Volkswagen recently announced the launch of the L4 (Level 4 autonomy with no human driver in designated locations, times and driving conditions) ID.Buzz AD microbus. The business model is to work with existing transportation and mobility companies (private and public), by providing the hardware (vehicle, sensors, compute) and autonomous driving software stack (including fleet management software). The first set of 500 vehicles is slated for delivery in 2026 to Uber for ride-hailing trials (with a human driver) in Los Angeles, Fully autonomous ride hailing is planned for 2027. The ID.Buzz AD is currently offering ride share services In Hamburg, Germany, through MOIA, a “technology company of the Volkswagen Group which develops on-demand ride-pooling services to redefine mobility for people in urban areas”.
Volkswagen has attempted to enter the autonomous mobility market since 2020 with ambitious plans and $Bs in investments – both, internal (the CARIAD software division), and external (co-investment in Argo.ai). Unfortunately, these were delayed or dissolved, led to board tensions and senior executive changes, and slowed the effort. This is similar to the experience at other OEMs like Ford and GM Cruise. Volkswagen is re-entering the space at a time when technology giants (Waymo-Alphabet, Zoox-Amazon and Tesla) are already significantly ahead. Even the Sam Altman-led OpenAI, (the company that started off as a non-profit venture and now also has a for-profit division) announced recently on a podcast that “We have some new technology that could just do self-driving for standard cars” (see minute 6 of the podcast).
So can a century old automaker (and other automakers) prevail in this market, dominated by technology companies? In Volkswagen’s case:
- The business model could help – it intends to partner with existing mobility providers who would use their platform to offer services (like Uber which does this today with human drivers and their cars).
- The hardware stack is much lower in cost since most components (vehicle, sensors and compute platforms) are also used in passenger cars that offer lower levels of autonomy (L2 and L3). The passenger car volume is significantly higher than L4 ride-hailing cars, enabling scale and lower costs.
- The manufacturing experience and supply chain for sensors and other hardware is a plus.
- Finally, sometimes its OK to come late to the party. 3D imaging sensors like LiDAR and imaging radar have improved dramatically and are more optimized today in cost, performance and footprint. These include new generations of LiDAR from Innoviz and Valeo, and imaging radar from Mobileye.
Valeo
As a global, publicly listed (~$25B/year in revenues), Tier 1 supplier to automotive OEMs, Valeo supplies a wide range of products in electrification, driving assistance systems, interior cabin experience and lighting groups. Products include mechanical, electrical, sensing and electronic components, spanning from motors and lighting assemblies to ultrasonic, camera, radar and LiDAR (Light Detection and Ranging). The SCALA™ series of LiDARs is one of these products, with the second generation (SCALA 2) version designed into the Mercedes S-Class (see Figure 1) to support its L3 autonomy (self driving under certain conditions with human driver ready to take over in 10 seconds) feature (Drive Pilot) at a maximum speed of 95 km/h (60 mph). SCALA 2 achieves a 80 m range for brick and tire debris size objects and 300,000 pps (points per second, a metric akin to resolution or pixels in a camera) in a ~ 600 cm³ volume with a square optical format.
Valeo’s next-generation version, the SCALA 3, uses a similar platform as the SCALA 2, but delivers significantly higher range (200m) and resolution (12.5M pps) performance in a ~1200 cm³ volume, with a 45 mm height and rectangular format, suitable for roof or behind-the-grill mounting, see Figure 2).
The higher performance of the SCALA 3 is expected to increase the allowable L3 speeds for the Drive Pilot to 130 km/h (80 mph). SCALA 3 is expected to start production in 2025, and has been selected by Stellantis (launch of Stella AutoDrive for L3 at initial speeds of 37 mph in 2026). It has also been selected by 3 other global OEMs to support L3 and L4 autonomy features. Per Antoine Claudepierre, V.P of LiDAR Business Development, one of these is a leading Asia-based automotive OEM (to start production in 2026 for L3 vehicle). Valeo has previously supplied LiDAR for Honda’s L3 efforts.
On the imaging radar front, Valeo announced a design win with a leading European OEM, with production slated for 2028. The goal is to increase allowable speeds for unsupervised Highway Pilot to 130 km/h. Apart from the imaging radar, the sensor suite will also use a LiDAR (supplier evaluation ongoing). Given the discussion above, its seems likely that this is for the Mercedes Drive Pilot.
