There’s a race to the bottom in making self-driving cars more affordable than ever, but Austin Russell has been running an entirely different race.
The 22-year-old college dropout has been sitting quietly for the last five years and watching as manufacturers rush to create cheaper versions of Lidar, a laser-based radar system that’s a key component in self-driving cars since it allows cars t0 “see” the road.
Russell, now the CEO of self-driving car technology startup Luminar, decided to skip working on a cheap system, instead spending the last five years with cofounder Jason Eichenholz putting together a Lidar system they hope is better than the rest.
“There’s a reason we’ve been in stealth so long,” Russell said. “If we announced our plans five years ago, everybody would be doing this.”
Comes down to seconds
Just four piers down from San Francisco’s famous seals at Fisherman’s Wharf is an empty cruise terminal filled with a smattering of fake deer, mannequins, and tires strategically stationed at points on the floor.
- Biz Carson/Business Insider
At one end of the warehouse, Luminar has installed its Lidar device into a car, pointed at its target: A blackboard as tall as an SUV, sitting next to a “200 meter” sign.
This had been Russell’s target to hit before publicly debuting his five-year-old startup, Luminar. 200 meters is longer than most Lidar systems for self-driving cars can see, and seeing dark black objects reliably at that distance is even harder.
During a press demonstration at the pier, Russell and Eichenholz’s system zoomed in on the black billboard in the distance. The Lidar was even able to pick up the pigeon walking in the middle of the fake deer on the floor, nearly 100 meters away from the car.
Russell was a kid of barely legal driving age when he decided to conquer self-driving cars. He’d been a tinkerer most of his life.
He memorized the periodic table by two, transformed a Nintendo game handset by the sixth grade, and made a holographic keyboard in high school. In 2012, Russell founded Luminar and soon after started doing independent research at the Beckman Laser Institute.
When he got to Stanford University in 2013, venture capitalist Peter Thiel paid the then-18-year-old Russell $100,000 to drop out of school. His Thiel fellowship bio said he had a “passion for developing innovative optoelectronic technologies”. For Russell, being innovative meant having the patience to build a better Lidar that would not only advance an industry, but do it in a field that could save people’s lives.
“We’re able to see seven seconds out instead of one second,” Russell said. “That’s a really big breakthrough.”
While some manufacturers say they can produce a Lidar that sees out to 200 meters, most advertised models can’t scan that far ahead. Going highway speeds, that means a car equipped with Luminar’s Lidar would have seven seconds of reaction time when it identifies an object at 200 meters compared to the typical few seconds most Lidar systems provide, Russell claims.
Of those that can hit the 200 meter mark, Eichenholz says Luminar is in a class of its own because it can see objects with 10% reflectivity at that distance. That low-reflectivity number is key because objects like black cars and even people wearing black t-shirts can be harder to spot since the color is less reflective.
“You don’t have to say everyone has to wear a white t-shirt,” Eichenholz joked.
An entirely different system
Russell and Eichenholz claim Luminar can do all this because they’re not using the same off-the-shelf parts as other Lidar companies. Instead, they spent the last five years designing a system from scratch.
For starters, the company eschewed traditional Lidar designs and build a new one on a different wavelength than traditional systems. While many Lidars systems operate at a 905 nano-meter wavelength, Luminar is at a 1550 nano-meter wavelength, which means it can emit 68 photons for every single one put out on the traditional wavelength, making it more powerful but still safe for eyes.
“Eye safety is a really big deal in this industry, It’s actually is a limiting factor of a lot of Lidar sensors today,” Russell said. “You can’t fry people’s eyes with sensors.”
The decision to switch wavelengths set off a cascade of events for the startup. Namely, there were no off-the-shelf products it could use. It’s had to do everything from creating its own chips and circuit design to manufacturing its own indium gallium arsenide, a chemical compound that it’s using as a receiver in its Lidar unit instead of silicon.
The company tried 2,000 different ways to construct a Lidar system before deciding on one that was both functional and manufacturable, Russell said.
The company raised $36 million from venture capitalists including GVA Capital, Canvas Ventures, and the 1517 Fund, a venture firm backed by Thiel. With the money, Russell acquired one of Luminar’s suppliers to lock up its supply chain from competitors, He also brought on Eichenholz’s previous company, Open Photonics, and named him cofounder and CTO.
While Russell oversees the company’s Silicon Valley office, a converted tank repair facility in Portola Valley, the majority of the company’s 150 employees are based in its Orlando, Florida office, which contains both R&D headquarters and 50,000 square foot manufacturing facility.
Even though Luminar has stayed quiet for the last five years, it’s already produced 100 units for its strategic partners, which include both unidentified automakers and software companies. Now, Luminar is ready to begin its first major commercial run of 10,000 units, set to be distributed among many customers instead of just putting it in the hands of a few.
“Our plan is to power every autonomous vehicle that’s produced and make them so they can truly be safe and autonomous,” Russell said of his long-term vision.
Our plan is to power every autonomous vehicle that’s produced and make them so they can truly be safe and autonomous
To do that, the college dropout’s startup will have to overcome major incumbents like Velodne as well as another onslaught of software companies – like Google and Waymo – that have developed their own in-house Lidar systems. But Russell is confident that Luminar’s system will become the new standard, and he’s not worried that everyone will soon be joining his race lane now that his company’s plans are public.
“The key thing here is that starting this five years ago we have a huge head start and advantage,” he said. “No matter how much money you throw at this problem, you don’t need hundreds of millions or billions to develop this. It’s about the time it takes.”