PUBLISHER: ResearchInChina | PRODUCT CODE: 1930698
PUBLISHER: ResearchInChina | PRODUCT CODE: 1930698
China Passenger Car Navigate on Autopilot (NOA) Industry Report, 2025
In 2025, NOA standardization was popularized, refined and deepened in parallel. In 2026, core variables will be added to the competitive landscape.
The evolution of autonomous driving follows a clear and step-by-step path of technological advancement: highway NOA -> commute urban NOA -> mapless city NOA -> experience deepening + safety refinement -> L3 + universal autonomous driving.
With its unique advantages like "low complexity and high value", highway NOA has become the "pioneer" in the implementation of autonomous driving. Since its advent in 2021, this technology has reached an explosive turning point in 2025. The monthly penetration rate of standard highway NOA soared from 5.8% in January 2025 to 19.6% in October 2025 with a 3-fold increase in 9 months, marking that domestic passenger car highway NOA officially entered the era of standardization.
After highway NOA laid the market foundation, commute NOA came into being in 2023. As a transitional solution for urban NOA, commute NOA is achieved rapidly thanks to repeated training on fixed routes, without waiting for the HD maps to be fully rolled out. The function can be unlocked in just a week for simple routes and 2-3 weeks for complex routes, effectively lowering the implementation threshold for urban autonomous driving. From the second half of 2023 to the second half of 2024, a fierce competition in the urban NOA field emerged, and OEMs accelerated their efforts herein. The implementation of end-to-end foundation models and mapless solutions pressed the accelerator button for the industry, helping it enter a new mapless stage of "full-domain development across the country".
In 2025, the development logic of urban NOA underwent a fundamental change. The industry transferred from the extensive competition to the all-scenario closed loop of "D2D", the safety refinement of complex scenarios, and the deepening of the value of driving experience.
Four Core Evolution Directions of Urban NOA in 2025
Evolution Direction 1: All-Scenario D2D
One of the core breakthroughs of urban NOA in 2025 is the transition from "road autonomous driving" to "all-scenario closed-loop autonomous driving", with "D2D" and "point-to-point full-domain navigation" becoming the mainstream upgrade path. The essence of this trend is to solve the separate experience of "parking lots - public roads - destination parking lots" in traditional autonomous driving, extend the coverage of autonomous driving to the "last mile" scenario, and ultimately realize autonomous driving from "P gear to P gear" without any breakpoints, reconstructing the users' full-process mobility experience.
Specifically, various brands have implemented relevant functional layouts one after another. As a pioneer, XPeng launched the complete D2D autonomous driving link in January 2025, covering the full process from parking spaces, cruising in multi-storey parking lots, automatic passage at turnstiles, to driving on public roads, and precise parking at destinations, truly enabling users to "get in cars and go". In June 2025, XPeng further upgraded it to cover all routes and scenarios with comprehensive functions, adapting to more vehicles models and parking lot types, completely breaking up the separation between parking lots and public roads, and realizing "zero breakpoint" full-link services. This function is built on the technical architecture and hardware platform of XNGP, extending autonomous driving capabilities from highways and urban roads to terminal scenarios such as parking lots and parks, covering all scenarios.
Based on the Cedar system of NT3.0, NIO unveiled a point-to-point full-domain navigation assist system - NOP+ in June 2025, which highlights seamless connection between highways and urban scenarios and lays the foundation for the "D2D" fully closed loop. ZEEKR achieved a core breakthrough in the second quarter of 2025, launching the "D2D" function, which directly realizes the entire closed loop of autonomous driving from P gear to P gear from the starting parking space to the destination parking space. Users can enjoy a worry-free autonomous driving experience with one-time activation, accurately echoing the evolutionary trend of all-scenario closed loop.
Evolution Direction 2: Transferring from the extensive competition to the security redundancy and refined development in complex scenarios
The core of the evolution of urban NOA shifted from the extensive competition in 2023-2024 to complex scenario development and security system strengthening in 2025. The value transition from "ability to drive" to "good driving, stable driving, and driving suitable for all scenarios" accurately solves users' autonomous driving problems in unconventional road conditions.
Specifically, the technology iterations of brands closely follow this core logic. In November 2025, XPeng relied on VLA 2.0 to launch the "small road NGP" function to solve problems such as narrow alleys and unmarked roads, and increase the Miles per Intervention (MPI) on complex small roads by 13 times. NIO upgraded functions such as General Obstacle Alert and Assist (GOA) and Rear Collision Warning (RCW), greatly improving the recognition accuracy and braking response efficiency. Denza N7 expanded the effective vehicle speed of AEB to 120km/h, added a new special obstacle sensing model to enhance safety protection in different scenarios in an all-round way.
Evolution Direction 3: The autonomous driving experience is refined and upgraded, and the core logic shifts from "available functions" to "good experience"
Differentiated competition focuses on users' real driving experience, and autonomous driving shifts from standardized function coverage to scenario-based and personalized adaptation. The rear anti-motion sickness mode launched by NIO in November 2025 specifically solves the problem of dizziness among rear seat passengers. It uses the edge of the screen to display dynamic dots that are synchronized with the body posture in real time, and uses visual compensation technology to improve riding comfort. ZEEKR optimized full-domain NZP in the second quarter of 2025, added the U-turn function and introduced the "smooth driving" function, which not only improves traffic efficiency through intelligent planning of continuous road changes, but also uses the adjustment of the suspension and braking systems to achieve comfortable passing of speed bumps and greatly reduce the bumpy driving experience. Denza N7 focuses on personalized adaptation, supporting multi-gear customization options for intelligent driving modes such as parking speed and lane-changing style. Meanwhile, it optimizes the number of maneuvers in valet parking and expands the speed range for autonomous driving functions, enabling mapless autonomous driving to accurately match different users' driving habits. These measures jointly confirm that the focus of competition in high-level autonomous driving has shifted from the competition in technical parameters to the in-depth care for users' real driving experience.
Evolution direction 4: Autonomous driving evolves from an "auxiliary tool" to an "intelligent partner"
In 2025, autonomous driving moved from passive execution to active interaction. The VLA-based command function launched by Li Auto in September 2025 makes language a new interface for driving control, upgrading the car from a tool for passively executing instructions to an autonomous driving partner that can understand, think and act, lowering the threshold for enjoying autonomous driving.
Currently, this function is adapted to Li Auto's MEGA and L vehicles equipped with the AD Max system. Users need to manually turn it on in the CID settings. Its core capabilities cater to multiple types of driving scenarios: including mobility planning, basic control command execution (such as left turn, right turn, lane change, etc.), flexible adjustment of vehicle speed/vehicle distance/lane (such as "drive faster"), memory and management of driving preferences (such as "take the leftmost lane on this road from now on" "check all speed memories and delete the third one"); it also have active communication attributes. For example, when the roaming function of NOA is turned on in the park or underground garage, the system will actively ask the driver's intentions, and will give instructions such as "park nearby" and "pull over" as per the scenario, demonstrating the ability to think proactively beyond "passive execution".
