- The autonomous-driving startup Zoox has a promising but challenging business model.
- The company is developing a new kind of electric vehicle designed specifically for autonomy.
- Making a custom vehicle will give customers a safer and more enjoyable experience, Zoox CTO and cofounder Jesse Levinson told Business Insider.
- The company is also focused on urban ride-hailing, starting with San Francisco and Las Vegas, rather than beginning in a less challenging suburban environment.
- If you spend too much time and effort perfecting your technology for suburban roads, it might not be able to handle the transition to cities, where there is more demand for ride-hailing services, Levinson said.
- Zoox vehicles are already better than humans at some parts of driving, Levinson said, like obeying stop signs and speed limits.
- And their cameras, radar, and lidar sensors give them “super-human” sensing capabilities that allow them to have a 360-degree view of their surroundings, and even what’s behind the objects next to them, at all times.
- But Zoox vehicles are still worse than humans at handling what those in the autonomous-driving industry call corner cases, unusual situations that drivers rarely encounter.
- Sign up for Business Insider’s transportation newsletter, Shifting Gears, to get more stories like this in your inbox.
- Visit Business Insider’s homepage for more stories.
Waymo may have more experience and access to more money than any of its rivals who are working on self-driving cars. In 2018, it became the first company to launch a commercial autonomous ride-hailing service, and it is seen by experts as the leader in the autonomous-driving industry.
Zoox thinks it has a better strategy for solving the enormous challenges posed by self-driving vehicles.
The startup, founded in 2014, is developing a new kind of electric vehicle from the ground-up, rather than installing its technology on a Chrysler Pacifica or a Volvo XC90, and focusing on offering rides in cities, starting with San Francisco and Las Vegas, rather than the suburbs Waymo has used to launch its Waymo One ride-hailing service.
Zoox’s ambitions are exciting in theory, but they come with a high degree of difficulty. The experiences of Tesla and the failed or struggling electric-vehicle startups that have attempted to follow it have demonstrated the enormous challenge of vehicle assembly, even on a small scale. And programming a car to drive in a crowded city is harder than doing the same in a quiet suburb.
But there are benefits to having a distinctive business model. Jesse Levinson, Zoox’s cofounder and CTO, told Business Insider that he doesn’t need to spend much effort convincing potential employees to work for the company instead of rivals like Waymo and Cruise, which is owned by General Motors.
“I don’t actually have to pitch them that hard, and I try not to, because we want people who are kind of intrinsically excited about our approach,” he said.
A difficult business model could have a big payoff
In the long run, taking a more challenging route could result in a greater payoff for Zoox. If you spend too much time and effort perfecting your technology for suburban roads, it might not be able to handle the transition to cities, where there is more demand for ride-hailing services, Levinson said. And building a vehicle designed specifically for autonomy will give customers a safer and more enjoyable experience, he added.
It’s significant that Zoox is making its vehicles with ride-hailing, rather than personal ownership, in mind, because doing so could allow the company to avoid some of the pitfalls faced by electric-vehicle startups. When you sell cars to customers, the profit margin you earn on each vehicle will be so low that you’ll have to make hundreds of thousands of them to turn a profit, Levinson said. Last year, Tesla CEO Elon Musk said the challenges of increasing production volumes almost bankrupted the company.
But when you’re making a vehicle for a ride-hailing service, you can spread its manufacturing costs over the large number of rides that will be given when the average passenger car would be sitting in a garage or parking lot. Zoox won’t have to “optimize every penny out of the supply chain and manufacturing process,” Levinson said, though the company will still be attentive to production costs. Levinson said he is happy with the progress the company has made on its vehicle, which includes testing how well it would fare in a crash.
“Over the last year, we’ve significantly increased our confidence in our ground-up vehicle design and its performance,” he said.
Zoox’s vehicles are already better than human drivers in some ways
A promising but challenging business model needs investors with enough have faith in its long-term prospects to weather years of cash burn. Zoox doesn’t plan to begin giving rides to paying customers until 2021, and if its experience is anything like Waymo’s, that service could debut on a small scale.
Getting investors to believe in Zoox requires the company to prove both that it is among the leaders in autonomous-driving technology, and that the technology will be usable in the near future, Levinson said. He thinks Zoox can meet both of those standards.
“If people decide that this is going to take 30 years, then we could be the best, but people aren’t going to pay for that,” Levinson said. “If it’s going to take five years but we’re 17th-best, then probably people aren’t going to pay for that either.”
Zoox’s vehicles are already better than humans at some parts of driving, Levinson said, like obeying stop signs and speed limits. And their cameras, radar, and lidar sensors give them “super-human” sensing capabilities that allow them to have a 360-degree view of their surroundings, and even what’s behind the objects next to them, at all times. Over the past year, the company’s vehicles have gained the ability to drive in the hardest parts of San Francisco, areas with unprotected intersections, high-speed roads, steep hills, and challenging behavior from pedestrians and other vehicles.
“As far as we know, we’re driving parts of the city that no other company’s even attempting to drive autonomously,” Levinson said.
But Zoox’s vehicles still have weaknesses
But Zoox’s vehicles are still worse than humans at handling what those in the autonomous-driving industry call corner cases, unusual situations that drivers rarely encounter. While a human might be able to quickly identify an erratic driver and understand how to respond to him, it’s very difficult for an autonomous vehicle to predict the next move an irresponsible driver will make, even if it understands what is happening on a moment-to-moment basis, Levinson said.
Corner cases are important, because for autonomous-driving technology to succeed, its creators will need to prove that it is significantly safer than human drivers, which means mastering the tricky scenarios that might come up only once per year for the average car owner.
“If you rush the technology and replace humans killing people with robots killing people at remotely similar levels, I don’t think society is going to accept that, and I don’t think they should,” Levinson said.
The urgency that Levinson and others in the autonomous-driving industry feel to expand the use of self-driving vehicles stems from the more than 1 million people each year who die in vehicle-related incidents, many of which of which involve human error. Zoox and its rivals will have to make sure that in their quest to solve that problem, they don’t create bigger ones.