2025年末收官之际，智驾领域各类榜单密集出炉。其中，高工智能汽车研究院最新发布的2025年智驾计 算芯片榜单，清晰勾勒出当前行业的竞争图景：整体市场马太效应显现，智驾芯片市场头部优势高度集 中，优势玩家借助软硬一体切入高阶智驾市场，增速明显。 支持城市NOA的高阶智驾计算芯片市场，则在2025年迎来大爆发。数据显示，中国市场（不含进出 口）乘用车前装标配（含限时免费促销）城区NOA车型交付量达207.09万辆，同比激增155.83%，渗透 率升至15.18%；城市NOA功能正逐步向中高端主流价格带渗透，有效拉动高阶智驾计算平台放量。 在城区NOA计算芯片市场，头部集中态势更为突出，英伟达、华为、地平线前三名合计占据90%的市场 份额。其中英伟达仍保持领跑，领先第二名华为约26%的市场份额；第三名地平线则依托征程6M/6P在 2025年集中量产上车，成为城区NOA市场份额增长最快的厂商之一，与华为市场份额差距并不算太 大。值得关注的是，随着地平线HSD持续量产交付，及携手国内外Tier1厂商共同推进基于单征程6M的 城区NOA量产渗透至10万元级国民车市场，或将引发2026年高阶智驾市场更为激烈的竞争。 随着 ...

Core Viewpoint - The article discusses the first Horizon Technology Ecosystem Conference held in Shenzhen, focusing on the development of smart vehicles and robots, and the collaboration among various industry players to promote technological advancements and accessibility [3][12]. Group 1: Market Achievements and Product Launches - Horizon has successfully shipped over 10 million units of its Journey series chips within five years, with the latest HSD model achieving over 12,000 activations in just two weeks after its launch [5]. - The company aims to reach a production target of 10 million units for the HSD model in the next 3-5 years [5]. Group 2: Technological Advancements - The fourth-generation compiler technology has significantly improved compilation speed from hours to minutes, with a 20% enhancement in model performance [6]. - The new Riemann architecture is expected to deliver a tenfold increase in computing power and a fivefold improvement in efficiency compared to the previous generation [6]. Group 3: New Service Models and Collaborations - Horizon introduced the HSD Together algorithm service model, which promises to reduce human, computing, and time costs by 90% for smart driving partners [7]. - The company announced that its urban auxiliary driving solution based on the Journey 6M model will soon be mass-produced, making it accessible for vehicles priced around 100,000 yuan [8]. Group 4: Expansion into Robotics - Horizon is expanding its technology from smart driving to general robotics, launching two foundational models: HoloMotion for embodied intelligence and HoloBrain for cognitive capabilities [9]. Group 5: Industry Challenges - The automotive industry is facing a potential storage chip supply crisis due to rising demands for higher performance and larger capacity products, with projections indicating that the supply satisfaction rate may fall below 50% by 2026 [11][12].

Core Viewpoint - The development of the autonomous driving L3 technology has increased the demand for high-performance visual perception systems, particularly in terms of signal transmission integrity, anti-interference, and stability [2]. Group 1: Company Developments - The company, Weifeng Electronics (301328.SZ), has leveraged its 20 years of experience in the connector industry to quickly develop high-frequency and high-speed connector products such as FAKRA, MINI FAKRA, HSD, and automotive Ethernet [2]. - These products are designed to meet the data transmission needs of smart cockpits, autonomous driving systems, and vehicle networking, providing customized solutions for onboard radar and camera modules [2]. - The company has achieved bulk supply to several leading automotive manufacturers through the synergistic supply of medium and low-voltage connectors alongside high-frequency and high-speed products, creating an integrated service advantage [2]. Group 2: Future Outlook - The company plans to continue deepening technological iterations and accelerating the expansion of new business opportunities in the intelligent driving sector [2].

