Russell says he has been approached by a half dozen competitors asking if Luminar would be interested in acquiring them. The biggest self-driving developers, including Waymo, Cruise, Argo, and Aurora, have either acquired lidar companies or developed their own tech in-house. The flow of venture capital has slowed, according to Crunchbase data. And at least one player, Israel’s Oryx, flat out folded last summer.
So far, there’s been a lack of high-profile washouts, and who will fail, merge or survive is far from clear. That’s largely because the differences between one lidar maker consist not just of business strategies, but approaches to the technology itself. A developer can fire its laser pulses at 900 or 1550 nanometers (or anywhere in between), and make its receivers out of silicon or ingaas. Frequency Modulated Continuous Wave lidar, the tech of choice for Aeva, Aurora’s Blackmore, and Cruise’s Strobe, detects an object’s velocity along with its shape. Russell says Luminar has developed the same capability, with a different approach. Australia’s Baraja uses the physics of prisms to see the world. Sense Photonics is one company working on a “flash” lidar, which sends out many laser points simultaneously, taking in information much like a camera.
Moreover, the lidar market will create a honeycomb of niches, based on who is using a vehicle and how (to say nothing of the robotics and other industries). Luminar focuses on long-distance vision, key for highway driving. France’s Valeo has logged $564 million worth of orders for its lidars, which work best for shorter distances. Israeli’s Innoviz has partnered with auto industry supplier Magna and landed a contract to produce sensors for BMW. Velodyne, which pioneered lidar for automotive purposes starting in the 2005 Darpa Grand Challenge, produces a wide array of products, starting at just $100.
“The market leaders are pulling away a bit,” says Mike Ramsey, an industry analyst with Gartner. But the big differences among these companies, and the wide variety of opportunities for success, are keeping many of them alive.
Who gets culled may hinge less on choices about nanometers and materials than on drudgier realities, says Matt Johnson-Roberson, CEO of Refraction AI, which is developing an autonomous delivery vehicle. “I’m less interested in 1,500 nanometers versus 900 versus whatever,” he says, than in whether it works, is affordable, and is available. The industry has collectively failed to hit that mark so far, though it’s getting closer. Whoever makes it first will likely get Johnson-Roberson’s business. “I’ll use whatever’s best,” he says. “I have no brand loyalty.”
More Great WIRED Stories