在这波AI产业浪潮中，单纯的模型研发已不再是唯一焦点，"算法+硬件+场景"的全链条融合正成为商 业化落地的核心关键。作为全球LED显示产业的龙头企业，洲明科技（300232.SZ）凭借前瞻的战略布 局，早已与智谱AI形成"资本+产业"的深度绑定，成为智谱AI生态中"显示具身化"的重要落地者。然 而，从二级市场表现来看，公司当不到百亿的市值尚未充分反映其在AI领域的先发优势与高成长潜 力。随着港股AI板块的持续升温，洲明科技的投资价值正迎来基本面、成长性与估值逻辑的三重共 振，一场全面的价值重估已然箭在弦上。 2月24日上午的市场走势显示，港股市场人工智能板块整体延续量价齐升的火爆态势，智谱 （02513.HK）继续大涨20.89%，MINIMAX-WP(00100.HK)涨7.85%。这波行情的背后，是国产大模型 商业化落地进程的全面提速。OpenRouter最新发布的周度数据显示，全球平台前十模型总token量约达 8.7万亿，其中中国模型独占5.3万亿，占比高达61%，国产大模型在全球AI竞赛中已确立核心话语权。 具体来看，MiniMax M2.5以2.45万亿token空降榜首，Kimi K2.5以1. ...

记者丨杨坪 编辑丨李新江 连续三年财务造假暴露后，*ST 立方的退市结局终于落锤！ 2月14日晚间，安徽证监局公布的行政处罚决定书显示，立方数科股份有限公司（以下简称"*ST 立方"）通过开展代理业务、融资性贸易、虚 假贸易大幅虚增营业收入，2021年至2023年连续三年年度报告存在虚假记载。 安徽证监局决定对公司责令改正，给予警告，并处以1000万元 罚款。 同日，公司收到深交所下发的事先告知书， 深交所拟决定终止公司股票上市交易。公司股票将于2月24日起停牌。 连续三年财务造假遭罚 根据安徽证监局出具的《行政处罚决定书》显示，*ST立方通过三种方式实施财务造假。 第一，通过开展代理业务虚增营业收入、营业成本。 2021年至2023年，*ST立方及其子公司与北京某创世纪科技有限公司等12家公司开展代理 业务并采用总额法核算。在前述业务中，*ST立方及其子公司向下游转让合同标的前不拥有对相关商品的控制权。其不参与货物运输，不承担 向下游客户转让合同标的的主要责任，不承担合同标的存货风险，无权自主决定所交易商品的价格。*ST立方采用总额法对该部分业务进行核 算，不符合相关会计准则。 第二，*ST立方通过开展融资 ...

Core Viewpoint - The domestic AI large model industry is entering a period of accelerated commercialization, with Shenzhen Zhuhua Technology Co., Ltd. (referred to as "Zhuhua") and its strategic partnerships positioning it for significant growth opportunities in the AI sector [1][4]. Group 1: Company Overview - Zhuhua officially listed on the Hong Kong Stock Exchange on January 8 [1]. - Zhuhua has initiated strategic collaborations with leading large model enterprises, including a joint investment with Zhuhua and Beijing Yuanke Shijie Technology Co., Ltd. to establish Shenzhen Zhixian Robot Technology Co., Ltd. [1][2]. - The joint venture focuses on developing vertical industry-specific models, AI industry solutions, and intelligent interaction functions, providing an integrated solution that combines algorithm models, hardware terminals, and perception interaction [1]. Group 2: Strategic Collaborations - The collaboration between Zhuhua and leading large model developers like Zhuhua is not coincidental, as Zhuhua's GLM series models are internationally recognized for their capabilities in programming and logical reasoning [2]. - Zhuhua's AIAgent platform has successfully integrated with major models such as Microsoft Azure GPT-5 and Google Cloud Vertex Gemini 2.5, expanding the boundaries of technological cooperation [2]. Group 3: Product Development and Market Penetration - Zhuhua's AI intelligent products have achieved mass delivery and scene reuse, with successful implementations in notable locations such as Nanjing Deji Plaza and the National Museum [3]. - The company has signed a framework agreement for the delivery of 1,500 AI all-in-one machines with Hong Kong Science Group and a five-year supply agreement for 1 million AI smart hardware units with Guangdong Huahui Huangguang Technology Co., Ltd. [3]. - Zhuhua's AI products, which incorporate intelligent features like content generation and scene recognition, significantly enhance traditional LED displays, resulting in higher profit margins compared to conventional display hardware [3]. Group 4: Future Outlook - Industry experts believe that Zhuhua's strategic partnerships with leading enterprises solidify its technological foundation, opening up growth opportunities through a multi-scenario AI intelligent product matrix [4]. - As the "LED+AI" strategy continues to deepen and intelligent applications expand into more vertical fields, the company's performance is expected to achieve a qualitative leap, leading the smart transformation of traditional manufacturing [4].

