Renesas and Nullmax Partnership
Renesas and Nulllmax announced integration of Nullmax’s AI perception stack on Renesas’ R-Card V4 and X5 series chips. Nullmax has focused on a vision only end-end AI approach for autonomous driving. The Nullmax CEO came out of Tesla’s autopilot team and has worked with at least one of China’s top OEMs. The Renesas X5 SoC is based on 3nm chiplets with up to 32 ARM cores and 400TOPs of AI performance and can compete with the lower end of the Qualcomm and Nvidia offerings today. Honda announced a 2000 TOPS development with Renesas at CES.
Technavio forecasts automotive technology market will grow $263B between 2025 and 2029
The growth is driven by ADAS and AV adoption, including technologies such as compute, displays and sensors. ADAS penetration in China already exceeded 35% in 2024 and is projected to reach 95% by 2030.
Lyft to roll out Mobileye powered robotaxis by 2026
Lyft announced that it is working with Mobileye to launch robotaxis in Dallas in 2026. Marubeni, a Japanese conglomerate with experience managing fleets, will own and finance the Mobileye-equipped vehicles that will show up on the Lyft app. This is another good win for Mobileye and a key milestone in the race to roll out robotaxi services across the US.
WeRide gets approval to launch Robotaxi service in Beijing
WeRide received approval to launch its GXR paid driverless ride-haling services in Bejing and is WeRide’s second commercial launch, following Abu Dhabi with Uber last December. The service is allowed to run on certain highways, including the expressway to the airport.
Kia launches Syros SUV in India with 18% ADAS Adoption
Kia launched its small Syros SUV in India and states that 18% of “bookings” included the ADAS package, which costs about an additional $1,000. This represents about ½ of the ADAS adoption rate in China but shows that L2 technology is also coming to other markets.
Stellantis Previews AutoDrive
Stellantis has previewed its hands-free highway driving system, named STLA AutoDrive that allows hands-off driving on mapped highways, similar to GM’s Super Cruise. It also allows L3 driving under certain conditions (up to 60km/h). Stellantis has still not announced which vehicles will first get this feature, but the press release pictures suggest the Jeep Wagoneer EV may be among the first. The system appears to be camera/radar based and probably using Qualcomm’s Snapdragon processors. Stellantis had acquired aiMotive in 2022 for perception and Qualcomm has an L3 software stack based on the Veoneer/Arriver work, although it’s not clear how much of that is used by Stellantis in this system. Stellantis has been a little behind GM and Ford with its L2+ system but looks to be catching up now.
Daimler Previews Level 2++ Driving Features on the new CLA
Daimler is rolling out a new set of L2++ driving features on the 2025 CLA, starting in China. The term L2++ is used for a set of enhanced urban assistance functions. The driver must still keep their hand on the wheel and monitor traffic, but the car handles steering, braking and accelerating. It helps the driver navigate busy streets in the world’s biggest cities and is tailormade for local conditions.
Cooperative steering is unique to Mercedes-Benz and means that the driver can always intuitively adjust the position of the vehicle without disengaging the system. Mercedes-Benz will be the first international OEM to offer L2++ starting in China in 2025.
Geely Launches G-Pilot ADAS Suite
Geely’s new suite of intelligent driving functions will lunch on its Galaxy line of EVs and includes NOA and Park Assist. The cars will be able to do “door-door” autopilot. The G-Pilot system is available at 5 levels of capability, starting with a 100 TOPS solution and going up to dual Nvidia Thor U chips (2000 TOPS) for a L3 capable system.
Geely has established a supercomputing center with 10,000 GPUs and has developed a proprietary LLM, which has since been integrated with DeepSeek’s model and is used for scenario comprehension and decision optimization. Geely has more than 7 million L2 equipped vehicles on the road and can collect large amounts of data from them for training.
Xiaomi EV Hyper Autonomous Driving requires User Training
Xiaomi, the Chinese mobile phone company, has announced the full rollout of its end-to-end all-scenario AD feature, which enables seamless integration from departure to automated parking at the destination. Interestingly, their press release states that users have to complete 1000km of driving before the system can be fully activated and parking lot rules have to be learned before the automated parking can be used. We have seen a number of China systems require “user training” before operating autonomously, which seems to reduce AI model sizes and hardware requirements.
Uber is partnering with Waymo in Austin TX
Uber users can now select a Waymo car in Austin TX. They are also working on rolling out this option in Atlanta. Uber is used for roughly 28M trips per day, which far exceeds Waymo’s app users, so this sort of partnership can be beneficial to both companies as robotaxis scale out. At launch there is no additional cost versus a standard Uber ride and you won’t have to tip the driver!
Wayve announces new AI Hub in Germany
Wayve has announced a new development and test hub near Stuttgart, following a $1B series C funding round. Germnay allows Wayve to train its ADAS systems on German specific driving conditions, including high-speed autobahns.
May Mobility starts driverless rides in GA
May Mobility is another US based robotaxi service, operating driverless vehicles in Arizona, Michigan and now Peachtree Corners, GA. The vans operate along pre-planned routes, between hotels, retail, offices and other city landmarks. Rides can be booked through May’s own app. May is focused on working with transit agencies, cities and similar organizations to bring AV capability to life. May is using Nvidia DRIVE hardware in its vehicles and has used lidar from companies including Ouster and Aeva along with its proprietary software.
WeRide introduces new Delivery Robovan
WeRide has introduced the Robovan W5, a Level 4 autonomous delivery vehicle designed to revolutionize logistics. This next-generation unmanned solution integrates WeRide’s self-driving technology, enabling 24/7 operations in all weather conditions. The van has a 220km range and a 360 degree perception system based on a modular sensor suite of cameras, radars and lidar and an end-end AI model. WeRide had 10,000 indication orders from an express delivery company in China for its previous generation Van based on a Transit like vehicle and it will be interesting to see how demand grows for these custom platforms.
Kodiak Robotics now commercially driving
Kodiak has been working with Atlas Energy since 2024 for off-road driverless truck testing. They recently announced that the first 100 commercial delivery loads (of proppant) have been completed with Atlas owned trucks and the Kodiak Driver self-driving system. Atlas will now be able to expand commercial driverless operation in Texas and New Mexico. This is an important step to expanding the “safety case” to highway operation. The trucks include Kodiak’s “SensorPods™”, which are pre-calibrated modular units that house all the sensors for autonomous driving, that allow for faster and easier repairs. The previous generation trucks had 18 sensors, including 10 cameras, 4 ZF Radars, two Hesai scanning LiDARs, 2 Luminar Iris LiDARS and used Ambarella CV2 SoCs for camera data processing. The 6th generation truck adds two additional cameras and radars.
Waabi Partners with Volvo on Autonomous Trucks
Waabi, a Canadian software company founded by Uber ATG’s former chief scientist, and Volvo Autonomous Solutions (VAS) have partnered to develop and deploy autonomous trucks. The partnership combines Waabi’s AI technology with Volvo’s expertise in safety and automation. Waabi’s virtual driver system, Waabi Driver, will be integrated into Volvo’s VNL Autonomous truck. The VNL Autonomous will be manufactured at Volvo’s New River Valley assembly plant in Virginia. Commercial pilots for the VNL Autonomous are scheduled to begin in Texas in 2025. There are a number of autonomous truck trials coming to Texas in 2025, including Aurora, Kodiak, Plus Torc and TxDOT and this partnership also aims to accelerate the rollout of self-driving trucks in the United States. Commercialization of these services is currently expected to start in 2027.
