Core Insights - The 2025 China Unicorn Rankings report highlights a slowdown in primary financing, a surge in secondary listings in Hong Kong, and explosive growth in three key sectors: AI, new energy, and robotics [3][5]. Financing and Market Trends - As of May 2025, the total scale of Chinese unicorns reached 8.46 trillion yuan, a slight increase of 0.23% from 2024 [3]. - The number of new unicorns is insufficient to fully replace those that have gone public, with notable companies like Mixue Ice Cream and Horizon Robotics achieving pre-IPO valuations exceeding 60 billion yuan [3]. - The trend of unicorns choosing to list in Hong Kong is significant, with 60% of unicorns opting for this route between April 2024 and April 2025, and 67% of these companies seeing an increase in market value post-IPO [4][33]. Sectoral Growth - The unicorn landscape is dominated by three core areas: AI applications, emerging technologies (commercial aerospace, biotechnology, AI+AR), and cultural exports through social media and gaming [5][14]. - In the AI sector, there are 39 unicorns, with a total valuation of 262.2 billion yuan, making it the leading area of growth [6][8]. - The intelligent driving sector is also thriving, with companies like Yihang Intelligent achieving valuations in the hundreds of billions [11]. Regional Insights - Hangzhou and Shenzhen are emerging as strongholds for new unicorns, while Beijing and Shanghai remain foundational for innovation [4][20]. - Beijing leads with 65 unicorns valued over 3 trillion yuan, accounting for 35.6% of the national total [23]. - Shenzhen has 30 unicorns with a total valuation of 926.9 billion yuan, showing a 13% increase from 2024 [29]. IPO Trends - The number of unicorns going public has decreased, with 40 companies listing between April 2024 and April 2025, totaling a market value of 104.2 billion USD, down from 54 companies and 168.2 billion USD in 2024 [33][34]. - The average market value of listed unicorns has also declined, with Hong Kong becoming the preferred market for smaller unicorns [36][38]. Cultural and Gaming Exports - The cultural export sector is gaining traction, with companies like ByteDance and Xiaohongshu leading the way in global user engagement and revenue growth [17][18]. - The gaming industry is also a significant contributor to cultural exports, with titles like "Black Myth: Wukong" achieving nearly 1 billion USD in revenue [18][19].

香港交易及結算所有限公司、香港聯合交易所有限公司與證券及期貨事務監察委員會對本申請版本的內容 概不負責，對其準確性或完整性亦不發表任何意見，並明確表示概不就因本申請版本全部或任何部分內容 而產生或因倚賴該等內容而引致的任何損失承擔任何責任。 Hai Robotics Innovation Group Co., Ltd. 深圳市海柔創新智能科技集團股份有限公司 （「本公司」） （於中華人民共和國註冊成立以不同投票權控股的股份有限公司） 的申請版本 警告 本申請版本乃根據香港聯合交易所有限公司（「聯交所」）及證券及期貨事務監察委員會（「證監會」）的要求 而刊發，僅用作提供資料予香港公眾人士。 本申請版本為草擬本，其內所載資料並不完整，亦可能會作出重大變動。 閣下閱覽本文件，即代表 閣 下知悉、接納並向本公司、本公司的聯席保薦人、整體協調人、顧問或包銷團成員表示同意： 本公司招股章程根據香港法例第32章《公司（清盤及雜項條文）條例》送呈香港公司註冊處處長登記前，本 公司不會向香港公眾人士提出要約或邀請。倘在適當時候向香港公眾人士提出要約或邀請，有意投資者務 請僅依據於香港公司註冊處處長登記的本公司招股章程作出投資決 ...

事关"人工智能+"，深圳最新发文。 今日（2月12日），据深圳工信微信公众号消息，深圳市工业和信息化局近日印发《深圳市"人工智能+"先进制造业行动计划（2026—2027年）》（以下简 称《行动计划》），从打造重点支撑平台、赋能重点产业集群、强化工作保障三个方面提出了13条具体举措，抢抓智能化与工业化交汇融合的历史机遇， 加快人工智能技术与制造业全过程、全要素深度融合。 《行动计划》明确，推动人工智能技术应用于半导体产业链的关键环节，利用AI优化芯片设计、软件代码等领域和环节的效率。以AI芯片为突破口做强 半导体产业，面向AI手机、AI眼镜、智能机器人等各类AI终端需求，研发高性能、高能效专用SoC主控芯片，支持存算一体、存内计算等新型架构处理 器。 深圳重磅发布 2月12日消息，深圳市工业和信息化局近日印发《深圳市"人工智能+"先进制造业行动计划（2026—2027年）》，其中提出，到2027年，在"人工智能+"先 进制造业领域，建成国家人工智能应用中试基地（消费领域移动终端方向），建设工业智能体创新中心，组建工业知识联盟，开放百个应用场景，打造百 个垂直行业模型及工业智能体，推广百个示范应用，形成"一基地 ...

