The race to win in self-driving hit a snag in 2024, but things were still strong at CES. CES isn’t really a consumer electronics show anymore – even the big classic consumer electronics vendors like Samsung, LG, Sony and others made their booths mostly about AI that wants to serve you (not t ‘they sell you physically). A carriage show has been made, and the new West Hall and part of the North Hall were devoted to it.
American automakers were largely absent this time around, though a few new Chinese ones did make an appearance. The most important company in self-driving, Waymo, showed two versions of their 6th generation hardware, one on the promised Zeekr minivan platform and another on the non-Chinese Ioniq 5 platform that is intended to replace it due to of the 100% set fee. Chinese import cars. Waymo continues to develop for Zeekr, and some legal analysis says they will be exempt from fees until the Ioniq is ready. The Zeekr is a longer and wider minivan, which may be preferred by customers over a crossover SUV.
Zeekr had its own booth and showed a version of the vehicle with the front seats flipped back to provide head-to-head seating, with the steering wheel locked. Waymo may offer this setup in the future, it is speculated. Indeed, Waymo can still use Zeekr after fees in many other countries around the world outside the US if they allow Chinese-made vehicles.
Chinese-made vehicles have strong advantages for robotaxis, due to low cost, speed of production, lack of need for recognizable license plates or even high mechanical reliability.
On the other hand, Korean manufacturers such as Hyundai can also make good value vehicles and are known around the world. The Ioniq should also be popular and produce a good taxi ride at a reasonable price.
Lidars and Imaging Radars
CES has always featured a ton of LIDAR companies, and this year they were back in force, despite the dwindling number of self-driving teams for sale. Indeed, most of them now see the initial volume of sales coming from regular automakers hoping to produce higher quality driver assistance tools and ADAS pilot tools that drive with human supervision, or perhaps so-called “level” 3 cars driving on highways while a human is able to ignore the road but must take over before going to city streets or certain complex situations. The volume of robotics being done today is far less than the LIDAR companies hoped it would be at this point.
However, there is a large array of offerings, with Chinese vendors Hesai and Robosense leading the way in volume. There were long-range FMCW 1550 nm LIDARs and plentiful lower-cost, shorter-range units. An interesting word came from Israeli companies, which hope they can gain a distinct advantage simply by being neither Chinese nor American. The US can resist Chinese LIDAR and China can resist US LIDAR, leaving other countries alone able to supply for use globally. Or so they hope.
One company that didn’t speak LIDAR was one of the world’s largest automotive suppliers: Bosch. Previously, Bosch had promised to release a LIDAR that would surpass all others on the market, and apparently it would sell well considering Bosch’s strong position in the industry. They have removed this product, according to a spokesperson.
There was even more growth in the field of imaging radars. Traditional automotive radars have very low resolution. So low that they are not useful for detecting static objects such as road debris or stalled cars. The whole world is static objects. Radar is great for seeing moving objects because it detects speed. Imaging radars get enough resolution—some claim below 0.5 degrees—that they may even replace LIDAR. This is good because they can cost less, and also because they are immune to fog and other weather. However, most teams aren’t looking to save money by replacing a sensor just yet. They want the best, even at the highest cost, and will work to be free later. Some imaging radar suppliers have special hardware, some only offer software. Like almost every booth at CES, the big buzzword was AI—in this case the use of AI pattern matching techniques to better classify radar targets and even improve resolution to understand their location.
May
With Cruise gone, Argo gone, and Apple and Motional in doubt, the robotics space in the US is now largely limited to Waymo and Zoox. Zoox was present on the show floor and gave tours to the press, but their planned public launch in Las Vegas for 2024 was delayed. Press leaks suggest it should happen in 2025, however, and not just “stay in Vegas” as San Francisco is also planned.
The other player present was May Mobility, which operates driverless security vehicles in confined spaces in Ann Arbor, Michigan. May has put the focus on ships, with more limited routes with limited stops. That makes the problem easier for a smaller company like May, but it means they’re not quite ready for robotaxis. May is conducting operations with security leaders in several other cities such as Miami, Martinez, Grand Rapids, Arlington, Sun City AZ and Detroit.
Also present were MobilEye (which I’ll cover later) and Helm.AI, who announced a partnership with Honda (and thus Nissan.) Helm.AI doesn’t have a functional self-driving staff at the moment, but they will be hard at work helping Honda make “tier” 3 cars.
The strange one
A Chinese company offered a prototype with a large truck with a special cargo bay containing a short-range multirotor aircraft. Owners of this vehicle, which costs around $340,000 if and when it ships, will be able to drive to faraway places and have fun flying around. Mainly for fun and sightseeing, although if you had a house on top of the hill, it might make sense as a quick way for one person to quickly travel to it. Pivotal (formerly Opener Blackfly) was also on display and it’s true – they had actual customers on hand talking about their flying experiences. Although they won’t say how much they sent, it’s not much.
This motorhome was still full of rough edges, but it was all electric and included the most spacious slide-out living room I’ve seen in an RV. They had specifics, but at this point it’s not clear how meaningful they are, since they’re years away. The Pebble electric trailer, which I’ve covered before, was also there, a little further along, as well as another competing electric trailer, in which the entire top of the trailer sits down (like a pop-up trailer on steroids) in they make it half height while pulling. The big problem with towing in EV is that it cuts the range in half or more, and that’s a real drawback. Reducing the cross section of the trailer is the most effective way to reduce drag and increase range. The Lightship trailer isn’t ready for production yet, though they expect it to go into production in 2025, at a similar cost to the Pebble.
Of course, there was a lot more – you can’t tour all of CES over its 4 days, although traffic was still a bit down from the Covid days. The showroom space, on the other hand, had begun to match the previous era, so the show is making a comeback. By far, the biggest theme was a claim that products are now AI-enabled. Unfortunately, most of the claims were hype, but that won’t end anytime soon.