At the safety level, the VLA-based command function achieves the synchronization of "experience upgrade" and "safety upgrade" by reducing human interference and strengthening emergency response. This is also the key to autonomous driving moving from "available functions" to "good experience". Specifically, the elimination of screen operation allows drivers to focus their attention on road conditions throughout the journey, reducing the risk of distraction. In emergencies, drivers can quickly intervene using simple commands such as "emergency stop," improving emergency response efficiency. At the same time, the system can automatically identify blind spots and take measures such as slowing down and yielding to pedestrians in advance, further expanding the safety margin of driving.
At the technical level, the "driver agent" is the core "intelligent cerebrum" that supports this capability. It first receives the user's voice instructions (no matter complex or simple), and at the same time collects real-time traffic conditions through multi-view cameras and LiDAR, and combines navigation data to encode the information into self-recognizable "Token information blocks", and then the cloud 32B VLA model disassembles the instructions into detailed operations. Finally, the automotive 4B VLA model (4 billion parameters) integrates all information (disassembled operations, encoded road conditions/navigation, pavement details), analyzes driving intentions and specific actions, generates driving trajectories and guides the vehicle to complete driving.
The maturity of technology and the improvement of scenarios have directly promoted the market penetration of NOA. From the NOA-related data of newly launched vehicle models from 2023 to October 2025, it can be seen that the configuration priority of highway NOA and urban NOA has seen a phased transition of "trial -> development -> standard configuration" with the maturity of the technology. Entering 2025, the penetration rate of urban NOA in new vehicles climbed to 28.1%, officially starting a new industry cycle of "full popularization + large-scale penetration". Urban NOA gets close to the explosion in 2026-2027.
In China, urban NOA is no longer an exclusive label for high-end vehicle models, but tends to cover those with varying price ranges. In 2022-2023, urban NOA was concentrated in high-end vehicle models priced at RMB300,000-350,000, exclusive to a small number of early users. After 2024, the penetration into vehicles priced at RMB100,000-150,000 and RMB150,000-200,000 accelerated. From January to October 2025, 150,200 vehicles valued RMB150,000-200,000 were sold with urban NOA, which had a penetration rate of 4.2%.
As per the current competitive landscape of the urban NOA market, top OEMs still focus on independent R&D, but third-party autonomous driving suppliers play a main force in "increment " instead of a supporting role in 2025-2026. As of October 2025, Huawei, Momenta, DeepRoute.ai, Zhuoyu Technology, Bosch and WeRide had achieved large-scale mass production and application of urban NOA. In 2026, open chip platforms represented by Horizon Robotics formed alliances with top algorithm companies such as QCraft and SenseTime, creating a "third model" between full-stack independent R&D of OEMs and full-stack outsourcing by suppliers with the core mission of achieving technological equality and cost revolution.
Reshuffling of urban NOA competitive landscape in 2026: Horizon Robotics emerges, QCraft becomes one of the core growth engines
Prediction 1: QCraft will rely on Li Auto's AD Pro and multiple OEMs to make urban NOA available in millions of vehicles in advance
QCraft will rely on Li Auto's best-selling AD Pro to make urban NOA available in millions of vehicles in advance as a rising star among third-party autonomous driving suppliers. The trend is signified in the NOA system installations from 2023 to October 2025 (as shown in the figure below): Li Auto's AD Pro ranked fourth among autonomous driving systems with NOA installed in 614,343 vehicles cumulatively, accounting for 6.8% of the total market. The core software algorithm of Li Auto's AD Pro is offered by QCraft. The high sales volume of Li Auto's AD Pro has become a key fulcrum for the rapid application of QCraft's urban NOA solution. On January 21, 2026, Li Auto fully pushed the urban NOA function to its vehicles fitted with AD Pro, so that QCraft's urban NOA technology landed in millions of vehicles. Millions of vehicles not only generate massive data, but also form a positive cycle of "data-algorithm-experience" to continuously drive system evolution, significantly reduce the marginal learning cost of each vehicle, and provide a solid foundation for rapid iteration.
Entering 2026, as solutions from suppliers such as Horizon Robotics and QCraft are applied on a large scale, the urban NOA market structure will be reshuffled. QCraft has become one of the biggest dark horses in the third-party urban NOA market in 2026 with the combination of "J6M + cross-chip adaptation + Driven-by-QCraft 2.0 + designated mass production of cross-brand multi-level vehicle models".
(1) Technical support: core competitiveness of single-J6M solution and end-to-end architecture
First of all, QCraft has reached a technological level with the industry's first single-J6M (Journey 6M from Horizon Robotics, 128TOPS) urban NOA solution mass-produced for the intelligent refresh version of Li Auto's L series vehicles, refreshing the application boundaries of medium-computing-power chips. This solution, based on Horizon Robotics' Journey 6M, integrates explainable one-model end-to-end technology and reinforcement learning, and has successfully embraced the mapless version of urban NOA. Through extreme computing power efficiency mining and optimization, it breaks through the application boundaries of medium-computing-power chips, allowing a 128TOPS platform to offer an autonomous driving experience comparable to that provided by a 256TOPS platform, while ensuring the security and explainability of the end-to-end system. This is its core advantage. As for technological innovation, the solution brings about the optimal experience under limited computing power through in-depth collaborative optimization of software and hardware, truly maximizing the value of each TOPS.
QCraft's "safe and interpretable one-model end-to-end + reinforcement learning" high-level autonomous driving technology architecture can be adapted to multiple chip platforms such as Horizon Robotics' Journey 6M and NVIDIA Orin Y. The architecture first receives heterogeneous perception data from multiple sources, including multi-frame temporal images from cameras, LiDAR point clouds, navigation maps, and ego vehicle pose data. It then generates a unified BEV world representation through a 3D encoder and a spatiotemporal BEV fusion module, providing a globally unified environmental cognition foundation for subsequent decision-making and reasoning. Next, the multi-task decoder outputs a series of explicit and interpretable intermediate features, such as traffic participant states, road topology, drivable areas, occupancy grids, and traffic signs. This design not only supports the model's internal decision-making but also fundamentally solves the "black box decision-making" problem of traditional end-to-end technology. Subsequently, the unified world state latent encoder encodes the BEV world representation into latent space features, and combines the navigation route with the flow matching planner to generate the initial driving candidate trajectory. After multi-agent motion prediction and multi-modal trajectory sampling, the safety reinforcement learning module that integrates reward functions and rule constraints finally selects the optimal driving trajectory that complies with safety regulations, realizing a full-link end-to-end closed loop of perception - decision-making - planning - control. This architecture not only retains the link efficiency of end-to-end technology, but also accurately solves the core problems of safety and interpretability in the field of high-level autonomous driving through explicit intermediate representation output and regularized constraints of reinforcement learning.
(2) Core product: iteration and layout of Driven-by-QCraft
"Driven-by-QCraft" is an autonomous driving solution launched by QCraft. It was first released in November 2022. It uses self-developed innovative technology as the core engine and integrates QCraft's full-stack software algorithm to support point-to-point autonomous driving in multiple urban scenarios, highways and expressways. When it was first launched, it was available in high, medium and low configurations to meet the diverse needs of OEMs.
