● 本报记者 杨梓岩 7月11日晚间，星图测控披露投资者关系活动记录，强调太空管理业务重要性。随着Starlink等国内外星 座计划加速部署，卫星数量激增，碰撞风险上升，威胁太空系统稳定运行。公司将通过太空态势感知、 交通管理及在轨服务等板块，规划未来太空管理业务发展。公司控股股东中科星图未来将发展算力星 座，深化太空管理服务，助力全球用户应对商业航天安全挑战。星图测控董事长胡煜表示，随着航天技 术不断成熟，2025年全球航天产业将进一步多元化，向更高水平迈进。 太空管理业务重要性凸显 针对投资者就未来太空管理业务前景的提问，星图测控表示，当前，Starlink已部署超过7000颗卫星， 未来随着我国GW星座计划、千帆星座计划及其他国内外大型星座的逐步建设，可以预见，卫星间以及 卫星与太空碎片碰撞的风险将急剧上升。这不仅对卫星自身安全构成严重威胁，还可能影响整个太空系 统的稳定运行。 日前，中科星图总裁邵宗有在接受记者采访时也透露，未来五年，中科星图在商业航天方面最为核心的 规划是要构建大规模的算力星座，攻关"天地一朵云"解决方案，加强"天数天算"新服务模式研究，形成 天地一体化协调能力。 中科星图还将重点发 ...

Core Insights - xAI, led by Elon Musk, is reportedly negotiating a new funding round with a potential valuation of up to $200 billion, following a recent $10 billion capital raise split between debt and equity [1][5] - Musk has publicly denied these funding rumors, asserting that xAI has sufficient capital [1] - The company recently launched Grok 4, touted as the "world's strongest AI model," which includes significant updates and new features [1][2] Funding and Valuation - xAI's recent funding efforts include raising $10 billion, with plans to raise an additional $10 billion [1] - The company's valuation during the B round in May 2024 was only $18 billion, indicating a substantial increase in interest and valuation in a short period [1] - The upcoming funding round could mark xAI's third major financing event within two months [1] Product Development - Grok 4 is the fourth major iteration of xAI's AI models, featuring two versions: Grok 4 and Grok 4 Heavy, with the latter supporting multiple agents and a context window of up to 256k tokens [1][2] - Internal benchmarks show Grok 4 achieving approximately 25% accuracy on a comprehensive exam, comparable to OpenAI's tools [2] - The model introduces new voice modes and a programming model aimed at enhancing developer efficiency, addressing competition from GPT-4 and Meta Gemini [2][3] Controversies and Challenges - Grok 4's launch has been met with criticism due to past issues with the previous model, Grok 3, which faced backlash for inappropriate content [3] - Concerns have been raised regarding the need for more structured safety mechanisms and transparency in training methods for Grok 4 [3] - Musk envisions Grok 4 evolving into a truth-seeking AI, potentially integrating with humanoid robots to discover new technologies [3] Company Background and Market Position - xAI was founded in July 2023, with a team comprising members from notable organizations like Google DeepMind and OpenAI [3] - The company has released several models since its inception, continuously improving capabilities in reasoning, context length, and image understanding [4] - Despite political controversies surrounding Musk, investor interest in xAI and related ventures remains strong, as evidenced by SpaceX's plans to sell $1 billion in shares, potentially valuing the company at $400 billion [5]

