For years, Tesla CEO Elon Musk has said lidar sensors are too expensive for consumer cars. He’s called the sensor a “crutch,” although most experts believe its key to making cars drive themselves. “Anyone relying on lidar is doomed,” Musk said at Tesla’s 2019 Autonomy Day.
Lidars, which emit pulses of light and measure how long they take to return to create a 3D map of a vehicle’s surroundings, are commonplace on the self-driving test vehicles operated by the likes of Waymo and Cruise. Those companies think lidars are key to developing a full, accurate understanding of the world.
But while lidar prices have declined over the years, you won’t find the sensors, or their capabilities, on consumer cars. Today’s semi-autonomous offerings from Tesla, Cadillac, Nissan, and others must rely on cameras, radars, and (in some cases) maps.
In May, the company announced an agreement with Volvo to include its Iris lidar sensors, which cost about $1,000, in models built on Volvo’s new vehicle architecture starting in 2022. The lidars will help power an automated-driving system designed for highways, one the Swedish automaker promises will be more robust than comparable features available today.
Volvo’s endorsement of Luminar’s tech spurred interest from potential investors, Russell told Insider, helping him strike a deal to go public through a reverse merger with Gores Metropoulos, a special-purpose acquisition company, or SPAC.
Vertical integration was the key to slashing costs
The key to making the Iris affordable was bucking the lidar industry’s tendency to rely on suppliers, and instead build many components — lasers, receivers, scanners, and processors — in-house, Russell said. Vertical integration was also necessary to fulfill the performance and safety standards required for a consumer vehicle. It wasn’t possible to build a lidar which could both see at least 250 meters away and shoot lasers at a frequency that was safe for the eyes of nearby pedestrians with off-the-shelf parts, Russell said. That’s why Luminar spent its first five years in stealth mode, building a new lidar from the ground up.
The result is a sensor that Russell says has a fraction of the parts found in competing lidars, while offering better performance. The Iris uses one laser, while some competing models have more than 100.
“Systems that require huge arrays of lasers and receivers are not going to be cost-effective,” Russell said.
Luminar also had a style problem to solve. In the development vehicles where lidars are most often used, they tend to jut out from the roof like the circular sirens found on old police cars. A bulky design isn’t a problem for a company that cares more about gathering data than aesthetics, but the average car buyer won’t make that sacrifice, Russell said. (Nor the average car designer.) Luminar was able to arrange the Iris’ components into a compact design that creates only a slight bump above the windshield.
“This has to be something that consumers like,” Russell said. “There’s no way any kind of roof-rack-based system is making its way onto these consumer vehicles.”
Though Luminar has been most outspoken about its lidar sensors, the company also makes software and other hardware components for driver-assistance systems. The company won’t have to look far to find new customers, as Russell said Luminar is already working in some capacity with seven of the world’s 10 largest automakers and expects them and other current partners to account for 80% of Luminar’s future revenue.
“If there’s any company in this space that really has reached a stage and level of maturity from both a product perspective and a commercial perspective, I think we’re probably in a lead position there,” he said.