|Lidar Is Finally Becoming a Real Business|
The success of lidar companies was unclear until this summer, when three leading makers offered glimpses of their finances.
TIMOTHY B. LEE, ARS TECHNICA
09.11.2020 08:00 AM
FOR YEARS, THE lidar business has had a lot of hype but not a lot of hard numbers. Dozens of lidar startups have touted their impressive technology, but until recently it wasn't clear who, if anyone, was actually gaining traction with customers.
That's starting to change. This summer, three leading lidar makers have done major fundraising rounds that included releasing public data on their financial performance.
The latest lidar maker to release financial data is Ouster, which announced a $42 million round in a Tuesday blog post. That post also revealed a striking statistic: The company says it now has 800 customers.
That's interesting because we can compare it fairly directly to two other prominent lidar companies that have released data in recent months. Velodyne, which has been considered the industry leader for the past decade, revealed in July that it had 300 customers.
The lidar startup Luminar hasn't revealed its number of customers, but it disclosed two other figures in August: The company has 50 commercial partners and expects to sell roughly 100 lidar sensors in 2020.
By the metric of total customers, then, Ouster seems to be well ahead of two of its better-known rivals. But saying that Ouster has become the industry leader would be too simplistic. In reality, the three companies are each pursuing different segments of the market.
At the high end of the lidar market are powerful sensors that sell for tens of thousands of dollars each. Self-driving-vehicle companies buy these units for their prototype vehicles. Because these companies are well-funded and are making only a few prototype vehicles, they're willing to pay piles of money to get the most powerful sensors available. This market has traditionally been dominated by Velodyne, which has charged as much as $75,000 for a single sensor.
At the opposite end of the spectrum are lidar sensors intended for mass-market automotive applications. Lidar sensors for this market generally need to stay under $1,000 to be viable.
The pioneer here was a little-known company called Ibeo, which partnered with auto supplier Valeo to provide lidar sensors for the 2018 Audi A8. The sensor was primitive, with only four vertical "lines" of resolution. But it was the best lidar Audi could afford given the financial constraints of the consumer car business.
This is the market Luminar is gunning for. Luminar's lidar is much more powerful than the sensor in those early Audis, and the company believes it can get the cost below $1,000 at scale. Back in May, Luminar announced a deal with Volvo to incorporate its lidars into vehicles beginning in 2022. It's the first deal to put high-performance lidar into consumer vehicles. Luminar hopes it will be an inspiration for other automakers.
Ouster makes spinning lidar that looks a lot like Velodyne's high-end sensors. But inside, Ouster uses solid-state chip technology to pack all of its lasers—16 to 128 of them, depending on the product—on a single chip. Ouster's sensors are much simpler than Velodyne's classic design, which involved packaging together 16 to 128 individual lasers and 16 to 128 individual sensors.
The resulting combination of strong performance and relatively low cost has opened new markets for lidar sensors. Ouster's latest generation of 32-laser sensors start at $6,000. That's way too expensive for mass-market automotive use, but it's much less than Velodyne charged for comparable sensors before Ouster came along.
Earlier this year, I talked to John Williams, chief technology officer at Kudan, which sells software to help robots track their own location (a problem known as SLAM in the robotics world). Williams said plenty of companies are building custom robots for niche applications in mines, warehouses, and other industrial environments.
Lidar sensors have obvious value for this kind of application. But before Ouster came along, high-quality lidar was simply too expensive.
"The fact that you can get a 64-channel spinning lidar for $12,000 was unheard of," Williams told me. While Ouster's lidars were "not quite as good as Velodyne" in his opinion, he argued that the company was "the up-and-comer in terms of disrupting the market."
So that may explain Ouster's 800 customers: Companies shipping robots to real customers can't afford to blow tens of thousands of dollars on a lidar sensor. But in many cases, their industrial customers are willing to spend a few thousands dollars for products that better understand the world around them.
Ouster also offers a compelling value proposition to university researchers. Researchers need high performance, but they don't have unlimited budgets. There are probably a lot more research labs with uses for a lidar sensor than there are self-driving projects.
This "broad middle" market is growing fast. Ouster says its third-quarter sales are already more than three times as high as they were in the third quarter of 2019. And Q3 2020 isn't over yet.
A big question for the next few years will be whether falling prices and improving performance will bring these companies into more direct competition.
The company with the most to lose is Velodyne, which has dominated the market for a decade but now faces growing competition from low-cost rivals—especially Ouster. History has shown that it's difficult for a company with a high-performance, high-cost product to deal effectively with a challenge from a cheaper, disruptive rival. It may prove much easier for Ouster to gradually improve its sensors than for Velodyne to slash its prices enough to compete with cheaper rivals.
In discussions with Ars, Ouster CEO Angus Pacala has argued that Ouster will benefit from progress in the broader semiconductor industry. If all goes well, Ouster's solid-state lasers and sensors will continue to get cheaper and more powerful in much the same way that computer chips have over the last 50 years.
Velodyne is certainly trying to meet the threat here. For example, the company has a product called the Velarray that's designed to be cheap enough for the automotive market. Whether Velodyne's lower-cost products have been gaining traction is not clear.
Meanwhile, Ouster and Luminar have very different go-to-market strategies. Luminar has focused on scoring big, multiyear deals with major automotive customers. Ouster has eschewed that market, focusing on selling lidar units a few at a time to a wide range of individual customers.
If Luminar is able to deliver on its Volvo contract and score similar deals with other carmakers, that could provide enough volume to achieve unrivaled economies of scale. Luminar might then be able to undersell higher-price rivals like Ouster and become the overall market leader.
Supplying equipment to automakers, on the other hand, is a notoriously cutthroat, low-margin business. If Luminar can't sign up other automakers, it could find itself yoked to a demanding but not especially lucrative customer.
Lidar is not just a three-horse race, of course. Velodyne is facing competition from Chinese rivals like Hesai and Robosense. Velodyne sued both companies last year for patent infringement.
Luminar, meanwhile, faces formidable competitors in the automotive market, including tier-one auto supplier Bosch. And there are a number of other independent lidar companies that could become significant players in the future.
This story originally appeared on Ars Technica.