At Auto China 2026 autonomous driving technology and robotaxis have taken the center stage with multiple announcements showcasing the leaps made in artificial intelligence, centralized computing and the growing ecosystem.
Purpose built robotaxis by OEMs like Geely are redefining the future where robotaxis are set to become a vehicle category of their own and one that promises profitability for OEMs.
The narrative in China is shifting from L2+ to Level 3 autonomy, where the driver can truly enjoy hands-free driving with OEMs like Chery and suppliers like Bosch spearheading their technology readiness.
Significant progress is being made in improving L2+ capabilities and experience by OEMs like BMW, Volkswagen and suppliers like Momenta, Valeo, QCraft, and ecosystem players like HERE.
At Auto China 2026, also known as the Beijing International Automotive Exhibition, autonomous driving technology and robotaxis took the center stage, having evolved both in ambition and deployment readiness. Companies made several announcements, including product displays and demonstrations from global automotive OEMs, Tier 1 suppliers and technology firms. The event showcased leaps made in artificial intelligence, centralized computing, and the growing ecosystem across the autonomous driving competitive landscape. Counterpoint Research was on the ground in Beijing covering key announcements at the show — the latest trends in automotive technology that are redefining smart mobility.
Global penetration of advanced driver assistance systems (ADAS) and autonomous vehicles is expected to grow to 94% by 2035 from 66% in 2025 with autonomous vehicles defined by the Society of Automotive Engineers (SAE) as Level 3 and Level 4 vehicles expected to make up 20% of vehicle sales in 2035 per Counterpoint Research’s Global ADAS and Autonomous Vehicle Forecast.
Global ADAS and Autonomous Vehicle Sales Forecast by SAE Level
Source: Counterpoint Research Global ADAS and Autonomous Vehicle Forecast 2025
Further, the global market for sales of robotaxis is projected to reach $54 billion in value by 2035, with annual robotaxi vehicle sales expected to reach 1.3 million units in 2035, according to Counterpoint Research’s new Global Robotaxi Vehicle Sales and Services Market Forecast.
Global Robotaxi Vehicle Sales Volume and Market Size Forecast
Source: Counterpoint Research
The forecasts demonstrate the increasing importance and high growth potential of autonomous vehicles for private passenger vehicles and robotaxi use. In this blog we capture and highlight the key trends in smart mobility and autonomous driving that echoed at Auto China 2026.
OEMs and Mobility Players:
Geely’s Robotaxi Prototype: Geely captured significant attention with the unveiling of China's first purpose-built robotaxi prototype, the Eva Cab, which supports face-to-face seating. Developed as a purpose-built platform rather than a converted passenger car, the vehicle does not have conventional controls like a steering wheel or pedals and has full redundancy. it uses Geely’s AI 2.0” E/E architecture with more than 3,000 TOPS compute (NVIDIA Superchip + NVIDIA Thor U + Qualcomm Snapdragon 8397), a high-resolution 2160-line LiDAR (600 m range) and Geely’s Qianli Haohan’s G-ASD L4 autonomous driving system. Geely has indicated a plan to start building Eva Cabs in 2027 and by 2030 build 100,000 Eva Cabs for CaoCao Mobility.
Geely Eva Cab and Xpang GX are both factory -integrated level 4 vehicles
Sources: Geely, Xpeng
Xpeng Motors Robotaxi: Xpeng unveiled its GX flagship 6-seat SUV which it described as ‘robotaxi ready’. It uses four Xpeng Turing AI chips equating to approximately 3000 TOPS and Xpeng’s second generation VLA system, Xpeng VLA 2.0, that translates visual inputs directly into actions, removing the need for translation to language. As such AI can interpret and respond to the physical environment faster and more intuitively, providing more natural human-like behavior.