As a Tier 1 supplier, Valeo has collaborated with Mobileye since 2015 on front-facing cameras and other driver assist solutions. In September 2023, the companies announced a collaboration to industrialize imaging radar for autonomy where “automakers gain access to the latest cutting edge technology from Mobileye that they can trust will exceed industry expectations as we have proven before, while benefiting from the customization, industrialization, testing and support capabilities brought by Valeo”. The two companies have also collaborated on delivering ADAS features to OEMs like Volkswagen, with Valeo as the Tier 1 and Mobileye as the Tier 2 supplier. Mobileye has been developing its imaging radar (Mobileye Imaging Radar™) since 2018 and separately announced a design win with a major OEM for L3 capability (at or > 130 km/hr) to launch in 2028 (Mercedes again??). More on this later in the Mobileye section.
Innoviz Technologies
Innoviz has been chosen as the LIDAR supplier for L4 autonomous driving in the Volkswagen ID.Buzz microbus. With launch of this capability planned for 2026 in select cities like Los Angeles and Hamburg, Innvoiz is currently ramping up production of its InnovizTwo short and long range LiDARs. Each ID.Buzz will use a total of nine InnovizTwo LiDARs (6 short and 3 long range).
Innoviz uses 905 nm SPAD (Single Photon Avalanche Diodes) based detectors and edge emitting laser arrays, along with MEMs (Micro Electro Mechanical) mirror scanners in the InnovizTwo LiDAR family. The InnovizTwo short range LiDAR has a range of 100 m and a high vertical Field of View (FoV) of up to 90°. The long range version has a range of 300 m, a 120°x43° maximum FOV, minimum angular resolution of 0.05°x0.05°, 20 Hz frame rate and a physical volume of 825 cm³. Figure 3 shows the InnovizTwo, and point clouds at short and long range:
InnovizTwo is also resilient to optical path obscurations (mud, rain, insect splatter), a critical requirement for autonomous operation. Per Mr. Kielaf, having such capabilities in an L4 vehicle (no human driver) is an absolute requirement since loss of perception at high speeds can be lethal. The optical design of the InnovizTwo LiDAR ensures redundancy in the optical path for the transmit and receive path for each pixel. As an example if the window is 25% covered, no single pixel will be “dead” and some pixels will lose about 7% of the range performance.
Innoviz was selected by Audi (a Volkswagen brand) to deliver LiDAR as a Tier 1 supplier in 2023. Achieving this designation has many benefits – direct relationships with the OEM, transparency and integration smoothness, visibility into other program opportunities, direct performance and product feedback, higher profit margins and credibility for other OEMs. But it has its own challenges as well – intensive audits on multiple fronts – manufacturing, costs, quality and production strategy.
Innoviz has a manufacturing facility in Israel. However, it is not automotive-certified, and achieving this takes time and investments. Innoviz announced a partnership with Fabrinet, a publicly listed contract manufacturer with automotive grade certification, and a specialist in high volume manufacturing of precision opto-mechanical-electrical products. Its factories in Thailand can scale manufacturing to automotive volumes and quality as demand grows. This positions Innoviz to deliver as a Tier 1 supplier to the Volkswagen ID.Buzz L4 program as it scales up.
Apart from the Volkswagen ID.Buzz and Audi selection, Innoviz is also the preferred LiDAR supplier to Mobileye for their different autonomy applications. The company also announced recently that “it has signed a Statement of Development Work (SoDW) agreement with a Top 5 passenger automotive manufacturer. Through this SoDW, Innoviz will supply advanced LiDAR units for the OEM’s L3 highly automated series-production program with target SOP of 2027”. Mercedes Again??
According to Omer Kielaf, CEO of Innoviz, the company’s success in getting designed into major OEM platforms “are our history of deep working relationships, disciplined project management, investing in becoming a Tier 1 designated supplier, the cost-performance-durability-size of the InnovizTwo LiDAR, and the ability for customers to source short and long range lidars from the same supplier”.
Mobileye
One of the biggest success stories of high tech from Israel in the last decade, Mobileye is today a Nasdaq listed company (occurred in 2022 after its spin-off from Intel which had acquired the company earlier). With a market capitalization of ~$15B and revenues of $1.8B/year, it is easily the leading player in the pure play autonomy landscape. Its customers include 50 automotive OEMs who use their ADAS-based solutions (cameras, chips, software) across 1200 car models and 190M cars. The goal is to support ADAS, L2, L3 and L4 capabilities through a combination of semiconductor chips, crowdsourced data from its mounted camera units, sensor integration (internal and external) and AI based software for path planning and drive policy. These products are listed below:
- SuperVision™ addresses L2+ systems which need constant human driver attention and control. SV52 is based on a proprietary EyeQ5 chip and launched in 2022. In 2026 the SV62 (based on the EyeQ6 chip) which delivers higher capabilities and robustness will deployed in the Porsche and Audi product lines. Mobileye is a Tier 1 supplier in this case.