Since then, the solution has continued to iterate around chip platform upgrades and core technology implementation. In April 2024, a new autonomous driving solution was launched based on the Journey 6 from Horizon Robotics to solidify the performance foundation for hardware adaptation. In April 2025, the urban NOA solution based on a single Journey 6M chip and end-to-end technology was implemented, achieving a key breakthrough in core scenario functions. On January 23, 2026, "Driven-by-QCraft 2.0" was officially released, with comprehensively upgraded capabilities.
So far, "Driven-by-QCraft 2.0" has been finalized into three solutions: Driven-by-QCraft Air (ultimate highway NOA), Driven-by-QCraft Pro (standard inclusive urban NOA), and Driven-by-QCraft Max (advanced ultimate urban NOA). It has cross-chip platform compatibility and can be adapted to domestic and foreign mainstream autonomous driving chips such as Horizon Robotics Journey 6M and NVIDIA Orin Y. It can be flexibly deployed according to requirements of OEMs, greatly reducing the cooperation and adaptation threshold and accelerating the mass production and application. "Driven-by-QCraft 2.0" covers all vehicles ranging from the entry level (RMB100,000) to high end (RMB400,000).
(3) Large-scale application of urban NOA: from a single brand to multiple OEMs
The list of vehicle models equipped with QCraft's urban NOA solution in 2026 indicates QCraft's cooperation matrix has covered many mainstream OEMs such as Li Auto, SAIC, GAC, Geely, ROX, etc.. QCraft has not only leveraged Li Auto's more than ten vehicle models fitted with AD Pro to form a large-scale delivery foundation, but also involved in the main product lines of SAIC Roewe, GAC Aion, Geely (including Galaxy, Xingyue and the like), ROX and other brands, covering vehicles from family cars to mid-to-high-end vehicles.
As these vehicles are mass-produced and delivered in 2026, QCraft's urban NOA solution will be applied to multiple brands and vehicle models of all levels on a large scale. Combined with millions of Li Auto's vehicles fitted with AD Pro sold, QCraft will become a core growth engine with both delivery scale and brand coverage in the field of urban NOA in 2026.
Prediction 2: Horizon Robotics will play a significant role in building an inclusive ecosystem for autonomous driving through an open model
Horizon Robotics is evolving from an autonomous driving chip supplier into a core builder of an autonomous driving inclusive ecosystem. Its open cooperation model dubbed "HSD Together" accurately addresses the core problem of current OEMs: in the context of sufficient supply of high-computing-power chips, most OEMs lack the technical capabilities and development efficiency to transform computing power into high-experience, high-level autonomous driving functions.
Based on this open cooperation model, Horizon Robotics opens its high-level autonomous driving algorithms that have been proven in mass production to its ecological partners, helping the latter significantly reduce research and development costs and shorten project cycles. Under the new cooperation framework, the partners can not only purchase chips, but also obtain all algorithm services in one stop. What is particularly critical is that the core model algorithm supports white-box delivery, leaving ample room for secondary development for OEMs. At the end of 2025, Horizon Robotics officially commenced the mass production of HSD which is a high-level autonomous driving solution. It was first installed on Deepal L06 and Chery ET5, and was launched and delivered simultaneously. This milestone marks that urban NOA with practical value has officially spread to vehicles priced below RMB150,000, and penetrated fast to vehicles priced at RMB100,000. The popularity of autonomous driving has entered the substantial application stage.
Prediction 3: SenseAuto defines a new level of autonomous driving safety with an AI native technology paradigm
Strategically, SenseAuto is positioned as a strategic partner to accelerate autonomous vehicles into the artificial general intelligence (AGI) era. It has built a "cockpit-driving-cloud" trinity AGI technology architecture, and formed a diversified product system consisting of intelligent driving, intelligent cockpits and AI cloud. In the field of autonomous driving, SenseAuto has unveiled AD Pro and AD Max based on Horizon Robotics J6E/J6M, and AD Ultra based on NVIDIA Orin/Thor to make NOA available in vehicles with varying price ranges. SenseAuto's multi-platform mass production solution has been steadily implemented. In March 2025, its purely visual autonomous driving solution based on Horizon Robotics J6M was mass-produced and applied to GAC Trumpchi S7. The company had fulfilled the mass production and delivery of urban NOA and end-to-end solutions based on multiple chip platforms such as J6M and NVIDIA Thor by the end of 2025.
SenseAuto's autonomous driving technology paradigm has always been centered on accurately predicting industry problems and accelerating the closed-loop implementation of technology. Each step of evolution has hit every key upgrade of autonomous driving: when the industry still relied on multi-module split architectures in 2022, SenseAuto was the first to launch the UniAD (one-model end-to-end solution) with integrated perception and decision-making, which directly outputs vehicle control instructions, solving the problems of information delay and data fragmentation in traditional architectures. It not only won the CVPR 2023 Best Paper Award, but also made autonomous driving officially enter the end-to-end era, and it took only 18 months to complete the verification from technical prototype to mass production; in 2024, in response to the common industry problem of "scarce long-tail scenario data and high labeling costs" for end-to-end models, SenseAuto launched the SenseWorld (autonomous driving world model), building pre-training capabilities by generating high-fidelity virtual scenarios, and built an "end-to-end data factory" with SAIC IM within only 3 months after the launch to achieve full-link simulation of 20+ core hazard scenarios, significantly reducing testing costs; in 2025, in order to break through the safety boundary from "driving" to "safe driving", SenseAuto further introduced reinforcement learning and launched the R-UniAD, which adopts the five-step method of "cold start->virtual enhancement->real fine-tuning->model distillation->automotive deployment" to allow the model to independently explore the safety boundary in the virtual scenario. In just half a year, the technology was promoted to mass production in cooperation with Dongfeng Motor, completing an efficient closed loop from academic breakthrough to industrial application. This evolutionary path of "one-model end-to-end -> generative world model -> multi-stage reinforcement learning" not only redefines the underlying logic of autonomous driving technology, but also becomes a Chinese paradigm for the transformation of the global autonomous driving industry from "function-oriented" to "safety-oriented" with a 1-2 year lead in technology.
SenseAuto's self-developed SenseWorld is the core engine of its end-to-end autonomous driving solution. Its crucial value lies in generating high-fidelity, multi-view simulated driving scenario videos to provide efficient training materials for the end-to-end autonomous driving model, breaking the crux of "real long-tail scenario data scarcity" in autonomous driving research and development, and building technical barriers with "generative AI-driven data efficiency" as the kernel.