Core Viewpoint - End-to-End Autonomous Driving is a key algorithm for intelligent driving mass production, with significant salary potential for related positions, and it has evolved into various technical branches since the introduction of UniAD [2] Group 1: Overview of End-to-End Autonomous Driving - End-to-End Autonomous Driving can be categorized into one-stage and two-stage approaches, with the core advantage being direct modeling from sensor input to vehicle planning/control, avoiding error accumulation seen in modular methods [2] - The emergence of BEV perception has bridged gaps between modular methods, leading to a significant technological leap [2] - The academic and industrial focus on End-to-End technology has raised questions about whether UniAD is the ultimate solution, indicating ongoing developments in various algorithms [2] Group 2: Challenges in Learning - The rapid development of End-to-End technology has made previous solutions inadequate, necessitating knowledge in multimodal large models, BEV perception, reinforcement learning, visual transformers, and diffusion models [4] - Beginners often struggle with the fragmented nature of knowledge and the overwhelming number of papers, leading to challenges in extracting frameworks and understanding industry trends [4] Group 3: Course Features - The newly developed course on End-to-End and VLA Autonomous Driving aims to address learning challenges by providing a structured approach to mastering core technologies [5] - The course emphasizes Just-in-Time Learning, helping students quickly grasp key concepts and expand their knowledge in specific areas [5] - It aims to build a framework for research capabilities, enabling students to categorize papers and extract innovative points [6] Group 4: Course Outline - The course includes chapters on the introduction to End-to-End algorithms, background knowledge, two-stage End-to-End methods, one-stage End-to-End methods, and practical applications [11][12][13] - Key topics include the evolution of End-to-End methods, the significance of BEV perception, and the latest advancements in VLA [9][14] Group 5: Target Audience and Expected Outcomes - The course is designed for individuals aiming to enter the autonomous driving industry, providing a comprehensive understanding of End-to-End technologies [19] - Upon completion, participants are expected to achieve a level equivalent to one year of experience as an End-to-End Autonomous Driving algorithm engineer, mastering various methodologies and key technologies [22]

第一，关注政策受益的三大方向。一是金融科技，险资入市和AI+金融的结合，可能会催生新的业务模式；二是高端制造，人形机器人、算力设备这些细分 领域，已经出现订单兑现的迹象；三是消费升级，食品饮料里的高质价比商品和创新药，正在展现结构性亮点。 第二，把握低吸机会。对于那些前期涨幅过大的板块，比如稀土永磁，短期可能会有回调压力，可以等回踩5日均线再考虑介入。而像电力、医药这些调整 充分的板块，反而可能成为资金轮动的下一站。 外围市场方面，美股三大指数周五全线收涨，英伟达更是突破4万亿美元市值，这说明全球资金对科技主线的认可度依然很高。不过大家要注意，美联储7月 议息会议虽然要到月底才开，但市场已经开始博弈降息预期。目前看，降息概率只有两成左右，所以下周外围市场大概率维持震荡格局，对A股的直接冲击 有限。倒是北向资金的动向值得关注，最近他们持续加仓金融和稀土板块，这种"聪明钱"的选择，往往能给我们提供重要参考。 技术面上，上证指数目前处于箱体突破后的回踩阶段。从日K线看，3500点附近的支撑力度较强，5日均线和10日均线形成金叉，MACD指标虽然出现顶背 离，但成交量依然维持在万亿级别。这说明市场整体趋势还是向上的， ...

Core Viewpoint - The company, Koyuan Smart, is expected to maintain strong performance and long-term growth potential driven by its multi-national certifications, comprehensive technology R&D system, platform-based products, and decentralized channels [1]. Investment Highlights - The investment recommendation is to maintain a "Buy" rating, with the target price raised to 31.66 yuan (+0.98). The projected EPS for 2025-2027 is 1.27 (-0.21), 1.73 (-0.22), and 2.06 yuan respectively. The average PE for comparable companies in 2025 is estimated at 26.79X. Given the company's focus on control systems and continuous expansion in the industrial internet product matrix, it is expected to sustain high growth, justifying a 25x PE for 2025, leading to the target price of 31.66 yuan [2]. - In 2024, Koyuan Smart achieved total revenue of 1.682 billion yuan, a year-on-year increase of 19.55%, with a net profit attributable to shareholders of 252 million yuan, marking a significant growth of 56.64% compared to 2023. In Q1 2025, the company reported revenue of 420 million yuan, a year-on-year increase of 14.60%, and a net profit of 61 million yuan, reflecting a 42.30% increase, showcasing its technological breakthroughs and ability to leverage policy benefits for enhanced profitability [2]. Technological and Market Positioning - The company has established a technological ecological barrier through multiple national certifications and a comprehensive technology R&D system. As a national-level manufacturing champion and a key high-tech enterprise, it holds certifications such as ISO 9001/45001/14001 and CMMI Level 5, demonstrating its credibility in the industrial intelligence sector. The company integrates cutting-edge technologies like AI, large models, and real-time databases, with R&D investment accounting for 10.44% in 2024, adding 16 new patents and holding a total of 236 effective patents [3]. - Koyuan Smart focuses on platform-based products and decentralized channels to efficiently convert technological advantages into market penetration. The company has built an IT/OT integrated platform that deeply merges industrial automation systems with industrial software, providing full lifecycle services from planning to operation for traditional industries. Its marketing network, covering 17 divisions nationwide, enhances customer loyalty and market share, accelerating the large-scale implementation of smart manufacturing solutions [3].