Core Viewpoint - The development of the autonomous driving industry imposes higher requirements on the integrity, anti-interference, and stability of visual perception system signal transmission [1] Group 1: Company Developments - The company has leveraged its 20 years of expertise in the connector field to rapidly develop high-frequency and high-speed connector product lines, including FAKRA, MINIFAKRA, HSD, automotive Ethernet, and TYPEC [1] - These products are precisely tailored to meet the data transmission needs of smart cockpits, autonomous driving systems, and vehicle networking [1] - The company provides customized solutions for automotive radar, cameras, and other modules, enhancing its service offerings [1] Group 2: Market Position and Future Outlook - The company has achieved bulk supply to several leading automotive manufacturers through the synergistic supply of medium and low-voltage connectors alongside high-frequency and high-speed products [1] - Future plans include continuous deepening of technological iterations and accelerating the expansion into new business opportunities within the intelligent driving sector [1]

Core Insights - Horizon Robotics achieved a market capitalization of HKD 126 billion by December 16, 2025, reflecting a growth of over 90% since its IPO, which is the largest IPO for a Chinese tech company in the past three years [3][9][8] - The company positions itself as a "Horizontal Platform" for the Robotics era, emphasizing an open and win-win ecosystem [4][12] Company Overview - Horizon Robotics, founded by Yu Kai, is described as a non-typical chip company, focusing on "processing hardware" rather than traditional chip terminology [4] - The company has shipped over 7 million units of its "processing hardware" [5] IPO Significance - The IPO not only marked a significant financial milestone but also attracted long-term international funds, injecting confidence into the Hong Kong market [9] - The timing of the IPO coincided with a market recovery, which Yu Kai humorously attributed to "luck" [8] Ecosystem Ambitions - Horizon's vision includes creating a vibrant ecosystem within the automotive industry, aiming to foster innovation and improve safety and quality of life [12][14] - The first technology ecosystem conference showcased participation from major automotive manufacturers, highlighting the collaborative nature of Horizon's approach [13] Technological Foundations - The company introduced its fourth-generation BPU architecture, named "Riemann," which aims for significant performance improvements through algorithm and architecture optimization [18] - The new architecture enhances tensor calculations and supports vector and floating-point computations, achieving a fivefold increase in energy efficiency [18] HSD Together Model - Horizon launched the HSD Together model, which aims to reduce partners' costs and time by 90%, transforming the business model from selling products to providing services [28][30] - This model is designed to attract more automotive manufacturers, particularly mainstream brands, to join Horizon's ecosystem [29] Competitive Strategy - Horizon's strategy contrasts with competitors like Tesla and Huawei by adopting a semi-open approach, focusing on collaboration rather than strict control [31] - The goal is to establish a trust-based ecosystem that fosters innovation and efficiency, positioning Horizon as a key player in the smart automotive and physical AI markets [32] Future Vision - Horizon aims to leverage its ecosystem to drive technological advancements and ensure long-term growth, with a focus on making technology accessible to everyday users [36]

Group 1 - The cumulative shipment of the Journey series has exceeded 10 million units, with the self-developed BPU architecture showing a performance improvement of 246 times over six years [1] - The Journey 6 series, set to begin mass production in 2025, is based on the new Nash architecture, covering computing power from 10 to 560 TOPS, and is the only domestic solution meeting the needs for L2 to full-scene urban driving assistance [1] - The Journey series has over 40 domestic and international automotive clients, with more than 400 models and serving over 6 million vehicle owners [1] Group 2 - The HSD is entering a large-scale mass production phase, with its technology path gaining widespread market validation and recognition [2] - The HSD employs a one-piece end-to-end architecture with reinforcement learning capabilities, significantly enhancing the safety, efficiency, and comfort of driving assistance [2] - The new flagship vehicle computing solution, Journey 6P, achieves a performance leap with up to 560 TOPS, supporting the large-scale implementation of HSD [2] Group 3 - The company expects revenues of 3.59 billion, 5.60 billion, and 8.14 billion yuan for 2025, 2026, and 2027 respectively, with net profits of -703 million, -151 million, and -83 million yuan [2]

