Group 1 - The Hong Kong stock market saw a strong performance on January 12, with significant gains in tech, media, and aerospace sectors, particularly AI concept stocks [1] - Notable performers included Zhihui, which surged nearly 38%, and MINIMAX, which rose over 25% [1] - Other companies that experienced substantial gains include Lion Group (up over 28%), Kingsoft Cloud (up 13%), Meitu (up 11%), and several others in the tech sector [1][2] Group 2 - Bloomberg industry research indicates that the earnings growth of Chinese tech giants is expected to surpass that of the "seven giants" in the US stock market for the first time since 2022, with a significant turning point anticipated in 2026 [3] - DeepSeek is set to release its next-generation V4 model in mid-February, which is expected to excel in programming tasks, long code prompt processing, and improved data pattern understanding [3] - The development of large models is accelerating the commercialization of AI applications, particularly in search & marketing, coding, multimodal applications, agents, and AI for science [3] - The evolution of large models is seen as a key tool for companies with deep industry know-how, particularly in AI healthcare, which is transitioning from a single auxiliary tool to comprehensive lifecycle management [3]

Core Viewpoint - The stock price of Beijing Zhiyun Huazhang Technology Co., Ltd. (referred to as "Zhiyun") has increased by 31.90%, reaching HKD 209.20 per share, with a total market capitalization of HKD 92.096 billion following the announcement of a strategic partnership with Didi to explore key technologies in general artificial intelligence (AGI) and their applications in the transportation sector [2]. Group 1 - Zhiyun's stock price rose significantly, reflecting positive market sentiment [2]. - The total market capitalization of Zhiyun reached HKD 92.096 billion, indicating strong investor interest [2]. - The strategic cooperation with Didi focuses on advancing AGI technologies and their application in intelligent agents within the transportation industry [2]. Group 2 - The collaboration aims to promote the implementation of intelligent agent scenarios and talent development in large model fields [2]. - The partnership will enhance the alignment of intentions and reasoning capabilities in transportation scenarios [2]. - The initiative seeks to validate and implement intelligent agents in more complex business environments [2].

Core Viewpoint - The stock price of Zhipu (2513.HK), known as the "first global large model stock," surged by 45.34% to 230.6 HKD on its third trading day, reaching a market capitalization of 101.4 billion HKD [1] Group 1: Strategic Partnerships - Zhipu has announced a strategic partnership with Didi to explore key technologies in General Artificial Intelligence (AGI) and their applications in the transportation sector [1] - The collaboration aims to advance the implementation of Agent scenarios and talent development in the large model field, enhancing intention alignment and reasoning capabilities in transportation contexts [1] Group 2: Company Achievements - Zhipu is one of the earliest companies in China to engage in large model research, having introduced the GLM pre-training paradigm and released several significant models, including the first hundred-billion model and the first open-source trillion model [1] - The company is recognized for maintaining a comprehensive model system that aligns with global top-tier standards, earning it the title of "China's OpenAI" [1] Group 3: Financial Projections - Zhipu is projected to exceed 100 million USD in revenue by 2025, with local deployment revenue accounting for 85% and cloud revenue increasing from 0% to 15% over the past two years [1] - The cloud revenue share is expected to continue rising this year, driven by high-cost performance coding tools [1]

Group 1 - The AI application and software sector is experiencing significant growth, with leading ETFs such as the software leader ETF (159899) rising over 8% and the cloud computing ETF (159890) increasing by 6.76% [1][2] - The cloud computing ETF (159890) tracks the cloud computing index, with IT services, horizontal general software, and vertical application software accounting for approximately 65% of its composition, indicating a deep investment in AI applications [1][4] - Major stocks in the software leader ETF include companies like Tuowei Information and Kingsoft, which are benefiting from the acceleration of AI applications and domestic replacements [2][6] Group 2 - The global AI computing power platforms are advancing, with Nvidia and AMD unveiling new platforms that are expected to transform AI computing capabilities [3] - The introduction of open-source models like DeepSeek is anticipated to enhance the development speed of vertical applications, allowing companies to explore and adjust applications independently [3][4] - The software leader ETF closely tracks the CSI All-Share Software Index, covering various software sectors, and is positioned as an efficient tool for investing in the AI and software sectors [6] Group 3 - The upcoming launch of the next-generation flagship model DeepSeek-V4 is expected to enhance code generation and long code context processing capabilities [7] - New player MiniMax has seen its market value exceed 100 billion HKD after a significant stock price increase, setting a new benchmark for AI model and application companies in China [7] - The competition among China's "AI four giants" has shifted from "Chat" to the "Agent" phase, indicating a new direction in AI model development [7]

Market Overview - As of January 12, the Hong Kong stock market's major indices showed an upward trend, with the Hang Seng Index at 26,456.75 points, up by 0.86% [1] - The Hang Seng Tech Index rose by 2.08% to 5,805.57 points, while the State-Owned Enterprises Index increased by 1.08% [1] Sector Performance - Technology stocks were the main support for the market, with Meituan's stock price rising over 7%, Alibaba increasing by more than 5%, and Kuaishou and Bilibili both up over 5% [3] - Baidu Group saw an increase of over 4%, while JD Group and Tencent Holdings both rose by more than 2% [3] - The artificial intelligence application sector was particularly active, with stocks of Zhipu and MINIMAX leading the gains, rising over 30% and 31% respectively [3] - Other sectors such as heavy machinery, semiconductors, mobile gaming, and Apple-related concepts also performed positively [3] - Conversely, real estate stocks, insurance stocks, and brain-computer interface concept stocks showed relatively weak performance [3] New Listings - OmniVision Technologies, a global image sensor supplier, made its debut on the Hong Kong Stock Exchange, becoming the first "A+H" listed company of the year, with its stock price rising over 15% at midday [3] - The company issued a total of 45.8 million H shares at an offering price of HKD 104.80 per share [3]

