Pierrick Boulay is a Senior Technology and Market Analyst in the Photonics and Sensing Division at Yole Intelligence. He works in the fields of solid-state lighting, carrying out technical, economic, and marketing analyses. In addition, he leads the automotive activities within the company.
Boulay has authored several reports and custom analyses on topics such as vehicle lighting, lidar, sensing for ADAS vehicles, and VCSELs. Prior to Yole, he developed his knowledge at a variety of companies, primarily in R&D departments on LED lighting applications.
He holds a Master of Science in Electronics degree from ESEO (Angers, France).
DVN-Lidar: Yole is a leading market research and strategy consultancy with solid understanding of the lidar market and technologies. What can you tell us about your departments?
Pierrick Boulay: Indeed, Yole Group is a consulting firm specializing in the strategic analysis of markets, the supply chain, and technological developments related to the semiconductor and adjacent industries. Benefiting from direct access to major players in the field of semiconductors and the multidisciplinary profile, including the scientific, technical, and market orientation of its analysts, the company supports its clients in understanding markets, the evolution of technology, and their commercialization.
At Yole Intelligence and Yole SystemPlus, we have been following the lidar market and related technical innovations since 2017. Since then, we have published numerous market and technology reports and developed an impressive collection of lidar teardowns. All analyses are available on our corporate website. Our group today has an in-depth understanding of the technological choices made by the leading automotive companies and the market changes that drove them.
DVN-L: Could you tell us more about the current state of the automotive lidar market? Which countries are leading?
P.B.: The lidar market is at a crossroads. Indeed, 2022 was the first year that the lidar market for passenger cars exceeded the one for robotaxis. In the beginning, only Audi had multiple car models implementing lidar. In 2021, at Yole Group, we saw more OEMs had implemented lidar, like Honda, Lucid Motors, Lexus, and Mercedes, among a few others, but it was limited to high-end cars, mostly from the E and F segments, and therefore low volume. It was only in 2022 that volume started to grow significantly with a strong push from Chinese OEMs with cars in the D segment. That means more affordable cars, more volume, and, therefore, a lidar market that is taking off. In 2022, according to the lidar 2023 report, experts estimated that the lidar market for passenger cars was valued at $169m.
In 2022, 100 per cent of the cars released were from Chinese OEMs. Since 2018, 61 per cent of the cars released with lidar were from Chinese OEMs. So, it is quite clear that China is leading the race in terms of implementation. Among the others, 25% of cars were released by European OEMs, 11 per cent from Japanese OEMs, and 3 per cent from U.S. OEMs.
DVN: What is your volume forecast for 2028?
P.B.: In 2028, the lidar market is forecasted to be $3.9bn, and will still be driven by Chinese OEMs. Our analysts are monitoring the different design wins, and between 2023 and 2025, we have counted more than 120 cars embedding lidar(s) to be released. Among them, more than 100 will be from Chinese OEMs. This clearly shows that there are currently two markets for lidar: the Chinese market and the rest of the world. They have different dynamics. In China, OEMs have already made the transition from ICE to EV cars, and this transition started 20 years ago. So, now OEMs can focus on implementing more features for ADAS and autonomy. The high number of OEMs implementing lidar leads to higher volumes and, therefore, a lower lidar price for the OEM. For the rest of the world, the transition from ICE to EV is just starting, and OEMs are focusing most of their resources on this transition. It is even more obvious in Europe, where in 2035, only battery-electric vehicles will be allowed to be sold. This is not preventing European OEMs developing automated driving features, but this is limited to high-end cars like the Mercedes S-Class or EQS, the BMW 7 series, or the Volvo EX90. When this transition is more advanced, more cars with lidars from European OEMs could be expected.
DVN-L: What are the main applications today? Do you see the applications changing in the next five years? Is Autonomous Driving expected to be the main application?
P.B.: Currently, most of the applications are related to automated driving on highways. First eyes-off applications in Germany and a few states in the United States will allow the car to drive by itself on highways up to 60 km/h. The speed limit is expected to increase to 130 km/h with the release of a new generation of lidars like the Valeo Scala 3. In China, OEMs are developing Navigate on Autopilot (NOA) features so that the car can complete lane changes and enter and exit ramp junctions, but only on highways. But, due to a lack of regulations, all these cars are still considered hands-off, and the driver must remain alert on the driving task.
In the next five years, other applications related to valet parking and possibly ADAS applications like the AEB functionality should be seen. Currently, AEB relies mainly on the forward ADAS camera and radar. This combination works well in good lighting conditions and in straight lines, but when conditions deteriorate with darkness, children, and curves, this system performs poorly. This is why the NHTSA in the United States is working on regulations to make AEB testing more stringent, with testing in complete darkness, amongst others. If this new AEB regulation is approved, this could mean the use of an additional sensor for OEMs. In this case, that could be an opportunity for lidar, but there will also be competition with thermal cameras or gated imaging cameras that are much cheaper.