Group 1 - The core viewpoint of the report is that Horizon Robotics-W (09660) has achieved cumulative shipments of over 10 million units for its Journey series, maintaining a "Buy" rating [1] - The company's self-developed intelligent computing architecture, BPU, has undergone three iterations, improving its computing performance by 246 times over six years [1] - The Journey 6 series, set to begin mass production in 2025, is based on the new generation BPU Nash architecture, covering 10-560 TOPS computing power, and is the only domestic solution that meets the needs for L2 assisted driving to full-scene urban assisted driving [1] Group 2 - The HSD (High-Performance Smart Driving) technology is entering large-scale mass production, with a diverse range of vehicle models being covered [2] - HSD employs a one-piece end-to-end architecture with reinforcement learning capabilities, significantly enhancing the safety, efficiency, and comfort of assisted driving [2] - The new flagship vehicle intelligent computing solution, Journey 6P, achieves a generational performance leap with up to 560 TOPS computing power, supporting the large-scale implementation of HSD [2]

Group 1 - The core viewpoint of the articles highlights the significant advancements and market presence of Horizon Robotics, particularly with its Journey series achieving over 10 million units shipped and maintaining a "buy" rating from Zhongyou Securities [1] - The company's self-developed intelligent computing architecture, BPU, has seen a performance increase of 246 times over six years, evolving through three generations from Bernoulli to Nash architecture [1] - The Journey 6 series, set to begin mass production in 2025, is based on the new Nash architecture and offers computing power ranging from 10 to 560 TOPS, making it the only domestic solution that meets the diverse needs of smart vehicles from L2 to full-scene urban driving [1] Group 2 - The HSD (High-Performance Smart Driving) technology is entering large-scale production, with a growing number of vehicle models being covered, indicating strong market validation and recognition [2] - HSD employs a one-piece end-to-end architecture with integrated reinforcement learning capabilities, significantly enhancing the safety, efficiency, and comfort of assisted driving [2] - The new flagship vehicle intelligent computing solution, Journey 6P, achieves a generational performance leap with up to 560 TOPS of computing power, supporting the large-scale implementation of HSD [2]

Investment Rating - The report maintains a "Buy" rating for the company [2][7] Core Insights - The company has achieved cumulative shipments of over 10 million units for its Journey series, with its proprietary BPU® architecture showing a 246-fold performance improvement over six years. The new Journey®6 series, set for mass production in 2025, is the only domestic solution that meets the diverse needs of smart vehicles from L2 to full-scene urban driving [5][6] - The company is entering a rapid growth phase with its HSD (High Safety Driving) technology, which has been validated by the market and is now in mass production. The flagship Journey 6P, with a computing power of up to 560 TOPS, supports the large-scale deployment of HSD, with partnerships established with over 10 automotive brands and more than 20 models [6] - Revenue projections for the company are estimated at 35.9 billion, 56.0 billion, and 81.4 billion yuan for the years 2025, 2026, and 2027 respectively, with expected net losses of 70.3 million, 15.1 million, and 8.3 million yuan for the same years [7][9] Financial Summary - The company is projected to achieve revenues of 35.89 billion yuan in 2025, with a growth rate of 51%, and a net profit attributable to the parent company of -7.025 billion yuan [9][12] - The company's operating income is expected to grow at a compound annual growth rate (CAGR) of 54% from 2024 to 2027, with a projected net profit margin improvement over the same period [12] - The company’s asset-liability ratio is projected to increase significantly, reaching 95.56% by 2027, indicating a high level of leverage [12]

Core Viewpoint - Horizon Robotics has made significant advancements in high-level assisted driving technology, aiming to make advanced driving experiences accessible to a broader audience [1] Group 1: Company Achievements - Horizon Robotics has been actively pushing high-level driving technology since 2021, achieving notable results in the past two years [1] - The company plans to launch the Journey 6 family of chips on April 24, 2024, followed by the Journey 6P and HSD on April 18, 2025 [1] - The HSD will officially start mass production on November 18, 2025, providing a luxury driving experience that lowers the barriers to high-level assisted driving [1] Group 2: Market Response - The first vehicles equipped with HSD, namely the Xingtu ET5 and Deep Blue L06, were launched just two weeks ago, with over 12,000 units activated, indicating a strong market response [1] - The positive reception highlights Horizon Robotics' robust capabilities and promising future in the high-level assisted driving sector [1]