Austin Russell has only been legal to drive for six years, but at 22, he might be breathing new life into the self-driving car industry by creating a safer navigation system. All of his hard work and dreams are about to get tested.
As a teenager, Austin Russell started thinking about creating better 3D imaging technology while spending his last couple years of high school at the Beckman Laser Institute at the University of California, Irvine; designing augmented reality and wireless power transmission projects.
Now at just 22, that early obsession has resulted in an advanced laser LiDAR sensor that he says paves the way for self-driving cars to “actually work and be safe.” Russell is CEO of Luminar Technologies Inc., a startup with operations split between Silicon Valley’s Portola Valley and Orlando, Florida, that’s been in stealth mode since he founded it five years ago.
Having hit key technical performance targets, Luminar is emerging from its shell with plans to make 10,000 automotive LiDAR units this year at an Orlando factory. These shoebox-sized devices can be mounted on a car roof or integrated into a front fender and will be sold to automotive and tech customers developing autonomous vehicles. In 2018, Luminar aims to scale up production to hundreds of thousands of the devices annually, and millions annually after that.
What distinguishes Luminar’s tech from existing LiDAR devices, made by companies including Velodyne, Ibeo and Quanergy, is how much farther they see and how much sharper the resolution is, according to Russell and Luminar co-founder Jason Eichenholz.
“We built a system from the ground up that delivers 50 times better resolution and 10 times farther range,” an enthusiastic Russell told Forbes. “We have the ability to see dramatically farther out, and we’re the only system that can be specced high enough to actually be the first to meet the industry requirements for safety and performance.”
Perfecting autonomous cars requires algorithms and artificial intelligence that can operate a vehicle as well or better than skilled human drivers, and relies on vision systems and sensors that provide instantaneous details of surroundings, preferably in 3D and in 360 degrees. LiDAR has become essential technology to do this, able to accurately monitor other cars, objects in the road, pedestrians, potholes and ice, in all weather and light conditions, though it remains costly and in need of refinement.
LiDAR’s importance is underscored by big deals, including $150 million invested in Velodyne by Ford and China’s Baidu last year, and Quanergy announced raising $90 million in 2016 to fund production plans. The vision technology is also at the center of a high-profile legal fight playing out in a San Francisco federal court, with Alphabet Inc.’s Waymo accusing Uber and its Otto self-driving truck unit of theft of trade secrets related to its proprietary LiDAR system.
Luminar’s LiDAR units detect and distinguish objects with precision at 200 meters compared with about 35 meters for competing systems, even black objects that conventional LiDAR can miss, according to the company. That distance is critical as 200 meters gives a car traveling at highway speed about seven seconds to respond to dangers ahead, versus a single second if the system only sees accurately 35 meters ahead.
Luminar achieves that by using a unique chip design utilizing indium gallium arsenide, rather than silicon, and a more sensitive laser that operates at the 1550-nanometer wavelength, instead of the 905nm wavelength used by competing LiDAR sensors, which provides 40 times more power.
And while competitors talk about massive cost reduction for their devices, with prices promised to fall from thousands of dollars per unit currently to hundreds of dollars each and less in the next few years, Luminar believes they’re moving too fast to cut the cost of technology that’s not sufficient for the task of autonomous driving.
“There’s a huge race to the bottom in the race for commoditization of LiDAR,” Russell said. “It’s getting to the point where these sensors are more appropriate for Roombas (robotic household vacuums) than autonomous vehicles. They don’t have anything close to good enough to drive an autonomous car.”
Luminar has already supplied prototype devices to four companies that Russell declined to identify, and it’s in talks with “80 percent” of companies in the autonomous vehicle space. After it completes this year’s 10,000-unit run, “we’ll be working with everyone.”
He declined to provide the current cost of a Luminar LiDAR unit and a target price over time, though.
To date, Luminar has raised $36 million of seed capital from investors including Canvas Ventures, GVA Capital and 1517 Fund. (Not $150 million, as reported in a recent story.) Russell himself got a $100,000 fellowship in 2013 from PayPal founder Peter Thiel, which allowed him to suspend his studies of applied physics at Stanford University after his freshman year to devote himself full time to Luminar.
A “significant,” though unspecified amount additional fundraising is planned to support rapid expansion of production capability, “but we’re not talking about billions of dollars,” said the lanky, six-foot, four-inch Russell, who takes pride in the company being highly capital efficient up to this point.
Orlando, where Luminar has a 50,000-square-foot production facility, turns out to be an ideal location for a LiDAR company owing to the region’s extensive network of defense and aerospace companies with expertise in optics and photonics for NASA and military projects.
“Everybody is fighting over the same talent here in Silicon Valley,” Russell told Forbes. “But there’s a lot of amazing talent we’ve been able to pick up in Orlando.”
That talent includes Eichenholz, Luminar’s co-founder and chief technology officer. Russell describes him as the “Kevin Bacon” of the photonics industry, owing to his extensive connections to all its key players.
Though still quite young, having just turned 22 in March, Russell has a long-term vision for Luminar that doesn’t include it being sold.
“We’ve built this out to be a highly profitable, long-term company,” he said. “We’re not trying to do this to burn a bunch of venture capital and hope to get acquired by someone. We’re building up this business to have some substantial revenue.”