7月12日，2025中国汽车论坛在上海召开。会上，懂车帝内容与产品副总裁彭艳鋆表示，懂车帝已正式 上线AI选车。 （原标题：懂车帝首次公开AI全景图 多元场景AI智能体已实现应用） 与此同时，AI大模型呈现出新趋势，垂类大模型持续增长。彭艳鋆说，从汽车垂类领域来看，通用大 模型对行业的理解深度有待提升，比如对用户汽车需求理解不够精准、数据不够权威、只有模糊信息参 考而没有决策建议等。 "懂车帝活跃用户超过七成是高意向选买用户，因此率先推出AI选车，解决用户最高频刚需的选买问 题。"彭艳鋆介绍，懂车帝AI选车主要满足模糊选车、多车对比、信息查询、交易和服务等场景需求， 旨在成为"用户口袋里的汽车选买专家"。 据悉，懂车帝于2023年推进AI大模型应用研发，2024年上线AI选车助理功能，2024年12月懂车帝大模 型算法通过备案，2025年7月正式上线AI选车产品。 彭艳鋆表示，选车只是懂车帝AI智能体目前最常用的场景之一，随着AI应用深入，传统的流量中心和 消费体验将被彻底颠覆，汽车营销交易等全链条也将受益。"传统的汽车行业生态将被深度重塑，期待 与行业一起，通过AI为消费者提供更好的服务，推动汽车产业向新向 ...

Investment Rating - The report maintains a "Recommendation" rating for the communication industry, expecting the industry index to outperform the benchmark index by more than 5% in the next 3-6 months [27]. Core Insights - The report highlights that Broadcom's management has indicated a significant surge in AI inference demand, which is expected to drive a systematic reassessment of profit structures and industry capacity [13][15]. - The AI inference demand has exceeded current capacity, suggesting potential upward revisions in profitability forecasts for Broadcom, with a projected market space of $60-90 billion for three major AI clients by 2027 [13][15]. - The report emphasizes the importance of high bandwidth and low latency networks for AI inference workloads, with a current spending ratio of approximately 3:1 between computing and networking devices in AI systems [14]. Market Performance - The communication sector (Shenwan) rose by 2.13% in the week, outperforming the CSI 300 index by 1.31 percentage points and underperforming the ChiNext index by 0.23 percentage points [6][7]. - Year-to-date, the communication sector has increased by 9.54%, surpassing the CSI 300 index by 7.51 percentage points and the ChiNext index by 6.48 percentage points [6][7]. Stock Performance - The top five gainers in the communication sector for the week were: SanChuan Wisdom (+27.91%), GuoYuan Technology (+27.72%), BoChuang Technology (+19.84%), ShiJia Photon (+14.85%), and ZhongXin SaiKe (+12.46%) [10][12]. - The top five losers were: YouFang Technology (-18.52%), RuiSiKangDa (-13.48%), XuanJi Information (-7.31%), HeZhong ShiZhuang (-5.09%), and RunZe Technology (-4.10%) [10][12]. Investment Recommendations - Key recommendations include focusing on major operators such as China Mobile, China Telecom, and China Unicom [21]. - For optical modules and devices, the report recommends companies like XinYiSheng, TianFu Communication, and ZhongJi XuChuang, while suggesting to pay attention to YuanJie Technology and ShiJia Photon [21]. - In the military/satellite communication sector, recommended companies include HaiGe Communication, Shanghai HanXun, and QiYiEr [21].

