市場調查報告書
商品編碼
1513565
ADAS/自動駕駛一級供應商,中國公司(2024年)ADAS and Autonomous Driving Tier 1 Suppliers Research Report, 2024 - Chinese Companies |
中國智慧駕駛市場在L2時代由國外供應商主導。隨著進入L2+及以上時代,包括L2+、L2.5、L2.9,國內ADAS供應商已開始佔據主導地位。因此,相較於兩年前,國內ADAS Tier 1陣容逐年擴大。
1.Tier 1加速先進智慧駕駛功能發展,汽車產業進入智慧駕駛 "標配時代"
隨著軟體演算法的不斷迭代以及大規模算力晶片、感測器等硬體成本的降低,智慧駕駛的落地在軟硬體上獲得了堅實的支撐。目前,國內乘用車ADAS系統功能(L1至L2.9)的數量及搭載率穩定提升,L2、L2++自動駕駛採用率快速提升。與2022年相比,2023年安裝的L2、L2+、L2.5和L2.9的數量分別增加了37.0%、71.9%、124.9%和63.1%。截至2024年1月至2024年4月,配備L2以上等級的乘用車滲透率從2022年的34.8%提升至2024年(1月至4月)的53.8%。
目前,智慧駕駛市場正處於加速落地的窗口期,為各大車企打造 "最強智慧駕駛" 創造了重大發展機會。從新登記車輛來看,L2以上ADAS功能搭載率快速提升,其中L2.5、L2.9增幅明顯。 2022年,中國新登記乘用車L2.5功能搭載率為13.25%。截至2024年(1月至4月),已增至19.86%。 L2.9功能的安裝率從12.36%成長到2022年的23.4%。這符合國內整車廠和Tier 1重視先進ADAS、大規模部署駕駛停車一體化、NOA解決方案的趨勢。
從公司類型來看,目前L2級ADAS以合資品牌為主。 2023年,47.48%的合資車型搭載L2,2024年(1-4月)搭載率達55.05%。對L2.9特性來說,中高端是重要的目標市場。 25萬至30萬元、30萬元以上的車型安裝率較高。其中,25萬元至30萬元價格區間安裝率最高。截至2024年(1月至4月),此價格區間乘用車搭載率由2023年的23.34%提升至40.89%,50萬元以上高階車型搭載率由2023年的6.73%提升至 2024年(1 2024年(1124年(1024年(1024)年(1024年 1024年。月至4月)將升至24.39%,成為成長最快的。
此外,隨著智慧駕駛技術路線日趨清晰,不少主機廠在技術路徑、城市規模、落地速度、成本等方面都聚焦在城市NOA,競爭已進入白熱化階段。 L2.9的安裝量開始呈現下降趨勢,20萬~25萬元價格區間的L2.9安裝率從2023年的5.15%上升到9.32%。這顯示使用者對高階自動駕駛的認知度和接受度正在逐漸增強,高階自動駕駛也體現了眾多主機廠的核心競爭力。隨著ADAS的水平提高,感測器的數量和種類也會增加。在硬體層面,主流的解決方案可以概括如下:
L1 ADAS方案:主要以1V或1R實現,功能主要包括ACC/AEB。
L2 ADAS方案:1R1V、3R1V為主流,佔80%左右。
L2+ ADAS方案:5R1V為主流,佔比逐年提升,2023年將達69.51%。
L2.5 ADAS方案:6V1R、1V5R為主流,佔比超過50%。相機數量顯著增加。
L2.9 ADAS 解決方案:大多數解決方案都部署基於 L2.5 的 LiDAR。
本報告提供中國汽車產業的研究和分析,包括一級ADAS/自動駕駛供應商的產品和解決方案比較以及競爭格局。
ADAS Tier1s Research: Suppliers enter intense competition while exploring new businesses such as robotics
In China's intelligent driving market, L2 era is dominated by foreign suppliers. Entering era of L2 + and above (including L2 +, L2.5 and L2.9), domestic ADAS suppliers have begun to dominate. Therefore, compared with two years ago, the lineup of domestic ADAS Tier1 is expanding year by year. The "Chinese ADAS and Autonomous Driving Tier 1 Suppliers Report, 2021-2022" studied 7 Tier1s, 2023 version studied 12 Tier1s, and this 2024 version expanded to 20 Tier1s.
