A handful of recent announcements and funding rounds suggest that autonomous driving is making a comeback. Elon Musk told an audience at a glitzy Tesla event in October this year that its self-driving Cybercab would be available “before 2027”. Less than two weeks later, Google led a $5.6bn round that valued Waymo at $45bn as the robotaxi producer – already a frequent sight on San Francisco’s roads – continues to expand into new cities.

The Waymo round is the largest so far in the autonomous driving sector, and it follows a billion-dollar round for self-driving technology developer Wayve in May 2024, and nine-figure rounds for Tier IV and DiDi Autonomous Driving since.

But these headline rounds obscure the fact that the self-driving car market is now largely divided between the few large companies pulling in rounds and the rest, for whom funding is drying up.

Raymond Zheng, senior managing director of carmaker Honda’s Xcelerator Ventures unit says that for most carmakers, the hype around achieving full or even high levels of automation has died down. They are focusing on the more modest goals of Level 2 automation – advanced driver assistance systems (ADAS) – and moving from there.

“There’s no doubt it’s slowing down from a few years ago,” he says. “OEMs are really focusing on what they do best, which is to build cars for people. Starting from Level 2 or 2+, trying to go to Level 3 or 4, that progressive path rather than revolutionary – that’s what is happening right now.”

Autonomous driving was one of the hottest areas in startup technology for a few years, culminating in Waymo and Cruise raising multibillion-dollar rounds in 2021. Further down, however, funding for startups focusing on products like sensors, software or lidar – the laser-based system that helps vehicles detect obstacles – has dropped sharply.

Tony Cannestra, director of corporate ventures for automotive component producer Denso, views the flurry of startups that emerged around the mid-2010s as a first wave, similar to how Web 1.0 first shaped the internet. But he can’t see a second wave on the horizon for this sector.

“Say, eight years ago…if you had called me up and asked what I thought about lidar, I would have literally had a list on my laptop of about 80-something lidar startups around the world,” he says. “And 95% of those lidar startups are dead now. The OEMs have not yet rapidly deployed lidar, and they haven’t figured out how to build it into the sensor suite because it’s not super cost effective, even though it provides fairly good computer vision benefits.

“Now if you’re a new startup and you’re coming to me with a lidar technology pitch…you better be a lot better technology wise and you better be much cheaper, because we know that’s the main problem with lidar. And you better be able to raise a lot of money to survive. I’m not sure the entrepreneurs understand that yet in automotive.”