Group 1 - The core point of the article is that Jia Mei Packaging has disclosed the capitalization path of Magic Atom Robotics for the first time, which is independent of the listed company [2][5] - Jia Mei Packaging's stock price increased by 632.24% from December 17, 2025, to February 12, 2026, closing at 33.39 yuan per share with a total market value of 36.61 billion yuan [2][9] - Magic Atom Robotics focuses on embodied intelligent robots and has completed two rounds of financing exceeding 100 million yuan, with plans to use the funds for core technology development and product deployment [5][7] Group 2 - Magic Atom Robotics is planning independent capital operations, with no intention to inject assets into Jia Mei Packaging [5][9] - The company aims to deepen strategic cooperation with Chasing Technology, leveraging supply chain synergies and brand development for market expansion [7][9] - Jia Mei Packaging's main business remains in food and beverage packaging, with no significant changes expected in the next 12 months [9][10]

事关"人工智能+"，深圳最新发文。 今日（2月12日），据深圳工信微信公众号消息，深圳市工业和信息化局近日印发《深圳市"人工智能+"先进制造业行动计划（2026—2027年）》（以下简称《行动 计划》），从打造重点支撑平台、赋能重点产业集群、强化工作保障三个方面提出了13条具体举措，抢抓智能化与工业化交汇融合的历史机遇，加快人工智能技术 与制造业全过程、全要素深度融合。 《行动计划》明确，推动人工智能技术应用于半导体产业链的关键环节，利用AI优化芯片设计、软件代码等领域和环节的效率。以AI芯片为突破口做强半导体产 业，面向AI手机、AI眼镜、智能机器人等各类AI终端需求，研发高性能、高能效专用SoC主控芯片，支持存算一体、存内计算等新型架构处理器。 深圳重磅发布 2月12日消息，深圳市工业和信息化局近日印发《深圳市"人工智能+"先进制造业行动计划（2026—2027年）》，其中提出，到2027年，在"人工智能+"先进制造业领 域，建成国家人工智能应用中试基地（消费领域移动终端方向），建设工业智能体创新中心，组建工业知识联盟，开放百个应用场景，打造百个垂直行业模型及工 业智能体，推广百个示范应用，形成"一基地、 ...

Core Insights - Kaierong High-Tech Co., Ltd. signed a strategic cooperation agreement with the Beijing Humanoid Robot Innovation Center to collaborate on the national-level general embodied intelligence platform "Wisdom Opens Things" focusing on environmental detection and intelligent manufacturing in the automotive sector [1][2] Group 1: Strategic Cooperation - The partnership aims to leverage each party's strengths to implement the "Wisdom Opens Things" platform in practical applications within environmental detection production lines and automotive parts manufacturing workshops [1] - The cooperation will enhance the intelligent upgrade of both sectors by integrating intelligent robots into industrial production practices [1] Group 2: Technological Advancements - The Beijing Humanoid Robot Innovation Center possesses core advantages in embodied intelligence technology, having developed a general embodied intelligence platform capable of "one brain, multiple functions" and "one brain, multiple machines" [1] - This platform addresses key software technology challenges for robots, enabling them to understand tasks, decompose actions, and autonomously correct errors, thus facilitating a transition from single-scene operations to multi-scene and multi-task adaptability [1] Group 3: Company Background and Vision - Kaierong High-Tech has extensive experience in the environmental detection equipment and automotive parts sectors, having entered the robotics field in 2025 with three core initiatives [2] - The company established a humanoid robot testing center in collaboration with research institutions, launched intelligent devices for precise detection of robotic hand performance, and invested in a humanoid robot manufacturing company, becoming its largest shareholder [2] - The chairman of Kaierong High-Tech emphasized the company's practical approach to transformation, focusing on real problems and seeking genuine technology solutions to integrate robots into their operations effectively [2]

