Image: Waabi
Waabi, the autonomous vehicle startup founded by former chief scientist at Uber’s Advanced Technology Group Raquel Urtasun, announced a new turnkey solution for semi truck manufacturers who want to transform their big rigs into robot trucks.
The product, which is called Waabi Driver, is intended as a full-stack autonomy solution for semi truck OEMs to integrate into their vehicles during the manufacturing process. “It’s very plug-and-play, very sleek, and you know, very lightweight,” Urtasun said in an interview last week, “so that it can be built or directly integrated into the assembly line for OEM integration.”
When it first launched last year, Urtasun said that Waabi’s approach will be to focus on trucking using its proprietary software to automate driving on commercial delivery routes. A little more than a year after coming out of stealth, that effort is now starting to bear fruit.
The company has its work cut out for it. Not only is it a little late to the game in terms of staking out a position in autonomous trucking — a space already crowded with the likes of Waymo, Aurora, TuSimple, Kodiak, and Plus as well as established manufacturers like Daimler, Volvo, and Navistar — but also it is doing so at a time of heightened uncertainty around autonomous technology. Robotaxi startup Argo AI recently shut down after its two main backers, Ford and Volkswagen, decided to pull funding. And other companies have experienced financial hardships, investigations, and high rates of turnover.
“If you look at the industry today, there is a little bit of the winter of AV happening and definitely some negative sentiments,” Urtasun said. She noted that early AV startups promised too much and spent too much and, in the end, couldn’t meet those high expectations. By contrast, Waabi’s AI- and simulation-focused efforts are smaller, cheaper, and more concentrated on specific use cases, like trucking, which gives it an edge.
So how is Waabi’s approach different exactly? According to the company:
The Waabi Driver is an end-to-end trainable system that automatically learns from data, speeding up development dramatically and enabling it to learn the complex decision-making needed for operating on the road safely. This contrasts the traditional approach that is brittle, overly complex, and requires painstaking manual code adjustments.
That training takes place in Waabi World, the company’s simulation program that it cites as a major cost saver. Urtasun has said she doesn’t want to rely on a large fleet of vehicles driving millions of miles on public roads and gathering data in service of training AI systems to drive better and safer than humans. That’s expensive, time-consuming, and ultimately doesn’t capture the seemingly endless number of edge cases that could confuse a self-driving vehicle. Urtasun claims that simulation is cheaper and more efficient than real-world testing.
Waabi wants its autonomy stack to be transparent and not a “black box” that no one can understand. But the same can’t be said for the company’s clients, which Urtasun wouldn’t reveal. “It’s a mystery,” she said with a smile.