Core Insights - The AI and computing industry is experiencing a price increase across various sectors, driven by rising costs in IT [1][2] - Hetzner, a major European data center operator, announced significant price hikes for its services starting April 1, 2026, affecting new orders and existing subscriptions [1] - Other cloud service providers are expected to follow suit, indicating a broader trend of increasing infrastructure costs in the cloud computing market [2] Pricing Trends - Hetzner's cloud service prices in Germany and Finland are set to rise between 30% and 38%, while prices for dedicated vCPU cloud servers in the U.S. are expected to increase by around 30% [1] - This marks Hetzner's second price increase within the month, with a previous hike in early February attributed to high hardware component costs [1][2] - OVHcloud, another European cloud service provider, has also announced price increases of 5% to 10% for products between April and September 2026 [2] AI Model Usage - There has been a significant surge in the token usage of domestic AI models in China, surpassing U.S. models for the first time [3] - From September 9 to 15, Chinese models reached 41.2 trillion tokens, compared to 29.4 trillion tokens for U.S. models, with further increases noted in subsequent weeks [3] - The rapid growth in token usage reflects an exponential increase in AI inference demand, with domestic computing power expected to gain a dominant position due to cost advantages [3] Market Outlook - The domestic computing power industry is anticipated to experience a "high prosperity trend," with AI training and inference demands rapidly increasing [4] - By 2026, the computing power demand is expected to transition from "cloud training" to a dual-driven model of "training + inference," leading to a significant release of computing power gaps [4] - The industry is predicted to enter a "full-chain inflation" cycle, with growth extending from core chips to AIDC, cloud and computing services, and supporting power equipment and servers [4]

Core Viewpoint - The article discusses the rapid rise and subsequent fall of the market value of Zhipu AI, highlighting the volatility in the tech sector and the contrasting valuation logic between traditional companies like JD.com and emerging AI firms [1][12][14]. Group 1: Market Performance - On February 20, Zhipu AI's market capitalization reached 323.2 billion HKD, surpassing JD.com's 302.4 billion HKD, raising questions about market exuberance [1]. - Following a significant drop on February 23, Zhipu's market value fell to approximately 250 billion HKD, trailing JD.com by over 80 billion HKD [3]. - The stock price of Zhipu AI experienced extreme fluctuations, with a peak increase of 43% after the launch of its flagship model GLM-5, followed by a sharp decline due to operational issues [5][6]. Group 2: Product Launch and Market Reaction - Zhipu AI launched its GLM-5 model on February 12, which was designed for complex system engineering and long-range tasks, and subsequently raised prices for its services by 30% in China and over 100% internationally [4]. - The launch led to a significant increase in usage, with a reported 460% week-over-week growth in calls to the GLM-5 model, indicating strong demand [5]. - However, operational challenges arose, leading to a public apology from Zhipu AI for service delays and user limitations, which contributed to a loss of over 70 billion HKD in market value [6][9]. Group 3: Valuation Dynamics - Zhipu AI's market valuation reached a price-to-sales ratio of over 750 times, indicating a speculative pricing model based on future growth expectations rather than current profitability [10]. - In contrast, JD.com had a price-to-earnings ratio of 9.10 and a price-to-sales ratio of 0.22, reflecting a more traditional valuation based on historical performance and cash flow [12]. - The article suggests that the valuation disparity between Zhipu AI and JD.com represents a clash between "new money" focused on future potential and "old money" grounded in established financial metrics [13][14]. Group 4: Industry Context - The article notes that while Zhipu AI's valuation appears excessive when compared to JD.com, it may still be justified when considering valuations of AI companies like Anthropic, which is valued at 380 billion USD [16]. - The narrative surrounding Zhipu AI's rapid rise and fall mirrors previous market behaviors, indicating that any deviation from industry fundamentals could lead to a swift correction in valuations [15].

Core Insights - Zhipu AI's stock price surged by 42.72% to 725 HKD, reaching a market capitalization of over 323.2 billion HKD, surpassing major domestic internet companies like Kuaishou and JD.com [1] - The company launched its flagship model GLM-5, featuring 744 billion parameters and significant advancements in technology, which has garnered substantial market interest and demand [2] Company Performance - Zhipu AI's revenue grew by over 300% year-on-year in 2025, with more than 20,000 enterprise clients across various sectors, and a high customer repurchase rate exceeding 80% [2] - The company has faced supply-demand challenges due to the overwhelming demand for GLM-5, leading to service delays and a need for expansion in computing power [3] Market Position - Zhipu AI's GLM-5 model ranks first among open-source models in key programming and reasoning benchmarks, closely competing with top closed-source models [2] - The Chinese large model market is projected to grow over 220% year-on-year by 2025, with a compound annual growth rate of over 150% from 2023 to 2027, indicating a favorable environment for Zhipu AI [3] Industry Trends - The AI industry is experiencing a high growth cycle, with increasing policy support from the government for AI development, which provides a stable environment for long-term growth [3] - Major brokerage firms express bullish views on Zhipu AI, highlighting its leading technology and commercial execution capabilities in the AI sector [4]