- Chauffeur™ (expected commercial launch in 2027), is targeted at highway L3 systems (up to 130 km/h speeds), and fuses cameras, its own proprietary imaging radar (Mobileye Imaging Radar™) and a front-facing LiDAR. Mobileye is the Tier 1 for the Chauffer™ and the EyeQ6 chips, although the sensor stack is handled through Tier 1 suppliers like Valeo for final vehicle integration, based on a reference design and validation by Mobileye. It also provides the Mobileye Imaging Radar™ to Valeo.
- Mobileye Drive™ is a L4 system, used by robotaxi and trucking customers. As a supplier for the Volkswagen ID.Buzz robotaxi discussed above, the reference design sensor stack is composed of 13 cameras, 5 Mobileye Imaging Radar™ and 9 InnovizTwo LiDARs (3 long and 6 short range). Mobileye supplies the radar units directly to Volkswagen as a Tier 1, and also the ECU composed of 4 EYEQ6H and other chip sets. The system provides a generalized driving policy, and Mobileye’s REM™ AV mapping technology. Mobileye directly supplies the platform to OEMs in the role of a Tier 1 supplier, including its radar. Other sensor components (like the LiDAR) are supplied directly by Innoviz to Volkswagen.
Mobileye had a development program for a proprietary 1320 nm FMCW (Frequency Modulated Continuous Wave) which it presented at CES 2023. In late 2024, it discontinued this effort claiming that this development was less essential to its future efforts, and “that the availability of next-generation FMCW lidar is less essential to our roadmap for eyes-off systems”. It cited as reasons the progress on its EyeQ6-based computer vision perception, performance of its imaging radar, and better-than-expected cost reductions in third-party time-of-flight (ToF) LiDAR units.
The Mobileye Imaging Radar™ draws on IP during its tenure within Intel. Mobileye believes that its imaging radar is a perfect complement to cameras, especially in bad weather and will play a big role in compensating for loss of LiDAR and camera sensitivity. The core of the radar is the proprietary design of radio-frequency integrated circuits (RFIC), embedded in a unique architecture where the entire radar signal is sampled and digitally processed by a dedicated proprietary processor. It has 1500 channels (virtual antennas), 20 Hz frame rate, 100 dB dynamic range and an angular resolution of 0.5°. The resolution is twice that of other imaging radars on the market enabling small object hazard detection, segmentation and identification at 130 m range (at speeds of 130 km/hr, presumably for L3 or L4, this is an OEM requirement). Larger objects like pedestrians and traffic poles can be detected at a 300 m range. Costs are expected to be in the $100 range. At short range, a combination of these radars can address a 360° FoV.
In essence, this radar delivers essentially the same type of 3D point cloud and axial velocity information as a FMCW LiDAR, but at 100X lower angular resolution. It can detect most objects and their axial velocities, but not identify or classify them. The big advantage of course, is that performance is not degraded in bad weather and it costs 5-10X lower than a LiDAR. It is also more compact and can be gracefully designed into the vehicle body. Figure 4 shows a point cloud generated by the imaging radar.
According to Dan Galves, Chief Communications Officer, “Our imaging radars were designed as a key technology pillar to enable high-precision hands-off, eyes-off driving in both robotaxis and privately-owned vehicles, with a cost that supports scale. By adding a perception layer based on imaging radars, fully independent of cameras and LiDAR, Mobileye delivers a scalable, safe and cost-efficient solution to support the global deployments of consumer AVs and robotaxis – no matter where they drive.”
Automotive OEMs are entering the autonomous car revolution once again. The challenge for them is not the technology – that exists. It is to find business models that unleash meaningful revenues and profits demanded by their shareholders (unlike the tech giants who have massive profits on their core business, and autonomous driving is just another BET). On the L3 front, it is a feature they can sell to consumers by the millions, and at reasonable cost and profits. Leveraging this into L4 capability for autonomous ride-hailing services will require competing aggressively with technology giants already deep in the game. It will be interesting to see how this war unfolds.