In July 2025, SenseAuto officially launched the first generative world model platform, which was simultaneously open to B/C-side users for trial use, marking the technology's move from the research and development stage to large-scale application. Currently, SenseAuto has reached an in-depth strategic cooperation with SAIC IM. They have jointly built an "end-to-end data factory", leveraging generative AI technology to accurately overcome three core problems - "scarcity of long-tail scenario data, low annotation efficiency, and high testing costs" in autonomous driving research and development, and accelerate the safe application and capability iteration of IM AD 3.0 +. Based on the cooperation, the WorldSim-Drive (customized data set exclusive to IM) created by both parties has seen key breakthroughs and can cover the full-link simulation of more than 20 core hazard scenarios such as cut-in, collision warning, road occupation emergency braking, roundabouts, etc., providing high-value test samples for autonomous driving systems. In the future, the two parties plan to further build a library consisting of tens of millions of generative scenario, a full-dimensional test sample system, full-domain coverage of various driving scenarios, and continue to enhance the scenario adaptation capabilities and safety performance of IM's Autonomous driving system.
From the application of highway NOA to the technological maturity of mapless urban NOA, the evolution of autonomous driving has always been driven by both user value and underlying safety. It is an important trend to realize cockpit-driving integration and promote autonomous driving to new heights with AI technology. In September 2025, the Ministry of Industry and Information Technology of China (MIIT) issued the mandatory national standard "Safety Requirements for Combined Autonomous Driving Systems of Intelligent Connected Vehicles", which clearly delineates the core rules for DMS and EOR: when DMS detects "the driver's eyes are off the road", it must trigger EOR within 5 seconds.
This standard mandates that L2 and above autonomous vehicles be equipped with DMS as standard, directly promoting this technology from high-end vehicles to RMB100,000 vehicles which are the mainstream in the market, and prompting the OEM shipments to grow two-fold soon. In this process, leading suppliers represented by SenseAuto (from January to November 2025, SenseAuto ranked first in the domestic DMS supply market with a market share of 27.7%) have a competitive advantage thanks to "fatigue + distraction" detection technology and adaptation to extreme scenarios such as backlighting and drivers wearing sunglasses. Small and medium-sized vendors with insufficient technical barriers are being eliminated at an accelerated pace, and the industry's concentration tends to intensify.
In the future, with the continuous iteration of L3 and higher-level autonomous driving technology, the constant improvement of cross-scenario data closed loops, and the deep integration of the industrial chain collaborative ecology, autonomous driving will completely reshape human mobility and promote the comprehensive transformation of the autonomous industry from "vehicle manufacturing" to "intelligent mobility services". A new era of safer, more efficient, and more considerate mobility is rapidly approaching.
Table of Contents
1 Sales Volume and Solutions of Passenger Car Models Equipped with NOA in China
- 1 Sales and Solutions of NOA-enabled Passenger Car Models in China
- Sales Volume and Penetration Rate of NOA-enabled Passenger Car Models in China, 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Highway NOA (by OEM Type), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Highway NOA (by Price Range), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Highway NOA (by Auto Brand), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Highway NOA (by Energy Type), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Highway NOA (by Model), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Urban NOA (by OEM Type), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Urban NOA (by Price Range), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Urban NOA (by Auto Brand), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Urban NOA (by Energy Type), 2022-Oct 2025
- Sales Volume and Penetration Rate of Models Equipped with Urban NOA (by Model), 2022-Oct 2025
- 1.2 Main Sensor Solutions for NOA
- Overall Sensor Solutions of Models Equipped with Highway NOA, 2023-Oct 2025
- Overall Sensor Solutions of Models Equipped with Urban NOA, 2023-Oct 2025
- Urban NOA Sensor Solutions for Passenger Car Models of Major Brands in China, 2025
- 1.3 NOA Solutions of Major Suppliers
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (1)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (2)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (3)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (4)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (5)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (6)
- Comparison of Autonomous Driving Solutions between Major Tier 1 Suppliers in China (7)
- NOA Solutions of Major Foreign Suppliers and Their Layout in China
- 1.4 Competitive Landscape of NOA Market
- Highway NOA: Independent R&D Models of OEMs (1)
- Highway NOA: Independent R&D Models of OEMs (2)
- Highway NOA: Domestic OEMs' differentiated competitive strategies in independent R&D transfer from "must-have" to refining
- Market Share of Third-Party NOA Suppliers, 2023-Oct 2025
- Competitive Landscape of Urban NOA Market (1)
- Competitive Landscape of Urban NOA Market (2)
2 Passenger Car NOA Market Trends
- 2.1 NOA Market Trends
- Trend 1: In 2025, domestic passenger vehicles were equipped with highway NOA as standard, reshaping the benchmark for autonomous driving in passenger cars
- Trend 2: The optimal solution for urban NOA was applied, with hardware pre-embedding + OTA iteration
- Trend 3: NOA officially penetrated the automotive market on a large scale from high-end cars to vehicles of all price ranges in 2025
- Trend 4: From the consumer side, urban NOA became the core consideration for car purchasers, with the transition from differentiated configuration to mainstream standard configuration
- Trend 5: German, Japanese, and American traditional OEMs officially dabbled in NOA in 2025 (1)
- Trend 5: German, Japanese, and American traditional OEMs officially dabbled in NOA in 2025 (2): Audi
- Trend 5: German, Japanese, and American traditional OEMs officially dabbled in NOA in 2025 (3): Mercedes-Benz
- Trend 5: German, Japanese, and American traditional OEMs officially dabbled in NOA in 2025 (3): Toyota
- Trend 6: Chinese independent brand OEMs made NOA standard instead of an option