Core Viewpoint - The article emphasizes the emergence of top-tier AI tools and the formal start of AI commercialization, marking a significant opportunity for entrepreneurs and job seekers in the industry [9]. Group 1: Company Performance - On the first trading day after the holiday, Zhiyuan's stock surged by 40%, with a total market value exceeding 320 billion HKD [1]. - Zhiyuan's stock price has increased nearly fivefold since its listing last month [2]. Group 2: Key Growth Drivers - The first logic driving the surge is "hard technology," highlighted by the release of Zhiyuan's flagship model GLM-5, which approaches the performance of top global closed-source models [4]. - The second logic is the announcement of a price increase for the GLM Coding Plan by at least 30% on the same day as the GLM-5 release, signaling a shift towards profitability in a previously price-competitive market [4]. - The third logic involves the overwhelming global demand for GLM-5, leading Zhiyuan to recruit "computing power partners," indicating real user engagement and monetization of the product [5]. Group 3: Implications for Entrepreneurs and Job Seekers - The open-sourcing of GLM-5 under the MIT license provides entrepreneurs with access to top-tier AI tools, creating a unique opportunity for startups to leverage advanced technology [7]. - The price increase in large models suggests a growing demand for AI-related positions, particularly for those skilled in deploying large models and developing applications [7]. - The current market dynamics present a time window for developing niche AI applications, especially in sectors like education, law, healthcare, and e-commerce, where user habits are still being established [8].

Group 1 - The stock price of Zhiyuan (2513.HK) has surged by 37%, reaching 696 HKD, marking a new historical high and pushing its market capitalization beyond 300 billion HKD [1] - On February 12, the company launched its next-generation flagship model GLM-5, which shows over a 20% performance improvement in programming development scenarios compared to its predecessor, closely approaching the performance level of Claude Opus 4.5 [1] - GLM-5 achieved the best performance in the open-source domain in three evaluation metrics: BrowseComp, MCP-Atlas, and τ2-Bench [1] Group 2 - Zhiyuan has increased the price of its GLM Coding Plan package by at least 30%, and the package sold out immediately upon launch [1] - The rapid sell-out of the paid package for the domestic AI programming model marks a historical first in the industry [1]

Core Insights - The article highlights the surge in global demand for GLM-5, leading to service issues such as queuing, response delays, and lag, which have affected user experience [1] - Despite multiple expansions of domestic chip clusters and the limited release of the GLM Coding Plan package, the company has not been able to fully resolve the supply-demand imbalance [1] - To enhance user experience, the company has launched a "Computing Power Partner" recruitment program aimed at optimizing GLM-5 through collaboration with chip manufacturers and computing partners [1] Group 1 - The company acknowledges the recent spike in demand for GLM-5, resulting in service disruptions [1] - The company has attempted to address the supply-demand issue through chip cluster expansions and limited product offerings, but challenges remain [1] - The "Computing Power Partner" initiative aims to collaborate with chip manufacturers and inference service providers to improve service performance [1]

Core Insights - The company has experienced a surge in demand for GLM-5, leading to service delays and user experience issues due to exceeding planned access limits [1] - Despite multiple expansions of domestic chip clusters and limited sales of the GLM Coding Plan package, the company has not been able to fully resolve the supply-demand imbalance [1] - To enhance user experience, the company has launched a "Computing Power Partner" recruitment program aimed at optimizing GLM-5 through collaboration with chip manufacturers and service providers [1] Group 1 - The global demand for GLM-5 has significantly increased, resulting in service queuing and response delays [1] - The company has expanded its chip cluster capacity multiple times but still faces supply shortages [1] - The "Computing Power Partner" program includes collaboration with chip manufacturers for core technology interface optimization and building a more efficient inference network [1]