1.Tier 1 accelerates the development of advanced intelligent driving functions, and the automotive industry enters "standard configuration era" of intelligent driving
With the continuous iteration of software algorithms and reduction of hardware costs such as large computing power chips and sensors, the landing of intelligent driving has solid software and hardware support. At present, the installation volume and installation rate of domestic passenger car ADAS system functions (L1-L2.9) have been steadily improved, and L2 and L2 ++ autonomous driving are in the stage of rapid improvement in terms of penetration rate. In 2023, the installation volume of L2, L2 +, L2.5, and L2.9 increased by 37.0%, 71.9%, 124.9%, and 63.1% year-on-year compared with 2022. As of January-April 2024, the penetration rate of passenger cars equipped with L2 and above rose from 34.8% in 2022 to 53.8% in 2024 (Jan.-Apr.).
At present, the intelligent driving market is in a window period of accelerated penetration, providing huge development opportunities for major automakers to build "the most powerful intelligent driving". From the perspective of newly listed cars, the installation rate of L2 and above ADAS functions has risen sharply, of which L2.5 and L2.9 have increased significantly. In 2022, the installation rate of L2.5 functions of newly listed passenger cars in China was 13.25%. As of 2024 (Jan.-Apr.), it has increased to 19.86%. The installation rate of L2.9 functions has increased from 12.36% in 2022 to 23.4%. This is consistent with the trend of domestic automakers and Tier 1s focusing on advanced ADAS, large-scale landing of driving-parking integration and NOA solutions.
According to enterprise type, joint venture brand is the main force of L2 ADAS currently. 47.48% of joint venture models in 2023 are equipped with L2; the installation rate reached 55.05% in 2024(Jan.-Apr.). Mid-end and high-end models are important target markets for L2.9 functions. The installation rate of models priced at 250-300,000 yuan and more than 300,000 yuan is higher. Among them, the installation rate of 250-300,000 yuan price range is the highest. As of 2024(Jan.-Apr.), the installation rate of passenger cars in this price range has increased from 23.34% in 2023 to 40.89%; the installation rate of high-end models priced over 500,000 yuan has the fastest growth rate, increasing from 6.73% in 2023 to 24.39% in 2024(Jan.-Apr.).
In addition, as the intelligent driving technology route becomes clearer, many OEMs are making intensive efforts in urban NOA in terms of technical path, city scale, landing speed and cost, and the competition has entered a white-hot stage. L2.9 installation has begun to show a downward trend, and L2.9 installation rate in the price range of 200-250,000 yuan has risen from 5.15% in 2023 to 9.32%; indicating that users' recognition and acceptance of high-level autonomous driving are gradually increasing, and high-level autonomous driving also reflects the core competitiveness of many OEMs.
As ADAS level increases, the number of sensors increases and variety becomes richer. At the hardware level, the mainstream solutions are summarized as follows:
L1 ADAS solution: mainly realized through 1V or 1R, and the functions are mainly ACC/AEB, etc.
L2 ADAS solution: 1R1V and 3R1V are the mainstream, accounting for about 80%.
L2 + ADAS solution: 5R1V is the mainstream, and the proportion is increasing year by year, reaching 69.51% in 2023.
L2.5 ADAS solution: 6V1R and 1V5R are the mainstream, accounting for more than 50%. The number of cameras has been significantly improved.
L2.9 ADAS solution: Most solutions introduce LiDAR based on L2.5.
1) In April 2024, Baidu officially released ANP3 Pro, another product of pure visual urban NOA, which reduces one NVIDIA DRIVE Orin compared with the ANP3 Max version; the Max series is aimed at segmented users that "pursuit extreme intelligent driving products, and the car price is over 250,000 yuan", the Pro series is aimed at "150-250,000 yuan "market segment, with more cost-effective high-end intelligent driving (" reduced configuration version ") solution.
2) In April 2024, Momenta announced that it will launch a variety of scalable automotive intelligent driving solutions in conjunction with Qualcomm. By leveraging Momenta's "one flywheel" core technology and Snapdragon Ride's SA8620P and SA8650P platforms, a complete set of functions can be reused. Adopting flexible sensor configuration solutions brings users more cost-effective solutions and a wider range of smart driving applications.