- Trend 7: The competition focus of urban NOA shifted from the extensive competition in 2024 to the deep cultivation of safe, anthropomorphic, smooth and efficient experience in 2025
- Trend 8: New vehicle models with urban NOA in 2025 mainly boasted 200-730TOPS in terms of autonomous driving computing power
- Trend 9: Autonomous vehicles will cover all vehicles, and urban NOA will spread to vehicle models priced at RMB100,000 in 2026
- 2.2 Evolution Directions of NOA in 2025
- H2 2023: Booming MNP (Commute NOA) Arena
- H2 2023-H2 2024: Urban NOA Becomes the Focus of Competition Which Emphasizes Extensive Coverage and Capability Enhance
- OEMs Accelerated the Application of NOA and Realized Mapless Urban NOA in 2024 (1)
- OEMs Accelerated the Application of NOA and Realized Mapless Urban NOA in 2024 (2)
- Autonomous Driving Vendors Accelerated the Application of NOA and Realized Mapless Urban NOA in 2024
- Mapless Urban NOA Case in 2024: Huawei
- Comparison of Mapless Urban NOA Solutions in 2024
- NOA Evolution Directions of OEMs in 2025 (1)
- NOA Evolution Directions of OEMs in 2025 (2)
- NOA Evolution Directions of OEMs in 2025 (3)
- Evolution Path: The evolution of autonomous driving follows the path: highway NOA -> mapless urban NOA -> deep experience cultivation + safety refinement -> L3 + universal autonomous driving
- Evolution Direction: The competition focus of urban NOA shifted from the extensive competition in 2024 to the deep cultivation of safe, anthropomorphic, smooth and efficient experience in 2025
- Evolution Strategy: Multiple brand OEMs adopt a layered autonomous driving strategy (independent R&D + multiple suppliers) to match brand positioning
- Popularization of DMS under the Urgent Need for Safety: Policies promote the standard configuration of autonomous driving, the DMS market experiences explosive growth, and leading suppliers enjoy benefits (1)
- Popularization of DMS under the Urgent Need for Safety: Policies promote the standard configuration of autonomous driving, the DMS market experiences explosive growth, and leading suppliers enjoy benefits (2)
- 2.3 End-to-End Foundation Models
- Overview of End-to-End Foundation Models Adopted by OEMs in 2025
- Comparison of Upgrade Solutions for End-to-End Foundation Models of Third-Party NOA vendors in 2025
- End-to-end Became the Mainstream Of Autonomous Driving Technology, and End-to-End Foundation Models Evolved into Multiple Routes in 2025
- VLA Route Analysis (1)
- VLA Route Analysis (2)
- Dual End-to-End Model Analysis
- Autonomous Driving Technology Paradigm of One-Model End-to-End + World Model + Reinforcement Learning
3 Passenger Car NOA Solutions and Application of OEMs
- 3.1 XPeng
- Profile
- Strategic Transformation
- Four Major Businesses
- Autonomous Driving Product Planning, 2025~2026
- Launch of VLA 2.0
- Analysis on VLA 2.0
- World Base Model Analysis (1)
- World Base Model Analysis (2)
- Cloud Factory
- Comparison of Cloud Computing Power and Autonomous Driving Data Reserves of OEMs
- Self-developed Turing Chip
- Autonomous Driving Evolution: Overview
- Autonomous Driving System: XNGP
- Autonomous Driving System: All-Scenario AI-Powered Autonomous Driving Becomes the Fourth Development Stage of XNGP
- Priorities of Autonomous Driving in 2025: D2D in All Scenarios
- Priorities of Autonomous Driving in 2025: From Developing New Cities to Optimizing Autonomous Driving Experience
- Autonomous Driving Experience Upgrade in 2025: D2D + Human-Machine Co-Driving + Small Road NGP
- Launch of L3 Autonomous Vehicles
- Realization of Mapless Urban NOA and L3 Autonomous Driving through Pure Vision
- Evolution of XNGP
- Key Technological Breakthroughs from XNGP 3.0 to 5.0
- Typical Vehicle Models with Urban NOA and Their Hardware Configuration Solutions, 2023-2024
- Typical Vehicle Models with Urban NOA and Their Hardware Configuration Solutions, 2025
- The First AI Vehicle Equipped with L3 Computing Platform in 2025: Hardware Configuration Solution of G7 Ultra
- Evolution of End-to-End Foundation Models
- End-to-End Foundation Models: Xbrain+Xnet+Xplanner (1)
- End-to-End Foundation Models: Xbrain+Xnet+Xplanner (2)
- 3.2 Li Auto
- Profile
- Organizational Architecture Adjustment
- Autonomous Driving Systems: AD PRO and AD MAX (1)
- Autonomous Driving Systems: AD PRO and AD MAX (2)
- AD Pro
- Evolution Direction of AD Max in 2024
- Evolution Direction of AD Max in 2025
- Typical Vehicle Models with AD MAX and Full-Stack Hardware and Software Configuration
- High-level Autonomous Driving Adds VLA-based Command Function, Marking a Watershed for Autonomous Driving from an "Assistant" to an "Intelligent Partner"
- Value of VLA-based Command Function
- Evolution of End-to-End Foundation Models for Autonomous Driving (1)
- Evolution of End-to-End Foundation Models for Autonomous Driving (2): VLA is expected to become a paradigm towards high-level autonomous driving
- VLA Driver Models (1)
- VLA Driver Models (2)
- Core Capabilities of VLA Models
- Evolution Direction of Capabilities of VLA Models
- Training Process of VLA Driver Models
- Technical Architecture of VLA
- Core Technology of MindVLA
- Application Scenarios of VLA Driver Models (1): The "Last 100 Meters" for Charging
- Application Scenarios of VLA Driver Models (2):
- End-to-end Solutions (1): Iterative Evolution of System 1
- End-to-end Solutions (2): System 1 (End-to-End Model) + System 2 (VLM)
- 3.3 NIO
- Profile
- Vehicle Model Platform Planning
- Technology Platform Evolution (1)
- Technology Platform Evolution (2)
- From Modeling to End-to-End, World Models Are Currently the Dominant Technology Paradigm
- Autonomous Driving System
- Autonomous Driving Technology Architecture: NAD Arch 2.0 (1)
- Autonomous Driving Technology Architecture: NAD Arch 2.0 (2)
- Comparison between Banyan and Cedar (1)
- Comparison between Banyan and Cedar (2)
- Panoramic View of Autonomous Driving Capability Upgrade of Banyan Based on NT2.0 in 2024-2025
- Upgrade Overview of Autonomous Driving Capabilities of Cedar Based on NT3.0 in 2025 (1)
- Upgrade Overview of Autonomous Driving Capabilities of Cedar Based on NT3.0 in 2025 (2)
- Iteration of NOP (1)
- Iteration of NOP (2)
- Organizational Architecture Adjustment of Autonomous Driving Team in 2024-2025
- NWM (1)
- NWM (2)
- World Model 2.0
- Comparison between End-to-End Model and World Model
- Comparison between VLA and World Model
- 3.4 BYD
- National Autonomous Driving Strategy in 2025
- Vehicle Model Layout and Planning (1)
- Vehicle Model Layout and Planning (2)
- Autonomous Driving 2025 and Future Strategic Priorities
- Three Autonomous Driving Solutions of "DiPilot" Achieve Comprehensive Coverage from Low-End Vehicles to High-End Vehicles, Accelerating Equal Access to Autonomous Driving
- An Innovative Front View Trinocular Camera Creates a "Super-Sensing Matrix".
