Core Viewpoint - BYD's intelligent driving business has entered a relatively stable phase after over two years of adjustments, with significant organizational changes reflecting its strategic focus on smart driving technology [2][4]. Group 1: Organizational Changes - Recent personnel changes in BYD's intelligent driving division indicate a shift in strategy, with key figures such as Liao Jie and Tang Minxin leaving the company [4][6]. - The restructuring has led to the integration of two major self-research teams, "Tianxuan" and "Tianlang," into a single team to streamline operations and enhance resource allocation [6][7]. Group 2: Strategic Focus - BYD has elevated intelligent driving to one of its core strategies, with Chairman Wang Chuanfu emphasizing its importance in internal meetings [6]. - The company aims to replace supplier solutions with self-developed technologies to reduce overall vehicle costs, as demonstrated by the successful transition to self-researched solutions in the "Tianshen Zhi Yan C" platform [7]. Group 3: Market Position and Goals - As the world's highest-selling new energy vehicle manufacturer, BYD has set an ambitious sales target of 5.5 million vehicles for the year, balancing the desire for technological leadership with cost-effectiveness [7].

Core Viewpoint - After more than two years of adjustments, BYD's intelligent business structure has entered a relatively stable phase, reflecting the company's strategic focus on intelligent driving and related technologies [2][3]. Group 1: Organizational Changes - Recent personnel changes at BYD's intelligent business unit highlight the company's strategic decisions and investments in intelligent driving, with key figures like Liao Jie and Tang Minxin leaving their positions [3][5]. - Liao Jie, who joined BYD in August 2023, was involved in the self-research of intelligent driving algorithms and has now returned to Horizon Robotics [3][5]. - The restructuring has led to a clearer division of responsibilities within the intelligent business unit, with a focus on integrating intelligent driving and cockpit functions under single management [6]. Group 2: Strategic Focus on Intelligent Driving - BYD has elevated intelligent driving to one of its core strategies, with significant investments and rapid recruitment aimed at catching up in the market [5][6]. - The company has formed two major self-research teams, "Tianxuan" and "Tianlang," which were later consolidated into a single team to streamline operations [5][6]. - The integration of intelligent driving and cockpit management is seen as a way to reduce costs and improve efficiency, aligning with the trend of combining these functionalities into a single computing chip [6]. Group 3: Technological Development and Market Position - BYD is focused on the self-research and mass production of the "Tianshen Zhi Yan B" platform, which aims to implement advanced driving features at a lower cost by replacing supplier solutions with in-house developments [6]. - The company has successfully validated this strategy with the "Tianshen Zhi Yan C" platform, demonstrating its capability to integrate advanced features into more affordable vehicle models [6]. - Unlike other emerging car manufacturers, BYD, as the world's top-selling new energy vehicle company, balances its desire for technological leadership with cost and return considerations, setting a sales target of 5.5 million vehicles for the year [6].

Summary of Conference Call on Intelligent Driving Chips Industry Overview - The domestic mid-tier intelligent driving chip market is primarily dominated by Horizon Robotics' Journey 6 series, which strikes a balance between performance and price, making it widely adopted by mainstream OEMs. Qualcomm and NVIDIA solutions are used less frequently [1][2] - The high-end market is mainly led by NVIDIA's Orin series chips, but Horizon has introduced a full-stack high-end solution, HSD, based on G6P, gaining some market share due to cost advantages and superior performance [1][2] - Qualcomm's 8,775 chip is positioned for the mid-to-low-end market, expected to be deployed in the second half of 2025, but OEMs show limited willingness to adopt it due to the challenges of simultaneous development of cockpit and intelligent driving systems [1][5] Key Insights - The domestic vehicle chip design sector has largely achieved independence, but the production side still relies on foreign foundries, limiting the localization rate of mid-tier and high-end chips. The future trend is to enhance design independence and seek more production collaborations [1][6] - NIO is making the fastest progress in self-developed chips, having implemented its self-developed ZAM chip in the ET9 model, improving computing power utilization and reducing costs, although it faces challenges in capacity and R&D costs [1][7] Market Dynamics - The domestic vehicle chip market can be segmented into low, mid, and high tiers. The low tier includes basic L2 functions, while the mid-tier has expanded to high-speed NOA, which has become mainstream following BYD's intelligent driving strategy [2] - The pricing trend for mainstream vehicle chips is generally declining, with low-tier chips around $20, mid-tier chips like Horizon's 6M at approximately $70, and high-end chips like NVIDIA's Orin X exceeding $300 [3][17] Competitive Landscape - The low-tier market is primarily led by Horizon, Mobileye, and Bosch, with Horizon's share expected to increase. The mid-tier market is dominated by the Journey 6 series, which is projected to capture a significant portion of the domestic OEM market in the next two years [18] - In the high-end market, NVIDIA still holds the largest share, with a small portion allocated to Qualcomm's 8,650 and Horizon's HSD solution [18] Self-Developed Chips - NIO is the fastest in self-development, with its ZAM chip already in use, while other companies like XPeng and Li Auto are at different stages of development. The advantages of self-developed chips include better software-hardware coupling and cost savings [7][9] - However, self-developed chips face disadvantages such as limited manufacturing capacity and high R&D costs, which could lead to higher costs compared to externally sourced chips if sales volumes are insufficient [8][9] Future Trends - The trend towards self-developed chips is not expected to be prevalent among traditional automakers like Geely and Chery due to their complex R&D systems and high investment risks. In contrast, emerging companies like NIO and XPeng are more inclined to pursue self-development as it aligns with their brand identity [9][10] - The integration of cockpit and intelligent driving functions is currently less common due to high performance requirements and rapid iteration speeds, with a trend towards separating these functions until higher-performance chips become available [15] Geopolitical and Regulatory Impact - Geopolitical factors and tariffs have limited effects on the vehicle chip industry, with few restrictions on chip exports from China to other countries, except the U.S. Companies like Horizon face slow overseas expansion due to insufficient data accumulation and mass production experience [20]