DVN-L: What types of lidars will launch in the next five years? What technologies will coexist?
P.B.: Long-range lidars used to be quite bulky, so they were implemented in the front grille or on the roof. The next generation will be thinner allowing them to be positioned behind the windshield, which also solves cleaning and thermal management issues. To enable this, windshield manufacturers had to improve their product with coatings to make the glass transparent enough to be used with 905- or 1,550-nm lidars.
Short-range lidars were virtually not used by OEMs for multiple reasons: lack of a real use case, high price, and limited performance. In 2024, new short-range lidars should be seen based on a combination of VCSELs and SPADs that will enable better performance and lower cost. The use case is not totally clear yet, but they could be used for lane change assist and parking applications.
Regarding technology, time-of-flight lidars at 905 nm are expected to dominate in the next five years due to lower cost and higher availability of components. The performance of these components is still increasing. Multi-junction VCSELs allow a higher output power while retaining the small die size. On the receiver side, SPADs and SiPMs are much more sensitive than APDs, and the photon detection efficiency (PDE) is expected to reach 30-35 per cent in the near future. So, 905 nm will dominate, but 1,550 nm lidar will also be used, though in lower volumes. 1,550 nm components are much more expensive, and the number of suppliers is limited. One of the consequences is that Luminar solved this issue by acquiring their suppliers so that they are now vertically integrated, can secure their production, and can work on improving the performance of these devices.
DVN-L: What are your thoughts on FMCW over the next five years?
P.B.: The feedback from lidar players working on FMCW technology is that they expect initial volume production not before 2027-2028. FMCW is seen as a very promising technology enabling the measurement of the velocity of objects instantly in contrast to time-of-flight. It is also more robust and resistant to interference from other lidars. But, before we see FMCW lidars on the road, the cost of the laser, the optics, and the electronics for calculations will need to fall drastically. FMCW lidars are based on Silicon Photonics technology, which is still emerging. Even in 2028, Silicon Photonics revenue will be quite limited (~$600M) and driven by datacom and telecom applications. Once these applications are more mature, with more players and sufficient manufacturing capacity, we could see FMCW lidars in large volumes.
DVN-L: Who are the main lidar suppliers, as you see it?
P.B.: In 2022, the lidar market leader was Innovusion, followed by Valeo, Hesai, and RoboSense. These four players generated 89 per cent of the revenue in 2022. Innovusion leads the market due to its partnership with Nio, which has many car models using lidar. Yole Group estimated that more than 56,000 lidars from Innovusion were on the roads in 2022. As the industry is quite young and evolving rapidly, analysts expect that these rankings will change in 2023, with Hesai taking the lead due to its large number of partnerships with Chinese OEMs.
DVN-L: What is slowing growth? Where is there significant cost reduction to be had?
P.B.: There are multiple roadblocks preventing the rapid deployment of lidar. Performance was a limiting factor, but with the arrival of the next generation in 2024 – like the Valeo Scala 3 – the range and resolution should be good enough for automotive applications. Once this level of performance is achieved, lidar manufacturers will be able to focus on cost reduction. Compared to cameras and radars, the cost of a lidar is an order of magnitude higher. So, this is still a significant roadblock for OEMs, explaining why lidars are only implemented in high-end cars. The cost of processors for lidar can easily reach more than $80. FPGAs are widely used in lidar, but once developed, an ASIC could reduce the cost. However, it comes with high upfront costs. Depending on the wavelength, the cost of the light source can also be problematic. Photodetectors are also expensive.
When opening a lidar, you see multiple boards inside, easily five or six, indicating that lidars are still not mature and optimization is needed to reduce the cost. This goes along with the automatization of production lines to reduce the number of manual operations.
Regulation is another roadblock as eyes-off driving is allowed only in a few countries: Japan, Germany, the United Kingdom, and a few states in the United States. Regulations allowing automated driving in more countries that clearly define the responsibility of the driver and the OEM (when driving in automated mode) are necessary. Can a driver play Tetris while the car is driving itself, or should he remain alert? OEMs will have to be clear about what can be done or not and not over-promising as Tesla did in the past.
DVN-L: Do you expect a reduction in the number of competitors?
P.B.: Yes, it is clear that a reduction in the number of competitors in the coming years can be expected. If we compare the lidar market with more mature markets like camera and radar, we see these markets are driven by four or five players having 75 per cent of the market with few other players. Yole Group expects a similar pattern in the lidar market. There are currently around 20 lidar players in or entering the automotive market. Not all of them will succeed. Some will die, others will be acquired, or change their focus to other markets than automotive.