- Layout in the Field of Autonomous Driving (1): Intelligent Computing Center
- Layout in the Field of Autonomous Driving (2): World Model Pre-research
- Organizational Architecture Adjustment of Autonomous Driving Team (1): Dual autonomous driving divisions integrate and pool resources to accelerate autonomous driving for all vehicles
- Organizational Architecture Adjustment of Autonomous Driving Team (2): An advanced technology research and development center was established to increase investment in AI and foundation models
- Evolution of DiPilot (1)
- Evolution of DiPilot (2)
- Evolution of DiPilot (3)
- Autonomous Driving Suppliers for NOA
- Typical Vehicle Models with Urban NOA (1): Denza N7 and Sensor Configuration
- Typical Vehicle Models with Urban NOA (1): Denza N7 Featured Mapless Urban CNOA at the End of 2024
- Typical Vehicle Models with Urban NOA (1): Autonomous Driving Function Iteration of Denza N7 in 2025
- Typical Vehicle Models with Urban NOA (2): Yangwang U8
- Autonomous Driving Ecological Strategic Partners
- 3.5 GAC
- "Smart Mobility 2027" Strategy
- Operation in H1 2025
- Product Matrix
- Autonomous Driving Business Layout: Investment and Cooperation (2017-2023)
- Autonomous Driving Business Layout: Investment and Cooperation (2024-2025)
- Evolution of ADiGO (ADiGO 1.0->ADiGO 6.0)
- Iteration of Autonomous Driving System: Hardware, Software Algorithms, And Sensor Configuration
- Five Autonomous Driving Platforms in 2025
- L2.9 Vehicle Models and Urban NOA Software Algorithms/Autonomous Driving System Suppliers
- Through the layout of "dual-gradient autonomous driving suppliers + scenario -accurate price matching", urban NOA is available in high-end vehicles and open to the public
- Farwon was established to target the high-end market and improve the brand matrix with the "Intelligent Manufacturing + Huawei Intelligence" model
- Farwon's first vehicle model F03 is expected to be launched in Q2 2026
- Momenta 5.0's one-model end-to-end algorithm spreads to RMB150,000 vehicle models and features urban NOA
- Trumpchi Xiangwang S7 Will Be Equipped with Momenta R6
- Autonomous Driving Ecological Strategic Partners: Multi-vendor Layered Binding + All-Scenario Technology Coverage" (L2+/L2.5/L2.9/L3/L4)
- 3.6 Geely
- Production Base Layout in China and Financial Data of BYD and Chery in 2024
- Autonomous Vehicle Full-Domain AI Strategy
- Product Portfolio by Brand (1)
- Product Portfolio by Brand (2)
- Sub-brands Are Reintegrated into "Geely"
- Geely Is Merge with ZEEKR into a Single Entity - "Geely"
- Organizational Architecture Adjustment of Autonomous Driving Team in 2025: Implementation of "One Autonomous Driving Strategy" (1)
- Organizational Architecture Adjustment of Autonomous Driving Team in 2025: Implementation of "One Autonomous Driving Strategy" (2)
- AFARI: The Core Carrier of Geely Group's Intelligent Strategy
- AI Computing Power Base: Xingrui Intelligent Computing Center 2.0
- Xingrui AI Foundation Model: Three Basic Models
- Xingrui AI Foundation Model: 6 Major Capability Models
- G-Pilot
- G-Pilot (H1-H9) Adopts a Differentiated Layering Strategy for Autonomous Driving Chip and Software Algorithm Suppliers
- ZEEKR's Autonomous Driving Solution Evolution (1): End-to-End Era and Highway NZP+
- ZEEKR's Autonomous Driving Solution Evolution (2): Mapless NZP + D2D
- ZEEKR's End-to-end System: Two-model Solution
- ZEEKR Officially Released End-to-End Plus: Digital Foresight Network Based on Multi-Modal Large Language Models
- ZEEKR's End-to-end Plus
- Iteration of ZEEKR's Autonomous Driving System
- Application of Geely's Autonomous Driving Foundation Model Technology
- Galaxy Adopts a Layered Deployment Strategy, and Some Vehicle Models Feature Mapless Urban NOA.
- Lynk & Co's Autonomous Driving Solution Evolution (1)
- Lynk & Co's Autonomous Driving Solution Evolution (2)
- Autonomous Vehicle Strategic Ecological Partners
- 3.7 Changan Automobile
- Vehicle Matrix
- Autonomous Driving Layout: Beidou Dubhe 1.0
- Autonomous Driving Layout: Beidou Dubhe 2.0
- Autonomous Driving Technology Strategy
- SDA
- Three Major New Energy Brands: "In-Depth Independent R&D + Development Cooperation" Autonomous Driving Mode (Function Progress/Time Nodes)
- Four Generations of Self-Developed Autonomous Driving Platforms: Nevo Q05/E07
- Nevo: Full-dimensional comparison of hardware configurations of typical vehicle models with highway/urban NOA
- Nevo is equipped with an end-to-end system: BEV+LLM+GoT
- Deepal's Autonomous Driving System
- Iteration of Deepal's Autonomous Driving System
- Future Evolution Plan of Deepal's Autonomous Driving System
- Evolution of Avatr's Autonomous Driving Functions
- Ecological Partners in the Field of Autonomous Driving
- 3.8 Great Wall Motor
- Operation
- Product Matrix
- Positioning and Planning of Autonomous Driving for Five Major Brands
- Autonomous Driving Computing Power Deployment and Prospect
- Development History and Planning of Coffee Intelligence System
- Autonomous Driving Foundation Models: SEE End-to-End Foundation Models (1)
- Autonomous Driving Foundation Models: SEE End-to-End Foundation Models (2)
- Autonomous Driving Foundation Models: VLA Models of DeepRoute.ai
- Application Progress and Planning of Urban NOA
- 3.9 Leapmotor
- Profile
- Independent R&D in Autonomous Driving
- Product Matrix
- Priorities of Autonomous Driving Layout in 2025
- Autonomous Driving Layout and Progress in NOA (1)
- Autonomous Driving Layout and Progress in NOA (2)
- Hardware Configuration and ADAS Function Comparison of Typical Vehicle Models (2024 VS 2026)
- "Change" of Autonomous Driving Chips: From NVIDIA Orin-X to Qualcomm 8650
- Independent R&D in Autonomous Driving System
- Evolution of Leapmotor Pilot (1)
- Evolution of Leapmotor Pilot (2)
- Evolution Features and Trends of Leapmotor Pilot
- Evolution of LEAP
- LEAP 3.5 (1)
- LEAP 3.5 (2)
- LEAP 3.0 (1)
- LEAP 3.0 (2)
- Autonomous Driving Strategic Ecological Partners
4 Passenger Cars NOA Solutions of Domestic Suppliers
- 4.1 Huawei
- Introduction
- Business of Intelligent Automotive Solution (IAS) Business Unit (BU) (1)
- Business of Intelligent Automotive Solution (IAS) Business Unit (BU) (2)
- ADS Full Stack Solutions
- ADS2.0
- Difference between ADS 2.0 and ADS 1.0: Sensors
- Comparison between ADS 1.0 and 2.0
- Comparison between ADS 1.0 and 2.0
- Algorithm of ADS 2.0
- Progress of ADS 2.0
- ADS 2.0: Intelligent Parking
- ADS 2.0: Obstacle Recognition
- Features of ADS 2.0 (1)
- Features of ADS 2.0 (2)