Core Insights - The domestic computing power industry chain has achieved a breakthrough in the logic chain of "industrial breakthrough - performance verification - demand verification," indicating that the industry is entering a catalytic phase [2][4] - Key domestic GPU manufacturers, such as Muxi Integration and Moer Thread, have had their IPO applications accepted, filling the gap in the A-share market for fully functional GPUs [5][6] - Huawei's computing power cluster performance is now comparable to NVIDIA, with significant advancements in AI model capabilities and infrastructure [6][9] Industry Overview - The performance of key companies in the industry chain, such as Industrial Fulian, has shown remarkable growth, with a forecasted net profit of 6.727 to 6.927 billion yuan for Q2 2025, representing a year-on-year increase of 47.72% to 52.11% [7] - The demand side is witnessing rapid growth in the consumption of tokens by domestic large models, indicating a significant increase in computing power demand [8] - The acceleration of the domestic GPU listing wave, combined with the performance of hardware manufacturers and the growing demand for AI large models, suggests that the domestic computing power industry is entering a high-growth cycle [4][8] Key Company Developments - Muxi Integration has sold over 25,000 GPU products and has achieved large-scale applications in various AI public computing platforms [5] - Moer Thread has successfully developed and mass-produced four generations of high-performance GPU architectures, launching multiple high-performance GPU accelerator cards and integrated solutions [5] - Huawei's CloudMatrix 384 AI cloud service has achieved a significant performance leap, surpassing NVIDIA's flagship products in several key metrics, indicating a milestone breakthrough in China's AI infrastructure [6][9] Market Dynamics - The rapid increase in token consumption for domestic large models, with daily usage reaching 16.4 trillion tokens, reflects a strong upward trend in computing power demand [8] - The performance of Industrial Fulian's AI-related business has seen over 60% revenue growth in AI servers, indicating a sustained positive outlook for the computing power industry chain [7] - The overall trend suggests that the domestic computing power market is likely to experience further catalysis, with recommendations to focus on domestic computing power chips, servers, PCB, and optical communication manufacturers [2][4][8]

Core Viewpoint - The company is recruiting for various positions related to embodied intelligence, focusing on multimodal large models, robotic multimodal interaction, and reinforcement learning, indicating a strong emphasis on innovation and application in the robotics field [1][3][5]. Group 1: Job Descriptions - **Embodied Multimodal Large Model Researcher**: Responsible for developing core algorithms for embodied intelligence, including multimodal perception, reinforcement learning optimization, and world model construction [1]. - **Robotic Multimodal Interaction Algorithm Researcher**: Focuses on researching multimodal agents, reasoning planning, and audio-visual dialogue models to innovate and apply robotic interaction technologies [3]. - **Reinforcement Learning Researcher**: Engages in exploring multimodal large models and their applications in embodied intelligence, contributing to the development of next-generation intelligent robots [5]. Group 2: Job Requirements - **Embodied Multimodal Large Model Researcher**: Requires a PhD or equivalent experience in relevant fields, with strong familiarity in robotics, reinforcement learning, and multimodal fusion [2]. - **Robotic Multimodal Interaction Algorithm Researcher**: Candidates should have a master's degree or higher, excellent coding skills, and a solid foundation in algorithms and data structures [4]. - **Reinforcement Learning Researcher**: Candidates should have a background in computer science or related fields, with a strong foundation in machine learning and reinforcement learning [6]. Group 3: Additional Qualifications - Candidates with strong hands-on coding abilities and awards in competitive programming (e.g., ACM, ICPC) are preferred [9]. - A keen interest in robotics and participation in robotics competitions are considered advantageous [9].

2025.07. 13 和每年进行的GDP核算不同，经济普查每五年进行一次。去年12月26日，第五次全国经济普查结果 发布，普查全面摸清了我国第二产业和第三产业家底，反映出我国经济社会发展取得的新成效。其 中，2023年末，全国共有信息传输、软件和信息技术服务业企业法人单位169.1万个，从业人员 1506.7万人，分别比2018年末增长85.2%和51.4%。 信息传输、软件和信息技术服务业包括：电信、广播电视和卫星传输服务；互联网和相关服务；软件 和信息技术服务。其中，2023年，软件和信息技术服务业的法人单位从业人员达到了1107.1万人， 占信息传输、软件和信息技术服务业的比重达73.48%。 那么，这个高薪行业，哪些城市的从业人员较多呢？ 今年以来，各地陆续发布了第五次经济普查相关数据。第一财经记者梳理发现，北京、上海、深圳、 广州、杭州的信息传输、软件和信息技术服务业法人单位从业人员超过60万人，位居前五，其中京 沪超过了100万人（需要说明的是，部分重点城市的五经普数据尚未发布，在此没有纳入）。 从具体城市来看，北京市五经普结果显示，北京在信息传输、软件和信息技术服务业的法人单位的相 关从业人员 ...