Group 1 - The core point of the news is that Jiechuang Intelligent plans to purchase IT equipment and components worth up to RMB 4 billion to advance its artificial intelligence strategy and business transformation, which represents 50% of the company's latest audited net and total assets, pending shareholder approval [1] - The purchase is aimed at expanding the company's cloud computing business and enhancing technology product development, with no implications for personnel placement, land leasing, or debt restructuring [1] - Jiechuang Intelligent emphasizes that the investment in AI is increasing as the company progresses with its AI strategy, which is crucial for integrating into the cloud computing industry ecosystem [1] Group 2 - On the same day, Jiechuang Intelligent announced a capital increase of RMB 170 million to its subsidiary, Guangdong Jiechuang, to enhance its capital strength and promote operational development, raising the subsidiary's registered capital to RMB 295 million [2] - The company is actively seizing opportunities in the AI computing power demand from internet, autonomous driving, and AI startups, providing flexible rental smart computing cloud services [2] - In the first half of 2025, the "AI + cloud computing" business generated over RMB 30 million in revenue, significantly surpassing the total for 2024, with expectations for continued rapid growth in 2025-2026 [2] Group 3 - Jiechuang Intelligent has been deeply involved in the cloud computing sector for years, establishing the independent brand "Changqing Cloud," which is currently in a rapid expansion phase after its incubation stage [3] - Changqing Cloud serves various vertical industries, including government, finance, and education, with notable clients such as the People's Bank of China and Guangzhou Port Group [3] - The products of Changqing Cloud are compatible with domestic operating systems and CPUs, and the company plans to enhance product competitiveness by closely monitoring developments in domestic technology [3]

"AI基础设施前期投入巨大，一套AI一体机的投资规模往往就达数百万元，甚至高达上千万元。"一位调 查机构人士向记者表示，从年度调研反馈及日常服务企业的情况来看，不少企业在AI布局过程中普遍 面临较大挑战。 尽管人工智能与传统产业的深度融合，正在推动生产流程向智能化、自动化转型，并在一定程度上提升 了生产效率和产品质量、降低了单位生产成本，但投资回报率仍是企业最为担忧的问题之一。 上述机构的调查数据显示，40%的中国受访企业将财务成本高、投资回报率偏低列为科技应用面临的首 要挑战；在中小企业中，这一比例更高，达到49%。 记者了解到，中型企业在推进定制化模型开发、云算力结合轻量化硬件应用时，初始投入一般在50万至 500万元之间，年运营成本约为30万至200万元，主要应用于AI质检、供应链预测等场景。大型企业若 开展大模型训练、搭建本地算力集群，初始硬件及研发投入通常需要5000万元至数亿元，年运维成本也 往往从千万元级起步。 从成本构成来看，数据工程、模型开发与训练、算力硬件、人才投入以及合规与运维构成五大核心板 块。其中，数据工程成本占比最高，约为30%—50%，涵盖数据采集、清洗、标注等全流程。 上述人士 ...

Group 1 - The company, Zhouming Technology, has signed a strategic cooperation agreement with TechTure to initiate an AI intelligent upgrade project for rail transit [2] - A framework agreement for the bulk delivery of 1,500 AI integrated machines has been established [2] - The delivery of the first batch of products for the project has already commenced, with production, quality inspection, and logistics processes progressing as planned [2]

Core Viewpoint - The company has initiated a strategic partnership with TechTure to enhance AI integration in rail transit systems, marking a significant step in its product development and market expansion [2] Group 1: Strategic Developments - The company signed a strategic cooperation agreement with TechTure on December 2, focusing on the AI smart upgrade project for rail transit [2] - A framework agreement for the bulk delivery of 1,500 AI integrated machines has been established, indicating a strong commitment to scaling production [2] Group 2: Production and Delivery - The delivery of the first batch of products for the project has commenced, with production, quality inspection, and logistics processes progressing as planned [2]

