client concentration, with its largest client contributing 19% of revenue and the top five clients accounting for 45.5% [10] - MiniMax is dealing with copyright lawsuits from major film studios, which could pose financial risks despite the company's belief in the lack of merit in these claims [14]