Raquel Urtasun, who is chief scientist at Uber Advanced Technologies Group (ATG) and heads the group's unit in Toronto, spoke about the challenges for self-driving development at a Reuters Newsmaker event in New York, Reuters reported.
“Self-driving cars are going to be in our lives. The question of when is not clear yet,” Urtasun said. “To have it at scale is going to take a long time.”
The more cautious tone marks a change from three years ago, when Uber embraced aggressive tactics to turbocharge its autonomous vehicle development in a bid to get more robot taxis on the street driving more miles. The company had been seen as an industry leader in the technology until one of its autonomous SUVs killed a pedestrian in Tempe, Arizona, in March last year.
Urtasun’s comments fall in line with the rest of the self-driving industry, which after much hype and bold promises has tempered expectations and pushed out timelines for deployment. The extreme technical challenges of building cars that can predict human behavior and respond appropriately proved greater than even some of the industry’s brightest minds had anticipated.
The progress of Uber’s self-driving car unit is in the spotlight as the company prepares to kick off its initial public offering this year. Uber, whose losses before taxes, depreciation and other expenses were $1.8 billion last year, has at times spent close to $200 million in a single quarter on its self-driving unit, sources told Reuters.
Uber, last valued at $76 billion in the private market, is seeking a valuation as high as $120 billion in its Initial Public Offering (IPO) and may kick off its investor roadshow before the end of April, Reuters has reported.
“It is true that when you go to an IPO, there is much more of a look into your finances,” said Urtasun. “That being said, again because Uber understands that self-driving cars at scale is not something that’s going to happen tomorrow, they understand the need for the science.”
Urtasun declined to offer any guidance on what mix of human-driven cars and autonomous cars Uber will have in the next 10 years, citing too many uncertainties in the industry.
“What is clear is that in a 10-year timeframe there will be a mix of both (self-driving and human-controlled cars),” she said.
UBER seeking partners
The business of building self-driving cars is extraordinarily expensive and the timeframe to payoff is increasingly uncertain. However, some consider Uber’s ability to successfully navigate the transition to autonomous vehicles as crucial to the company’s long-term financial prospects.
By removing drivers and adding automation, Uber will be able to pocket the full fare that passengers pay and create more efficient routes to move around people and packages.
Uber opened ATG in Toronto in 2017 and named Urtasun, who is also an associate professor at the University of Toronto, as head of the Toronto organization. Uber said last year it would invest more than $150 million in Toronto to grow its self-driving car operations, open an engineering hub and expand its work in artificial intelligence.
After the fatal Arizona crash, Uber removed its autonomous cars from the road, laid off hundreds of test drivers and shuttered operations in Arizona, its testing hub. Uber resumed very limited testing on public roads in Pittsburgh in December.
Uber has recently taken a more collaborative approach, inviting General Motor Co.’s self-driving car unit Cruise and Alphabet Inc.’s Waymo to put their cars on Uber’s ride-hailing network. Urtasun said on Monday that Uber encouraged every self-driving company to add their cars to the Uber network.
“We’ve been very, very open that we would like every other company to be in our (autonomous vehicle) network,” said Urtasun.
A group of investors led by SoftBank Group Corp and Toyota Motor Corp are in talks to invest $1 billion or more into Uber’s self-driving vehicle unit, Reuters reported last month.
Toyota previously invested $500 million to jointly develop self-driving cars with Uber.