Investment News - The Shanghai Stock Exchange is enhancing the "fixed income +" multi-asset index and ETF product system to provide diverse quality targets for wealth management and medium to long-term capital allocation, with an expected market benefit of approximately 1.113 billion yuan by 2026 [4] - Public fund issuance has exceeded 200 billion yuan this year, with 46 funds established on February 10 and 11 alone. Equity mixed funds have raised over 70 billion yuan, becoming the most prominent category, while mixed FOFs have raised over 45 billion yuan [4] - The China Securities Index Company announced the upcoming release of the CSI Hong Kong Stock Connect Robotics Theme Index and the CSI US Stock 30 Index on February 27, which includes major companies like Netflix, Goldman Sachs, Meta, Tesla, Amazon, Microsoft, Apple, and Nvidia [4] - The total margin financing and securities lending balance has increased by approximately 800 billion yuan over the past year, reaching 2.64 trillion yuan as of February 11, compared to 1.86 trillion yuan at the beginning of 2025 [5] Company Developments - BMW Group announced a global recall of several hundred thousand vehicles due to potential safety hazards in the engine start device [6] - Shuangliang Energy disclosed it received an order from SpaceX, leading to a stock price surge. However, the Shanghai Stock Exchange issued a regulatory warning due to insufficient risk disclosure regarding the order, which amounts to approximately 1.7 million euros and involves indirect supply [6] - ByteDance's Doubao video generation model Seedance 2.0 has been integrated into the Doubao App, computer, and web versions, with a significant upgrade expected on February 14. The model has gained international attention, including from Tesla's CEO Elon Musk [6] - Lenovo Group reported third-quarter revenue of 22.2 billion USD, a year-on-year increase of 18%, with AI-related revenue growing by 72%, accounting for 32% of total revenue. Net profit reached 545.5 million USD, with a gross margin of 15.1% [7] - Hua Hong Semiconductor anticipates a record high sales revenue of 659.9 million USD in Q4 2025, a year-on-year increase of 22.4% and a quarter-on-quarter increase of 3.9%. The expected sales revenue for Q1 2026 is between 650 million and 660 million USD, with a gross margin of 13% to 15% [7] - Hermès reported a 9.8% year-on-year revenue growth in Q4, exceeding market expectations, with a 14.6% organic growth in its core leather and equestrian department. The annual operating profit reached 6.57 billion euros, with an operating margin of 41% [7]

Market Overview - The three major indices collectively rose by midday, with the Shanghai Composite Index up 0.12%, the Shenzhen Component Index up 0.81%, and the ChiNext Index up 1.18%. Over 2,700 stocks in the two markets increased, with a total trading volume of 1.33 trillion yuan [1]. Sector Performance - The film and cinema sector experienced a significant decline, dropping 3.05% by midday, with stocks like Huayi Brothers, Hengdian Film, and Bona Film hitting the daily limit down. Other sectors such as grain economy, tax refund stores, community group buying, gaming, short drama interactive games, broadcasting, tourism concepts, and hotel tourism also followed suit with declines exceeding 1% [3]. - The electric grid equipment sector showed strong performance, with companies like Siyi Electric and Sifang Co. reaching new highs, and several stocks including Wangbian Electric, Shun Sodium, and Senyuan Electric hitting the daily limit up. The non-ferrous metals sector was also active, with Xianglu Tungsten gaining three boards in five days and Zhangyuan Tungsten achieving two consecutive boards [3]. - The CPO concept saw renewed strength, with Tianfu Communication rising over 10% to set a new historical high. The gas turbine sector experienced a surge, with Yingliu Co. and Changbao Co. hitting the daily limit up [3]. Price Movements - On February 12, Zhipu announced a structural adjustment to the pricing system of the GLM Coding Plan, with an overall price increase of over 30%. By 2025, China's transformer export value is expected to reach a record 64.6 billion yuan, marking a nearly 36% increase from the previous year [3]. - The price of praseodymium and neodymium oxide continues to rise, reaching 800,000 yuan per ton, with a week-on-week increase of 9.9% and a month-on-month increase of 28.8% [3].

Core Viewpoint - The competitive landscape for domestic AI models is shifting, with companies like Zhiyu taking bold pricing actions due to strong market demand and technological advancements in their products [3][4]. Group 1: Market Dynamics - Zhiyu's stock surged over 40% intraday, closing with a 28% increase, reaching a market capitalization of 170 billion HKD, driven by a price increase announcement for its GLM Coding Plan due to rising user demand and usage [3]. - The company reported that its GLM Coding Plan was sold out immediately upon launch, indicating strong market interest in its AI programming model [3]. - In contrast, a price war is expected among major model vendors from late 2024 to early 2025, with significant price reductions already observed in the market [3]. Group 2: Technological Advancements - The price increase is supported by the launch of GLM-5, a new flagship model that has gained attention for its capabilities in complex system engineering and programming tasks, comparable to leading closed-source models [4]. - User feedback indicates that GLM-5's programming capabilities are on par with top models, although some performance issues were noted due to computational constraints [4]. Group 3: Strategic Focus - The industry is transitioning from a focus on customer acquisition to leveraging product quality and service to generate commercial value, particularly in the programming sector [4]. - Zhiyu's founder emphasized the importance of strategic focus on the coding domain, which is seen as a critical battleground for AI model companies [4]. Group 4: Future Outlook - By 2026, the application of large models in enterprises and the widespread adoption of AI programming tools are expected to make programming capabilities a key competitive factor for model vendors [5]. - The pricing strategies of large models, particularly those from leading companies like OpenAI, reflect their superior performance and the market's willingness to pay for high-quality outputs [5]. Group 5: Pricing Trends - Analysts suggest that Zhiyu's price increase may influence other domestic model providers to follow suit, as the intrinsic value of foundational models becomes more apparent in response to actual demand [6]. - The decision for other models to increase prices will depend on their competitive capabilities and the willingness of customers to invest in them [6].