- Features of ADS 2.0 (3)
- Features of ADS 2.0 (4)
- ADS 3.0 (1): End-to-end
- ADS 3.0 (2): End-to-end
- ADS 3.0 (2): End-to-end
- ADS 3.0 (3): ASD 3.0 VS. ASD 2.0
- Application Cases of ADS 3.0 (1): STELATO S9
- Application Cases of ADS 3.0 (2): LUXEED R7
- ADS 4.0 (1): WEWA
- ADS 4.0 (2): Sensor Hardware Configuration
- ADS 4.0 (3): ASD 4.0 VS. ASD 3.0
- ADS 4.0 (4): Comparison of Autonomous Driving Versions
- Autonomous Driving Route (ADS 1.0->ADS 4.0)
- Production Models and Mass Production Customers
- 4.2 QCraft
- Introduction
- Development Vision: AGI will be realized ultimately, currently it is in the accelerating process from the second stage to the third stage
- Moving from a technology research and development enterprise to a comprehensive autonomous driving company with "technology + scale + globalization"
- Firmly adhering to the "parallel development strategy" of L2++ and L4 for application and iteration of autonomous driving
- "Dual Engine Strategy"
- Logic of "Dual Engine Strategy": from technology closed loop to global layout
- Product Matrix in the Field of Autonomous Driving
- Product Matrix in the Field of Autonomous Driving (1): Robobus Layout
- Product Matrix in the Field of Autonomous Driving (2): Status Quo of Large-Scale Deployment of Robobuses
- Product Matrix in the Field of Autonomous Driving (2): L4 Autonomous Delivery Logistics Vehicles
- L4 Robovan Layout Logic (1)
- L4 Robovan Layout Logic (2)
- Product Matrix in the Field of Autonomous Driving (3): "Driven-by-QCraft 2.0" with a three-level product matrix
- Mass Production of Single-J6M End-to-End Urban NOA Solution
- Core Technology of Single-J6M End-to-End Urban NOA: Explainable One-model End-to-End
- Core Technology of Ultimate Urban NOA Experience: VLA and World Model Architecture
- Autonomous Driving Solution Evolution
- Data and Model Training Closed Loop
- Ecological Partners (1)
- Ecological Partners (2)
- 4.3 SenseAuto
- Strategic Priorities in the Field of Autonomous Driving in 2025
- Autonomous Driving Layout
- "Cockpit-driving-cloud" Trinity Layout
- Autonomous Driving Solution
- Comparison of Autonomous Driving Solutions of Five Major Autonomous Driving Suppliers Based on Horizon J6M
- Comparison of SOP of Six Major Autonomous Driving Suppliers Based on Horizon J6M
- Technology Roadmap (1): End-to-End Autonomous Driving Evolution
- Technology Roadmap (2): Generative R-UniAD (1)
- Technology Roadmap (3): Generative R-UniAD (2)
- Technology Roadmap (4): Practical Demonstration of R-UniAD (complex scenario mining, 4D simulation recurrence, reinforcement learning, generalization verification)
- World Model: Generation of Virtual Training Data
- Parameter Performance of InfinityDrive
- Pipeline of InfinityDrive
- DiT Architecture and FID/FV for Video Generation Evaluation
- Release of "WorldSim-Drive"
- SenseWorld and Success Cases
- Strength Comparison of World Model Vendors in 2025
- Intelligent Simulation Toolchain, Technology Innovation and Cost Reduction
- In 2025, the "reinforcement learning + world model" deep integration model was introduced to open a new paradigm of end-to-end autonomous driving
- Mass Production
- 4.4 Desay SV
- Profile
- Core Business Layout
- Strategic Priorities in 2025
- Operation in 2024
- R&D Investment and Key R&D Projects in 2024
- Supply Chain Distribution and Core Clients in 2024
- Autonomous Driving Layout
- Autonomous Driving Sensor Layout
- Front View Camera: 17MP
- Forward Radar: FRD02/03
- Corner Radar: CRD03P, CRD03H
- Autonomous Driving Domain Controller (1): Domain Controller Evolution
- Autonomous Driving Domain Controller (2)
- Launch of Next-generation Central Computing Platform: ICPS02H
- Deep Cooperation with NVIDIA and Qualcomm in "Chip Platform + Intelligent Solution" Collaborative Model
- Cockpit-driving Integration Solution: ICPS01E
- Autonomous Driving Cooperation Model
- Smart Solution
- Major Customers and Ecological Partners
- 4.5 Jingwei Hirain
- Introduction
- Business Model
- Strategic Priorities in 2025
- Overseas Strategy
- Operation in 2024
- Progress of the Top Ten Research Projects in 2024
- Core Business: Electronic Products for Autonomous Driving
- Progress of Key Core Technologies in the Field of Autonomous Driving in 2024
- Autonomous Driving Layout (1)
- Autonomous Driving Layout (2)
- Main sensors
- Radar Layout
- Parameters of 4D Radar
- Launch of Long-range Imaging Radar: LRR615
- LiDAR layout
- High-precision Positioning Module
- IFC Evolution
- Overview of Driving-Parking Integrated Products
- Driving-Parking Integrated Controller: ADCU (1)
- Driving-Parking Integrated Controller: ADCU (2): ADCU II
- High-end Autonomous Driving Domain Controller: New Product
- High Performance Computing (HPC) Platform
- Central Computing Platform (CCP)
- Customers and Ecological Partners
- 4.6 Freetech
- Profile (1)
- Profile (2)
- Product Matrix
- Operation in 2022~2024
- Autonomous Driving Solution (1)
- Autonomous Driving Solution (2)
- Autonomous Driving Solution (3)
- ODIN: ODIN 1.0->ODIN 3.0
- Domain Controller Solution and Evolution
- Domain Controller Solution 1: ADC15, ADC1x
- Domain Controller Solution 2: ADC20
- Domain Controller Solution 2: Architecture of ADC20
- Domain Controller Solution 3: ADC25-J "Smart Sharing Universal Edition"
- Domain Controller Solution 3: Architecture of ADC25
- Domain Controller Solution 4: ADC30
- Domain Controller Solution 4: Architecture of ADC30
- Domain Controller Solution 5: Architecture of the Next-generation ADC-X
- Perception Solution Configuration
- Camera: Front View Camera for Passenger Cars
- Camera: FVC3
- Camera Module (1)
- Camera Module (2)
- Radar
- CMS, DMS
- Recent Dynamics and Mass Production Designation
- 4.7 Zhuoyu Technology
- Profile
- Chengxing Platform
- Chengxing Platform-based Autonomous Driving Solution Evolution
- End-to-end World Model
- GenDrive: Generative Autonomous Driving Solution Based on End-to-end World model
- Two-Model End-to-end Parsing
- One-Model Explainable End-to-end Parsing
- VLA Model and L3/L4 Autonomous Driving Planning
- End-to-end Mass Production Customers
- Cooperative Production Model
- 4.8 Momenta
- Profile
- Autonomous Driving Strategy
- Autonomous Driving Technology Evolution
- L4 Robotaxi Layout
- Autonomous Driving Solution (1)
- Autonomous Driving Solution (2)
- Algorithm Development Path
- End-to-end Solution (1)
- End-to-end Solution (2)
- Core Algorithm (1)
- Core Algorithm (2)
- 4.9 Yihang.AI
- Profile
- Strategic Layout in 2025
- NOA SOP