Group 1: LED+AI Product Strategy - The LED+AI strategy has two main paths: AI industry solutions focusing on vertical classification models and terminal AI intelligence, and AI edge products creating an ecosystem driven by hardware, IP, and scene services [1][2] - The company has invested 50% (CNY 25 million) in a joint venture with Beijing Zhipu Huazhang Technology Co., Ltd. and Beijing Yuanke Shijie Technology Co., Ltd. to establish a subsidiary for AI solutions [2] Group 2: AI Edge Product Orders - The company has signed a framework agreement for the delivery of 1,500 AI all-in-one machines with TechTure [3] - A partnership with Guangdong Huahui Huang Optoelectronics Technology Co., Ltd. will see the procurement of 1 million AI smart hardware units [3] Group 3: Nanchang Base Capacity Planning - The Nanchang base project has a total investment of CNY 600 million, focusing on Micro and Mini LED display and lighting production [4][5] - The new base has begun limited production, with plans for full-scale operations to enhance production efficiency and reduce costs [5] Group 4: Participation in National Games Facilities - The company has provided LED display products and integrated solutions for multiple venues in the National Games across Guangdong, Hong Kong, and Macau [6] - Key projects include installations at major sports centers in Guangzhou and Shenzhen, enhancing event presentation and lighting [6] Group 5: Future Profit Improvement Strategies - The company aims to focus on mid-to-high-end products to differentiate from low-cost competitors and enhance customer value perception [7][8] - Plans include exploring high-value application scenarios in smart museums and digital tourism, leveraging LED and AI technology [8] Group 6: M&A Plans and Financing Needs - The company maintains a cautious approach to mergers and acquisitions, focusing on integrating market, brand, and technology resources [9] - Key areas of interest include upstream and downstream integration in the optoelectronics industry and collaboration with government and financial institutions for strategic investments [9]

Core Insights - The company has solidified its leading position in the display industry, showing steady improvement in profitability with a revenue of 154.55 billion yuan for the first three quarters of 2025, a year-on-year increase of 7.5%, and a net profit of 4.6 billion yuan, up 39% [1] - The company is actively expanding into emerging markets, achieving significant results across four business segments, including smart terminals, MLED, sensing, and digital healthcare [2] - The company has embraced AI-driven transformation with its "1+4+N" strategy, enhancing operational efficiency and expanding its value chain to discover new growth points [3] Group 1: Financial Performance - In Q3 2025, the company achieved a revenue of 53.27 billion yuan, a year-on-year increase of 5.8%, and a net profit of 1.35 billion yuan, up 32.1% [1] - The gross margin and net margin for the first three quarters of 2025 were 14.4% and 2.9%, respectively, with the net margin increasing by 1.6 percentage points [1] Group 2: Business Expansion - The smart terminal business has successfully launched its first full-process product at the BOE Vietnam vehicle display base [2] - The MLED business saw a revenue increase of over 15% year-on-year in Q3 2025, with a nationwide dealer network covering all provincial administrative regions in China and expansion into 32 countries [2] - The sensing business reported a 57% year-on-year revenue growth in Q3 2025, with advancements in glass-based advanced packaging projects [2] Group 3: Technological Innovation - The company has launched AI factories, enhancing production planning, material supply, manufacturing, quality management, energy optimization, and environmental safety [3] - The company is advancing semiconductor display technology, IoT innovation, and sensor technology, establishing a comprehensive "screen IoT" technology architecture [3] - The profit forecast for the company has been adjusted, with expected net profits of 6.02 billion yuan, 8.64 billion yuan, and 10.56 billion yuan for 2025, 2026, and 2027, respectively [3]

Core Insights - The AI industry is shifting from a focus on computational power to application implementation, as evidenced by Hengwei Technology's acquisition of 75% of Shuhang Information, marking the first AIRaaS acquisition in the A-share market [1][4] - Hengwei Technology's previous focus was on hardware solutions, but the need for quantifiable business outcomes has led to a strategic pivot towards service-oriented models [1][3] Company Overview - Shuhang Information, established in 2017, specializes in binding large model technology with industry scenarios, offering a full-process solution from data preparation to deployment through its proprietary S-GPTAI engine and Langtree model orchestration platform [2] - The company has experienced rapid revenue and profit growth over the past three years, serving various industries including fast-moving consumer goods, automotive, and finance [2] AIRaaS Model - The AIRaaS model aims to address the growth challenges faced by traditional AI service companies, which often struggle with implementation and measurable results, leading to client hesitance [3] - Unlike traditional models that charge for software licenses or human resources, AIRaaS focuses on quantifiable business outcomes, charging clients only upon achieving agreed results, thus creating a "results-based" service model [3] - This model allows for access to a core profit pool that constitutes 20%-60% of enterprise revenue, significantly higher than the 1%-2% covered by traditional digital services, presenting a substantial value opportunity for companies like Shuhang Information and hardware manufacturers like Hengwei Technology [3] Industry Trends - The acquisition reflects a broader trend in the AI industry, where competition is expected to shift from computational power to the ability to implement solutions in specific business contexts by 2025 [4] - As AI technology becomes more accessible, the focus will increasingly be on understanding industry needs and adapting models to meet those demands, moving away from a purely parameter-driven approach [4] - The ultimate impact of this acquisition will require time to assess, particularly regarding the integration of technologies and resources, as well as the ability to create differentiated advantages in the market [4]