Core Viewpoint - Zhiyuan (02513) has seen its stock price surge over 45%, nearing a doubling from its IPO price of 116.2 HKD, with a market capitalization exceeding 100 billion HKD [1] Company Overview - Zhiyuan is one of the earliest companies in China to engage in large model research and development, having pioneered the general pre-training paradigm GLM based on autoregressive fill-in-the-blank [1] - The company has released several significant models, including China's first hundred-billion model, the first open-source trillion model, the first dialogue model, the first multimodal model, and the world's first device control agent, establishing a comprehensive model system [1] - Zhiyuan is recognized for maintaining a competitive edge in original technology, aligning with global standards, and is often referred to as "China's OpenAI" [1] Strategic Partnerships - Recently, Zhiyuan announced a strategic partnership with Didi, focusing on collaborative exploration of key technologies in general artificial intelligence (AGI) and their applications in the transportation sector [1] - The partnership aims to advance the implementation of agent scenarios and talent development in the large model field, enhancing intent alignment and reasoning capabilities in transportation scenarios [1] - The collaboration will also facilitate the validation and implementation of agents in more complex business environments [1]

本文源自：金融界AI电报 "全球大模型第一股"智谱(2513.HK)今日为上市第三个交易日，股价继续上涨，盘中一度飙升45.34%至 230.6港元，再创上市新高，市值升至1014亿港元。 ...

Core Viewpoint - Zhiyu (02513) has seen its stock price surge over 45%, nearing a doubling from its IPO price of 116.2 HKD, with a market capitalization exceeding 100 billion HKD [1] Company Overview - Zhiyu is one of the earliest companies in China to engage in large model research and development, having pioneered the general pre-training paradigm GLM based on autoregressive fill-in-the-blank [1] - The company has released several significant models, including China's first hundred-billion model, the first open-source trillion model, the first dialogue model, the first multimodal model, and the world's first device control intelligent agent, establishing a comprehensive model system [1] - Zhiyu is recognized for maintaining a level of original technology that is on par with global leaders, earning it the title of "China's OpenAI" [1] Strategic Partnerships - Zhiyu has recently announced a strategic partnership with Didi, focusing on collaborative exploration of key technologies in general artificial intelligence (AGI) and their application in the transportation sector [1] - The partnership aims to advance the implementation of agent scenarios and talent development in the large model field, enhancing intent alignment and reasoning capabilities in transportation scenarios [1] - The collaboration will also facilitate the validation and implementation of agents in more complex business environments [1]

大模型不赚钱，这事不多说，但简单回顾： OpenAI据说得2029年盈利，每年亏140亿刀。 Anthropic据说2028年能盈利，也亏损。 刚上市的MiniMax每年亏5个亿刀。 智谱和MiniMax差不多，也是每年亏这么多。 据说DeepSeek能赚钱，但这行当基础模型基本亏损是确定的。 不能赚钱就主要靠VC，国内VC不给力所以就导致了基模发展的一系列问题。 但基模属于非典型产品，不多说了，下面重点说智能体。（包括做基础模型也做应用的） 智能体也先说结果： 就拿我比较熟悉的Glean举例子，ARR最新报道过了2亿美金，我们就假设每年都是这么多收入，那现在 有1000个人，所以每个人对应收入是20万美金。这在AI行当基本会亏损。对应的事件呢，我觉得接下 来大概率会继续融资。 类似的，其他明星智能体公司也基本这样，大额融资的包括并购的... 如果就这么多，那就和上波AI有点像，还好不全是，比如： Midjourney一年赚了5亿美金的时候只有40个人，这就怎么都赚钱的。因为它流量基本还是自然流。 你比如Base44，一年350万美金，还就1个人，这怎么也是赚钱的，不太可能都是投流投出来的。 这样题目从结果看 ...

Core Viewpoint - The article highlights a significant shift in market dynamics, with AI application stocks experiencing a major surge while AI hardware stocks decline, indicating a potential turning point in investment focus towards AI applications [2][5][12]. AI Application Surge - AI application sectors, including Sora concept, AI+ marketing, and AI intelligent agents, saw substantial gains, with leading stocks like BlueFocus and Kunlun Wanwei hitting "20CM" limit up, and BlueFocus achieving a transaction volume of 193.2 billion yuan, the highest in A-shares [5][6]. - The market is establishing a dual mainline structure of "AI applications and commercial aerospace," with the Shanghai Composite Index rising by 0.75% and the Shenzhen Component Index by 1.31% [2][5]. Reasons for AI Application Growth - The recent strong performance of major model companies MiniMax and Zhipu, which have listed on the Hong Kong Stock Exchange, is seen as a key factor for the surge, providing a critical anchor for industry valuation and financing [7]. - Three new hotspots in the AI application field have emerged: 1. AI4S (AI for Science) with stocks like Zhizhi New Materials hitting "20CM" limit up and a year-to-date increase of over 198% [8]. 2. GEO (Generative Engine Optimization) focusing on optimizing content for AI search, marking a shift in user behavior towards direct answers from AI rather than traditional search engine results [8]. 3. AI+ healthcare, with OpenAI launching "ChatGPT Health" and significant user engagement reported by Ant Group's "Antifufu" [9]. Future Outlook for AI Applications - 2026 is anticipated to be a pivotal year for AI applications, with expectations of a "golden year" driven by technological maturity, supportive policies, and market demand [12]. - Key investment directions include: 1. Super entry points where large models become dominant traffic sources. 2. AI infrastructure focusing on software-defined computing. 3. High-growth areas in marketing and media leveraging AI capabilities. 4. High barriers in sectors like healthcare and manufacturing due to data and workflow integration [12].