- Product Matrix
- Duxing Autonomous Driving Platform Three Versions of Urban Autonomous Driving
- 4.10 NavInfo
- Introduction
- Four Main Business Lines (1)
- Four Main Business Lines (2)
- New Product Technology Layout amid the Five Major Trends in the Industry
- Strategic Layout in Autonomous Driving in 2025
- Operation in 2024
- Autonomous Driving Product Matrix
- Product Planning
- Driving-parking Integration Solution (1)
- Driving-parking Integration Solution (2)
- New Driving-parking Integration Product: Based on SA8620/SA8650
- Cockpit-parking Integration Product: SA8255
- Cockpit-driving-parking Integration Product: AC8025+J3
- Cockpit-driving Integrated Domain Controller Solution
- Application Service Capabilities
- Ecological Partners
- Major Customers and Partners
- Production Model
- 4.11 Horizon Robotics
- Profile (1)
- Profile (2)
- Key Strategic Layout in the Next Three Years (1): Chips
- Key Strategic Layout in the Next Three Years (2): Autonomous Driving
- Autonomous Driving Team Organizational Structure Adjustment: Focus on High-level Autonomous Driving R&D and Optimize Resource Allocation
- Autonomous Driving Solution (1)
- Autonomous Driving Solution (2)
- Autonomous Driving Chip Series
- J6 Has Become the Choice of Many OEMs
- Autonomous Driving Solutions Based on J6M
- Release of Fourth-Generation BPU Architecture
- Introduction of HSD Together
- Ecological partners
- 4.12 Neusoft Reach
- Profile
- Autonomous Driving Product Matrix
- Front View Smart Camera (1): X- Cube3.0
- Front View Smart Camera (2): X- Cube4.0
- X-Cube Evolution
- ADAS Domain Controller: M-box
- Driving-parking Integrated Domain Controller (1): X-BOX 3.0 & X-BOX 4.0
- Driving-parking Integrated Domain Controller (2): X-BOX 3.0 & X-BOX 4.0
- Driving-parking Integrated Domain Controller (2): X-BOX 5.0
- X-BOX Evolution (1)
- X-BOX Evolution (2)
- End-to-end Foundation Model (1)
- End-to-end Foundation Model (2)
- End-to-end Foundation Model Helps Upgrade the Autonomous Driving Product Matrix (1)
- End-to-end Foundation Model Helps Upgrade the Autonomous Driving Product Matrix (2)
- Central Computing Platform (1)
- Central Computing Platform (2)
- SOA
- Basic Software: NeuSAR (1)
- Basic Software: NeuSAR (2)
- Autonomous Driving Partners and Mass Production Customers
- 4.13 iMotion
- Introduction
- Strategic Layout and Planning
- Innovative Business Exploration: Robot Arena
- Operation in 2024
- Business Model
- Product Matrix
- Front View All-in-one: iFC series
- Domain Controller (1): SuperVison
- Domain Controller (2): iDC100/iDC300/iDC500
- Domain Controller (3): IDC 510Pro/IDC 510 Based on Horizon J6
- Autonomous Driving Solution
- Autonomous Driving Technology Evolution
- Autonomous Driving Algorithm and Software (1)
- Autonomous Driving Algorithm and Software (2)
- Autonomous Driving Algorithm and Software (3)
- Autonomous Driving Algorithm and Software (4)
- Production Models and Major Partners
- 4.14 DeepRoute.ai
- Product Layout and Strategic Deployment
- Autonomous Driving Evolution
- Autonomous Driving Solutions (1): Driving-parking Integration
- Autonomous Driving Solutions (2): DeepRoute IO Based on End-to-End Model 1.0
- Autonomous Driving Solutions (3): Equal access to autonomous driving accelerates the popularization of autonomous driving
- Frontier Technology Layout
- Launch of VLA Model
- Architecture of VLA
- Production model
- 4.15 Qianli Technology
- Profile
- Autonomous Driving Solution and Roadmap
- Positioning and Strategy
- 4.16 PhiGent Robotics
- Profile
- Positioning in Robot Industry Chain
- Core Technology
- Binocular Stereo Vision Technology
- Autonomous Driving Solution
- PhiGo Pro Autonomous Driving Solution Based on Journey 6
- PhiGo Pro Autonomous Driving Solution: Based on J6E
- End-to-end Foundation Models (1): General AI paradigm based on end-to-end and generative foundation models
- End-to-end Foundation Models (2): Designated projects based on progressive end-to-end
- End-to-end Foundation Models (3): Reinforcement learning link + cloud world model training
- Phigo Pro (single-J5 version) (1)
- Phigo Pro (single-J5 version) (2)
- PhiVision 1.0 (1)
- PhiVision 1.0 (2)
- PhiVision 2.0
- PhiMotion1.0
- PhiMotion2.0
- Dynamics and Partners
5 Passenger Car NOA Solutions of Foreign Suppliers
- 5.1 Bosch
- Profile
- Operation in 2024
- Main Financial Data (2019-2024)
- Layout in China
- Organizational Architecture Adjustment: Smart Mobility Group
- Strategic Layout in China (1)
- Strategic Layout in China (2)
- Strategic Layout in China (3): Cross-Domain Computing Solutions Division (XC)
- Strategic Layout in China (4): High-level Autonomous Driving Solutions
- Strategic Layout in China (5): Accelerate Cockpit-driving Integration Layout
- Multi-domain Integrated Computing Layout: Development Route of Cockpit-driving Integration
- Strategic Layout in China (6): Cross-domain Integration Platforms Have Become a Development Trend, and Bosch Has Launched Cockpit-driving Integration Solutions
- Assisted Driving Product Matrix (1)
- Assisted Driving Product Matrix (2)
- Autonomous Driving Algorithm Evolution Planning
- DiffVLA
- VLP
- Based on the End-to-end Development Trend, Bosch Autonomous Driving initiates the Organizational Structure Reform
- Assisted Driving Partners
- Mass Production of End-to-end Solutions
- 5.2 Mobileye
- Profile and Product Portfolio
- Core Autonomous Driving Technology: REM (Road Experience Management)
- Core Autonomous Driving Technology: RSS (Responsibility-Sensitive Safety)
- Core Autonomous Driving Technology: True Redundancy (TR)
- From the perspective of product orientation for consumers, autonomous driving is redefined as L4
- In 2025, L3 autonomous driving chips would be mass-produced, promoting the large-scale application of high-level autonomous driving technology
- EyeQ6H vs. EyeQ5H (1)
- EyeQ6H vs. EyeQ5H (2)
- L1-L4 Full-Stack Autonomous Driving Product Portfolio (1)
- L1-L4 Full-Stack Autonomous Driving Product Portfolio (2)
- Sensor Hardware Configuration of L3 Autonomous Driving Products (Highway/Urban/Rural, etc.)
- Comparison of L2/L3 Autonomous Driving Products in Mass Production Customers, Production Vehicle Models and Target Markets
- Chauffeur Algorithm Architecture
- Product Evolution Planning: CH (L3) products will account for 10% of the company's total revenue in the second phase
- Cooperation with Volkswagen in Releasing the First L4 Autonomous Vehicle (1)
- Cooperation with Volkswagen in Releasing the First L4 Autonomous Vehicle (2)
