Some lidar suppliers may run into obstacles

In the late 1960s, the Boeing 747 made intercontinental travel an everyday reality. The distinctive hump atop the aircraft became a symbol of 20th-century transportation transformation.

Might another distinctive hump atop a vehicle — that of a cylindrical lidar sensor — come to define transportation advances in the 21st century?

After more than a decade in development for automotive purposes, lidar sensors have become a linchpin technology for autonomy, using lasers to help vehicles determine where they are in the world and detect obstacles in their path.

But the path ahead for dozens of lidar companies remains blurry. As lidar has transformed itself into a multibillion-dollar business over the past year, a handful of companies have separated themselves from the pack. The rest likely face a reckoning.

"There's going to be three to five lidar companies within the next five years," said Angus Pacala, CEO of lidar supplier Ouster.

Over the past five months, five lidar startups have gone public via special-purpose acquisition companies or announced intentions to do so. Collectively, those five have reached a market valuation of $10.6 billion. Dozens of other startups have raised hundreds of millions more from venture capitalists.

Yet experts say there's a reality check on the horizon that's akin to a game of musical chairs. There are plenty of lidar companies; there are only so many automakers still looking for partners. Cruise Chief Technology Officer Kyle Vogt noted the five lidar companies that opted to merge with special-purpose acquisition companies have projected revenue from "entirely overlapping" potential customers.

"It's totally possible the numbers will be correct for one of these companies," Vogt tweeted in January. "But it's not possible for all to be correct."

Vogt outlined the promise and pitfalls ahead for the lidar industry and compared the pitches of some of the most high-profile lidar companies in a 13-tweet thread that culminated with an ominous warning for enthusiastic investors.

"We saw a consolidation/collapse of the robotaxi space over the last 24 months (down to a handful of players), and lidar is next," he wrote. "This probably means lower market caps for most of these co's, which sucks for everyone involved, but may the best product win!"

As the CEOs of lidar startups and others have learned, survival will require more than the best product.

Much to the consternation of Tesla CEO Elon Musk, who has lampooned the sensor as a costly and unnecessary component, lidar has evolved from a science project into what everyone else deems an essential part of any system that allows human drivers to take their eyes off the road, whether momentarily or for the entirety of a trip.

Autonomous driving companies such as Cruise, Argo AI and Aurora have acquired lidar startups, while Waymo and Mobileye have built internal lidar development programs.

Of course, lidar's uses extend beyond cars into trucking, delivery robots, warehouses, industrial applications, smart cities and more. But autonomous vehicle applications have commandeered attention, and that's potentially troubling.

"Investors have no clue, and it's a very difficult sector because each company's technology is very different from each other," said Omer Keilaf, CEO of Israeli lidar company Innoviz, which has partnerships with BMW, Magna and Aptiv, among others. "Everybody is coming up with claims and promises."

For the uninitiated, lidar can be complicated. There are different wavelengths at which the laser pulses can be transmitted — typically 905 nanometers or 1,550 nanometers. Some lidar devices use rotating parts, though for automotive purposes, the industry is gravitating toward solid-state and digital lidars.

Some use amplitude modulation — sending a beam of light and measuring the time it takes to return — while others use frequency modulation, which produces a continuous wave of light.

No matter the approach, lidar companies and automakers grapple with the same challenges. They're attempting to optimize the range, resolution, reflectivity, power consumption and cost, which are all at constant odds with one another.

"There's such a diversity in architectures, and each has a very long list of problems that developers need to overcome," said Tom Jellicoe, an analyst at TTP, an independent technology consulting firm in the U.K. "All of these companies seem to be founded on one piece of the jigsaw puzzle. It's the guys who have done a good job filling in the other pieces that can make a credible case that they might get to the end."

In the early days, lidar could cost more than $70,000 for a 360-degree spinning unit affixed atop a vehicle. Today, earning a contract for mass production demands that cost be below $500 per unit.

"Cost is king," said Keilaf. He estimated that a driver-assist system designed for Level 2 automation will be sold to customers as part of a $3,000 technology package and that Level 3 automated systems will cost $5,000; in either case, the lidar cannot exceed $500, he said.

Costs fall as volume increases, which is one reason why Innoviz has taken an auto-centric, traditional path to the market.

For the past three and a half years, the company has worked with BMW and Magna to ensure lidar can be manufactured at scale. Like many new technologies before, lidar's business path will be through luxury carmakers, Keilaf foresees, as their buyers are early adopters willing to pay a premium for these systems.

And once automakers choose a production lidar supplier, it would be arduous for them to switch for later generations of vehicles.

"Nobody has the appetite to change the hardware," he said. "So we're very focused on winning premium cars today and understanding that will give us the lead. Or at least we will have reserved a seat with the short list of companies that will eventually go forward."

COVID-19 has helped clarify that picture. A year ago, Jellicoe says, many automakers and suppliers were enamored with various lidar solutions, often tinkering with numerous sensor configurations on vehicles and rewriting specifications for their lidar suppliers. The pandemic has compelled automakers and Tier 1s to essentially etch their plans in stone.

"We've noticed a real acceleration," Jellicoe said. "They're cutting away the lightweight ambitions and laser-focused on what's important to them."

Along those lines, Ouster's sales grew 350 percent year over year in 2020, according to Pacala. He agrees with the notion that companies that have already manufactured lidar as part of production contracts will remain viable while competitors without that experience will falter.

"That's an extremely good metric," he said.

But compared with Innoviz, Ouster has taken a different approach to business development. Pacala is reluctant to marry the company to any particular customer and potentially narrow its market.

Ouster has more than 800 customers across multiple industries, per financial documents. It sells lidar for driver-assist systems at $600 per unit. But automotive is just 30 percent of Ouster's customer base. The company has gained traction in other applications, such as trucking, mining and defense. Beyond that, Ouster may supply lidar for everything from e-bikes to aerial vehicles. In Pacala's view, the real opportunity is not limited to one particular sector, but the chance to redefine transportation writ large.

"I'm a car nut," Pacala said. "But the opportunity is much bigger than automotive. Ouster's goal is to put a lidar sensor on every moving object on Earth."