Chery and NVIDIA + Qualcomm partnership: Chery Auto announced a partnership with NVIDIA to show jointly developed Physical AI through automated Level 3 and Level 4 vehicles, cockpit AI, robotics and smart factories. Chery will use NVIDIA Drive Hyperion for assisted driving hardware. (Read more about how NVIDIA redefines the autonomous driving game with “Three Computers and A Five-Layer Cake”). Chery also partnered with Qualcomm using its Snapdragon Cockpit Platform Premium, Snapdragon Ride Platform Premium and Snapdragon Rise Flex SoC. For driver assistance the two companies intend to co-develop intelligent solutions based on the Snapdragon Ride Platform with several models expected to utilize this platform. Chery intends to use the Snapdragon Ride Flex SoC on multiple models delivering unified upgrades in perception, decision making and interaction across both cockpit and driving systems.
WeRide and Lenovo collaboration: Chinese robotaxi player WeRide announced a collaboration with Lenovo to accelerate large-scale commercialization of Level 4 autonomous driving worldwide. As part of this collaboration both companies announced an ambitious target to jointly deploy 200,000 autonomous vehicles including robotaxis over 5 years. They jointly launched a high-performance-computing platform, HPC 3.0, powered by NVIDIA DRIVE AGX Thor SoC delivering 2000 tops and built on Lenovo’s L4 autonomous driving domain controller, AD1. WeRide and Lenovo claim a 50% lower total cost for the autonomous driving suite over HPC 2.0.
BMW Neue Klasse models: BMW premiered its Neue Klasse platform for China with the ix3 and i3 long wheelbase and updated its 7 series models and get the OEM’s new L2+ system with implementation expected in a further 12 models by end-2027. For China, BMW has partnered with Momenta to implement an ”address-to-address” driver assistance system which uses a locally trained end-to-end AI model for Chinese traffic. The system provides hands-free Level 2+ motorway driving up to 130 km/h and a city assistant for guided urban segments.
BMW and Volkswagen Group showcased their latest vehicles with L2+ advanced driver assistance capabilities
Sources: BMW, Volkswagen Group
Volkswagen Group: The OEM debuted four global vehicles including the VW ID. AURA T6, VW ID. UNYX 09 (co-developed with Xpeng), Jetta X, and Audi E7X. VW, through its VW China and Horizon Robotics JV CARIZON, is working on launching the first models equipped with its full L2 Advanced ADAS technology, including Navigation on Autopilot (NOA) in both urban and highway environments, plus automated parking. AUDI also announced it will make Level 3 autonomy capabilities available on the new AUDI E7X model. The Level 3 system will be powered by Momenta.
Tier 1 Suppliers and Technology Players
Bosch L2+ and L3 demonstration: The company showcased a new one-stage end-to-end L2+ architecture and an L3 highway pilot demo. The L3 system for highways up to 120 km/h uses dual redundant computing and actuators, enabling hands-off driving with fail-operational safety. Bosch said it has obtained an L3 test license in Wuxi, China. Bosch’s system using an advanced version already applied system on the Chery Exeed ES vehicle, the vehicle can now brake, steer and accelerate autonomously.
Bosch and QCraft pushing the boundaries on Level 3 and Level 2+ automated driving
Sources: Bosch, QCraft
Valeo L2+ Highway NOA: Valeo demonstrated an upgraded Valeo Smart Safety 360 VSS360 platform adding Highway Navigation-on-Autopilot (NOA). This L2+ system uses cameras, radars and their new VISONIC LiDAR to support hands-free highway driving and lane merging. It has been in series production for a major Chinese OEM since Oct 2025.
QCraft L2++ City NOA: The company unveiled its “Physical AI” model and its new QPilot MAX city NOA solution which delivers >500 TOPS on a unified architecture (cloud + vehicle) and claims an ultra-low AEB (autonomous emergency breaking) false-positive rate. It is already deployed in 25 vehicle models at L2++ (city NOA) with deployment on 50 more models scheduled in 2026. The hardware uses a world-model + reinforcement-learning framework and has a 500+ TOPS ECU using 11 cameras and 1 LiDAR (no radars) configuration. QCraft also previewed L4 robotaxi and “Robovan” projects.