3) Zhang Kai, chairperson of Haomo.ai, believes that China's intelligent assisted driving is in a big market explosion, and the more cost-effective driving-parking integrated domain control solution will become the mainstream. In October 2023, Haomo.ai launched the "three ultimate cost-effective" intelligent assisted driving product HPilot 2.0, seizing the smart driving market.
4) In 2024, PhiGent Robotics launched the ultimate cost-effective "PhiGo Pro" (for 100-200,000 yuan market segments) and "PhiGo Pro Plus" (for 200-300,000 market segments) smart driving solutions, and controlled the cost within 4,000 yuan and 5,500 yuan respectively.
With the acceleration of automotive intelligent connection wave, the automotive industry chain, technology chain, and value chain are accelerating deconstruction and reshaping. In terms of industrial chain, the original industrial boundaries are expanding, and more specialized new industrial entities are emerging. At the same time, the vertical chain industrial structure is shifting to a horizontal mesh structure, and new local industrial chains will be derived. The role positioning and development model of OEMs and Tier1 will undergo major changes. First, the automotive industry will further evolve into a multi-party ecosystem, with participants including OEMs, Internet companies, ICT companies, artificial intelligence companies, cloud computing service providers, big data companies, Tier2/Tier1/Tier0.5 suppliers, and governments. This multi-party model will promote the collaborative development and innovation of the entire industrial chain.
Secondly, the past vertical supply model will be broken, and the automotive supply chain will begin to develop into a mesh. In the past, automotive OEMs mainly defined functional architectures and participated in system integration work, and supply chain implemented the vertical supply model of Tier3 -> Tier2 -> Tier1. OEMs mainly cooperated directly with Tier1 suppliers. Under the trend of the overall change of "software-defined vehicles", in order to achieve more functional differentiation features, improve development efficiency, and realize code reuse, OEMs will also participate in the development of applications, and the automotive supply chain will begin to develop into a mesh.
At present, some automakers purchase software and hardware separately, and the procurement method has become very flexible. A set of integrated driving-parking system can be divided into sensors, controllers, system integration, application software development and other components. For the necessary and capable parts, automakers will choose to develop their own, while other parts will be purchased through suppliers.
At present, Tier1 is mainly provided to OEMs through gray-box or white-box mode.
White-box mode: Tier1 is responsible for hardware production, middle layer and chip solution integration, OEM is responsible for software part of application layer of autonomous driving, or Tier1 is only responsible for hardware production, OEM or its designated software supplier is responsible for the system architecture and application layer development of domain controllers. Typical cases: Desay SV + NVIDIA + Xiaopeng/Li Auto/IM, Zeekr + Mobileye + iMotion, etc.
Gray-box mode: OEMs put forward customized requirements, and Tier1 provides R & D services, which are finally presented in the form of independent R&D by OEMs. At the same time, OEMs may also develop their own domain controller system architecture and autonomous driving application layer algorithm software development. The final product logo is designated by OEM.
Black box model: Tier1 cooperates with chipmakers to achieve solution integration, develop a central domain controller, and sell it to OEMs.
In the past year, ADAS market has entered a stage of intense competition, and most suppliers are facing losses. At present, only the top Tier 1 can achieve profitability. Second-tier Tier 1 is striving to expand market share and achieve profitability in the next few years by expanding mass production.
In addition to competing for more ADAS production orders, many Tier1s are also entering the fields of AI robots, flying cars, and more.
In early 2024, Zongmu Technology announced that its new subsidiary Cancong Robot officially launched a new product - FlashBot Lightning Treasure. According to reports, this product is not only an energy robot, but also can be used as a "mobile power bank" for new energy vehicles. Equipped with L4 intelligent driving capability and 104KWh capacity, it can provide intelligent charging services for new energy vehicles in parks, parking lots and other scenarios.
In April 2024, iMotion Technology announced that the company had become the autonomous driving solution provider of "a well-known domestic flying car company" and would provide it with an integrated software and hardware autonomous driving solution with the core of autonomous driving domain controller iDC High.