从底层芯片技术到行业生态协同的全方位展示。 在车展首日， 黑芝麻智能创始人兼 CEO单记章在 发布会 开场演讲中回顾了公司成立八年发展历 程：凭借自研 ISP、NPU两大核心IP，构建了华山系列（辅助驾驶芯片）与武当系列（跨域融合芯 片）两大产品线，并成功登陆港交所成为"智能汽车AI芯片第一股"。过去几年，公司通过A1000 （本土首款车规级高算力芯片）、C1236（全球首款跨域融合芯片）等产品持续突破，2025年推出 的A2000系列更以全景通识高算力设计，率先支持大模型端侧推理，推动AI计算效率再升级。 面向未来，黑芝麻智能将以 "技术创新+开放生态"双轮驱动，聚焦芯片研发，赋能汽车智能化转型。 单记章指出，大模型进入端侧推理时代，计算效率与算法架构成为关键挑战。黑芝麻智能通过优化带 宽性能、探索混合模型架构，致力于突破"算法决定上限"的行业瓶颈。同时，公司立足汽车领域， 向机器人、边缘计算等场景延伸，以全栈技术能力开启智能时代的"计算革命"。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 据 IPO早知道消息，黑芝麻智能（2533. HK ）携 旗下明星产品华 ...

从底层芯片技术到行业生态协同的全方位展示。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 据 IPO早知道消息，黑芝麻智能（2533. HK ）携 旗下明星产品华山系列、武当系列芯片及其应用展 示，生态合作案例、合作产品及解决方案 等亮相 2025年上海国际车展。 值得注意的是，不久前的 3月31日 ，黑芝麻智能发布了 2024年业绩报告，这也是其成为"智能汽车 AI芯片第一股"后发布的首份年报 —— 财报显示， 2024年黑芝麻智能的营收同比增长至51.8%至 4.74亿元，毛利从2023年的0.77亿元大幅增长152.4%至1.95亿元，毛利率亦从2023年的24.7% 增加超16个百分点至41.1%。 截至目前，黑芝麻智能就旗下 A1000系列芯片和武当C1200系列跨域融合芯片已与超40家车企达成 合作。 财报发出后，多家知名券商纷纷给予黑芝麻智能 "推荐"/"增持"/"优于大市"等评级，以示对黑芝麻 智能长期价值的看好。 某种程度上而言，本次车展有助于黑芝麻智能再一次展现 从底层芯片技术到行业生态协同 的全方位 能力。 在车展首日， 黑芝麻智能创始人兼 C ...

汽车越来越向智能体进化，就越需要一套扎实的底层支撑。冲在前沿探索的新造车公司们，如蔚来、理想等都已经未雨绸缪，在整车操作系统层面，投入重 金，推出整车OS级产品。 3月27日，理想汽车CEO李想在中关村论坛上正式宣布开源其自研汽车操作系统"理想星环OS"。据介绍，理想星环OS是一款面向人工智能时代的汽车操作系 统，包含智能车控、智能驾驶、通信中间件、信息安全系统四大重要支柱。 日前，理想汽车CTO谢炎，也接受36氪在内的媒体访谈，进一步阐述理想星环OS的发展历程、战略使命和技术亮点。 据谢炎介绍，理想星环OS项目于2021年启动，投入了两百人的研发团队，这些年累计超过十亿的研发费用，并在2024年首次实现自研操作系统上车。 谢炎是国产操作系统届的老兵，在加入理想汽车前，他曾在华为担任消费者事业群软件工程副总裁，OS部部长，负责鸿蒙OS等操作系统技术的研发工作。 在此之前，谢炎于2014至2019年在阿里巴巴AliOS事业群担任首席架构师和轮值总经理，负责移动操作系统的技术架构设计，和研发管理工作。 l理想汽车CTO谢炎，图：理想汽车 2020年，疫情引发全球供应链紧绷，历时3年的芯片慌爆发。2022年7月，谢 ...