HERE Technologies Live maps and global NOA: HERE unveiled an AI-powered live road model for consistent L2++ NOA across international markets by orchestrating navigation, automated driving functions and infrastructure awareness through a single, AI-powered road model that continuously supports both human navigation and machine planning. It also announced an MoU with Baidu Maps to co-develop intelligent driving maps, and an MoU with Horizon’s neueHCT to bring production-ready ADAS globally with full‑scenario NOA expansion for global automakers, accelerating the progression from L2+ capabilities toward higher levels of automation, including L4 and robotaxi use cases across international markets such as Europe and the Gulf region.
HERE signing of MoUs for intelligent maps to support ADAS and Autonomy and Hesai’s 6D LiDAR launch are examples of ecosystem pushing boundaries of autonomy technology
Sources: HERE, Hesai
Hesai Imaging LiDAR: Hesai is the world leader in LiDARs that are widely used across automakers. The company launched the world’s first 6D full-colour LiDAR sensing chip, the Picasso SPAD-SoC. Unlike conventional LiDAR systems that capture only geometric information such as width, height, and depth, Picasso integrates RGB sensing and time-of-flight (ToF) ranging at the chip-level, enabling the direct generation of colorized point clouds without the need for complex sensor fusion. Hesai also launched the ETX platform built on the Picasso architecture, the world’s first 6D full-colour ultra-thousand-line LiDAR platform. It will enter mass production in the second half of 2026, with widespread deployment across multiple global flagship models starting in 2027. The company has already secured contracts from several top-tier automakers, positioning the platform as the “perception core” for L3 and higher autonomous driving systems.
Conclusion
The trajectory is clear that increasingly China is setting the agenda when it comes to advancing the autonomy narrative. Be it with L2+ (and variants of it) over the past couple of years to now pushing the boundaries not just on regulation but an ecosystem shift toward Level 3 and Level 4 autonomy so that not only is the technology maturing, but real world implementation in vehicles is now very much set to happen in the near term.
Analyst Take:
Three key trends are apparent from some of the key highlights on Autonomous Driving at Auto China:
Purpose-built robotaxis have started to emerge. While factory-fitted Level 4 technology on vehicles is already happening, we are seeing companies like Geely going a step further and redefining the future of robotaxi designs through the Eva Cab concept. From a revolution in vehicle seating to no in-vehicle controls like steering and pedals, and the use of not just superior autonomous driving technology but also redundancy, the robotaxi of tomorrow is set to become a vehicle category of its own and one that promises profitability for OEMs.
The narrative in China is shifting from L2+ (and variants of it) to Level 3 autonomy, where the driver can truly enjoy hands-free driving. Not only are OEMs like Chery building collaborations with technology players like NVIDIA and Qualcomm, but they are also doing so with traditional Tier 1 suppliers like Bosch who are also spearheading their technology readiness and getting partnerships in place to make Level 3 the next big growth story in autonomy in China.
Lastly, OEMs like BMW, VW Group and suppliers like Momenta, Valeo, QCraft, and ecosystem players like HERE continue to push the boundaries by improving L2+ (and variants of it) technology capabilities and experience. The improvements made to L2+ hardware and software are being made to the immediate benefit of Chinese vehicles, but the learning and improvements are not going to be restricted to China alone. We will see implementation of L2+ in Europe, including on exported Chinese vehicles (sales of which are growing at a fast pace and already have 10% market share in Europe). Players like Momenta and HERE are enabling this through their autonomous AI models, AI-powered Live maps and intelligent mapping for NOA.
Receive our insightful weekly newsletter and stay ahead of the competition.
Author
Murtuza Ali
Murtuza is a Senior Analyst at Counterpoint Research based out of the UK. In Counterpoint, he closely tracks the Automotive Industry and Markets with a focus on pivotal technologies such as Electric Vehicles, Autonomous Vehicles, Software Defined Vehicle, Infotainment & Digital Cockpit, Mobility and Connectivity. He started his career at Tata Motors developing Electric Vehicles graduating into Strategy roles. His most recent experience prior to joining Counterpoint Research has been as a Consulting Manager at the Transport & Mobility consultants Ricardo UK. He holds an Executive MBA from Warwick University, MSc in Automotive Systems Engineering from Loughborough University and a BEng. in Automobile Engineering from Mumbai University.