Perfecting driverless tech has taken longer than expected, even before the pandemic.
Perfecting the technology has taken longer than expected. The coronavirus pandemic has made it even more difficult. By Cade Metz and Erin Griffith
Tech companies once promised that fully functional, self-driving cars would be on the road by 2020 and on the path to remaking transportation and transforming the economy. But a decade after Google unveiled an autonomous car prototype with global fanfare, the technology is still far from ready, and many investors are wary of dumping more money into it — just when the world could benefit from cars that ferry people and deliver packages without a human driver.
The companies that made these promises are now in a jam: To perfect their technology, they need to test it on roads. But they need at least two people in the cars to avoid accidents. Because of social distancing rules meant to keep people safe during the coronavirus pandemic, that is often not possible. So many cars are sitting in lots.
“This is a difficult time for everyone,” said Bryan Salesky, chief executive of the startup Argo AI, which is backed by $1 billion from Ford and another $1 billion in promised funding from Volkswagen. “We want to get back on the road as soon as it is safe to do so. There is no substitute for on-road testing.”
The timeout caused by the pandemic has hastened an industry shakeout that was already starting to happen. Many selfdriving car companies have no revenue, and the operating costs are unusually high.
Autonomous vehicle startups spend $1.6 million a month on average — four times the rate at financial tech or health care companies, according to PitchBook, which tracks financial activity across the industry.
It’s a sharp turn from 2016, when an investment bubble in selfdriving technology started. General Motors acquired Cruise, a three-year-old, 40-person startup, for roughly $1 billion including performance incentives. A few months later, Uber announced that it would pay around $680 million for Otto, a six-month-old autonomous trucking startup.
The price tags for those deals worked out to about $10 million per engineer, and that became the going rate. A fledgling threeperson startup, for example, valued itself at $30 million.
Now one self-driving car startup has gone out of business, and another is for sale. Four have laid off employees. And bigger companies are hunkering down to wait out the delays.
Cruise said that although it had gotten some cars back on the road by making deliveries for two food banks in San Francisco, its testing had been curtailed. Earlier this month, Ford, which has temporarily closed factories because of the virus, pushed the launch of its autonomous service from 2021 to 2022.
At Waymo, the self-driving car unit of Google’s parent company, Alphabet, the pandemic has set work back at least two months because of social distancing rules and trouble getting hardware from other countries.
The startup Zoox, which investors have valued at $2.7 billion, recently hired the investment bank Qatalyst Partners to explore a potential sale while it also tries to raise new funding, according to two people familiar with the effort, who were not allowed to speak about it on the record.
The news was reported earlier by The Information.
There have been lay-offs at Zoox, at the autonomous trucking companies Ike and Kodiak Robotics, and at Velodyne Lidar, which makes the Lidar sensors that are an essential part of most autonomous driving. Lyft, which recently laid off or furloughed more than 1,000 employees, said its autonomous division was affected.
“It was appropriate and necessary to be conservative about our cash burn,” said Alden Woodrow, Ike’s chief executive. “That had to happen.”
Before the pandemic, Voyage, a startup in Silicon Valley, tested its autonomous vehicles inside retirement communities in California and Florida.
Realising the limits of self-driving technology, the startup was focused on situations where its cars would face less traffic — and less chaos, said Oliver Cameron, the company’s chief executive.
On a recent morning, Cameron and several of his engineers logged onto a Zoom videoconference call. A virtual recreation of The Villages, a retirement community in San Jose, California, appeared on their screens. The simulation was built from digital data collected over the past several years by cameras and other sensors installed on the cars.
Inside the digital simulation, the company’s autonomous vehicle slowed behind a parked car as traffic approached from ahead. It stopped to let the oncoming traffic pass, but then stalled, failing to proceed once the road was clear.
With simulated tests, companies like Voyage could make some progress, but not all scenarios could be tested.
“Simulation is not something you do in a vacuum, without any connection with the real world and real data,” said Davide Bacchet, Voyage’s vice president of engineering. “We can only progress to the point where the simulation is accurate.”
Cameron estimated the company’s latest autonomous vehicle had already been delayed by four months, partly because of hardware supply-chain slowdowns in China.
Voyage has raised $52 million, which he
‘‘
That was a clear moment in time where the whole industry went from being a bull market to a bear market. Covid-19 has taken us even further into the bear market. OLIVER CAMERON
CEO of Voyage
said would last until the end of 2021. But until the technology is ready, no revenue will be coming in.
Autonomous driving research was derailed, in part, by a death in Arizona. In March 2018, one of Uber’s autonomous vehicles killed a pedestrian in Tempe.
Many companies temporarily took their cars off the road, and after it was revealed that only one technician was inside the Uber car, most companies resolved to keep two people in their test vehicles at all times.
“That was a clear moment in time where the whole industry went from being a bull market to a bear market,” Cameron said. “Covid-19 has taken us even further into the bear market.”
It became clear that the technology was years from being ready. The cars still made mistakes in unexpected ways. And crossing the safety hurdles would take much longer than had been expected.
“Before the pandemic hit, there was a broad recognition, within the industry, that this was a much harder problem,” said Anand Gopalan, chief executive of Velodyne.
Waymo, for example, operated an automated car service in Phoenix, but rides still required safety drivers behind the wheel.
In the coming months, some of the betterfunded companies can draw on money that is already in the bag. Argo AI, the self-driving company started by Ford, for example, can most likely lean on the Volkswagen money.
Others aren’t as fortunate. As the pandemic approached in February, Starsky Robotics, an autonomous trucking startup, shut down. It tried to sell its assets, but potential buyers were skittish, said Stefan SeltzAxmacher, chief executive.
Last year, Drive.ai, a startup backed by $77 million, sold to Apple. One of Drive.ai’s investors, the venture firm NEA, narrowed its autonomous investment focus to robots in places like retail warehouses.
“With autonomous vehicles, you may find yourself in a company that requires billions of dollars of capital, with no clear timeline for building a large business or seeing a return on the investment,’’ said Aaron Jacobson, a partner at NEA.