国家地方共建人形机器人创新中心（以下简称"国地中心"）与北京理工大学联合研发的人形机器人"身外化 身"小脑模型—— "龙驭" ，目前已正式发布。 人形 机器人"身 外化身"通常是指，让机器人实时复刻人体全身高动态动作。 不同于传统机器人的遥操作， 仅在固定基座上复刻上肢准静态动作，"身外化身"需要在控制全身更多自由度的基础上兼顾本体的运动平 衡，甚至包含腾空起跳等高动态运动与落地平衡，对机器人运控算法提出了极高的要求。 此次推出的"龙驭"小脑模型，仅需前期少量动作数据训练，即可零样本泛化至真实场景的实时动捕、视频衍 生动作及全身遥操作等多样输入，在实体人形机器人上成功跟踪行走、奔跑、跳跃、舞蹈乃至武术等复杂技 能，并包含抗扰动，摔倒自动恢复能力，具备可用、耐用、能用的业界领先身外化身系统技术特征和优势。 "龙驭"具备三大核心特点： 1. 胃口小 ——打破数据等于智能的传统思维，利用模型设计精简与训练策略创新，仅需少量动作数据即可 达到高泛化性能力； 2. 动作精 ——一丝不苟地复现人体快速、剧烈的运动，无论是奔跑、跳跃、武术招式，还是搬运物品，都 能流畅跟随； 3. 响应快 ——遇到碰撞或干扰时可自行调整恢复 ...

Core Insights - The core focus of the report is the transition of AI development from experimental phases to scalable value creation, emphasizing the need for businesses to restructure processes and strategies for competitive differentiation [1][14][25]. Group 1: Innovation and Value Creation - The report highlights a compound effect of innovation, where advancements in technology, data, investment, and infrastructure create a self-reinforcing cycle, accelerating growth and necessitating a shift from mere automation to comprehensive business process redesign [15][25]. - Generative AI has seen rapid adoption, reaching 100 million users in just two months, showcasing the exponential growth potential of AI technologies [15][25]. Group 2: Physical AI and Robotics - Physical AI is transforming robotics, enabling machines to operate autonomously in complex environments, with applications in warehousing, manufacturing, and autonomous driving [16][30]. - By 2035, it is projected that 2 million humanoid robots will be deployed in workplaces, although challenges such as training gaps and cybersecurity risks remain [16][30]. Group 3: Digital Workforce and AI Agents - There is a significant gap in the application of digital employees (AI agents), with only 11% of companies implementing them in production due to challenges like legacy system integration and data architecture limitations [17][32]. - Leading companies are restructuring processes around AI agents, focusing on multi-agent collaboration and viewing AI as a core component of workforce management [17][32]. Group 4: AI Infrastructure Strategy - Despite a dramatic decrease in inference costs (down 280 times over two years), overall AI spending remains high due to increased usage, prompting companies to shift from a "cloud-first" strategy to a hybrid architecture combining cloud, on-premises, and edge computing [18][33]. - Companies are investing in AI-specific data centers and "AI factories" to support this hybrid approach, while also addressing challenges related to employee skill transformation and sustainable computing innovations [18][33]. Group 5: Cybersecurity and AI Risks - AI introduces a paradox in cybersecurity, where the same technologies that drive innovation also create new vulnerabilities, necessitating robust risk management across data, models, applications, and infrastructure [21][35]. - Companies are advised to enhance security measures through access controls and model isolation, while leveraging AI for automated threat detection and red team testing [21][35]. Group 6: Emerging Technology Signals - The report identifies eight key technology signals to monitor, including the potential plateau of foundational AI models, the application and risks of synthetic data, and the rise of neuromorphic computing and edge AI [3][22]. - The ability to rapidly perceive, assess, and respond to technological changes will be crucial for companies to maintain competitiveness in the AI era [3][22].

（文章来源：证券日报） 证券日报网2月12日讯 ，中联重科在接受调研者提问时表示，技术基础：以自身丰富的应用场景为突 破，结合在机械制造、大数据、人工智能方面的丰富积累，快速开发出多款人形机器人。研发主体：中 科云谷机器人研发中心为研发主体，同时依托中联重科股份有限公司总部中央研究院、传动公司、智能 公司、新能源公司开展关键零部件与共性技术的攻关。合作模式：采用核心零部件、机器人本体、小脑 与大脑关键技术自主攻关，同时与高校、AI大厂合作加快大模型相关技术的攻关。机器人领域属于前 沿科技、国家重点战略产业，目前行业整体研发及应用尚处于早期投资阶段，离规模性盈利仍需时间。 敬请广大投资者谨慎决策，注意投资风险。 ...

证券日报网2月12日讯 ，中联重科在接受调研者提问时表示，产品矩阵包括：1.9米全尺寸轮式人形机器 人、1.3米双足人形机器人、1.7m双足人形机器人、四足机器狗。产品策略：充分分析场景需求，根据 实际需求来选择合适的产品。开发逻辑：从产品与技术的可实现性逐步开发，先在自身丰富场景验证， 在逐步推广到其他场景。机器人领域属于前沿科技、国家重点战略产业，目前行业整体研发及应用尚处 于早期投资阶段，离规模性盈利仍需时间。敬请广大投资者谨慎决策，注意投资风险。 （文章来源：证券日报） ...