Self-driving startup Vay to launch commercial teledriven service in Hamburg in 2022

Vay, a company that is trying to get self-driving car services to market faster by pursuing a teledriven approach, has raised a $95 million Series B round to help it launch its first commercial service in Hamburg, Germany sometime in 2022. The startup plans to offer a transportation service where a user can order a car to be delivered to their door, drive the car themselves to their destination and then just hop out, leaving the job of parking it to the remote operator.

New investors Kinnevik, Coatue and Eurazeo participated in the round alongside existing investors Atomico, La Famiglia and Creandum. Other participants include Project A, Visionaries Club and Signals, as well as former Alphabet CFO Patrick Pichette and Spotify board member Cristina Stenbeck. The funds will also be used to triple headcount, particularly in the engineering department, as Vay scales out its teledrive and driverless tech and launches in more cities in Europe and the U.S., according to the company. 

The Berlin-based startup came out of stealth in September after working for the last two years on its teledriven technology, which allows remote operators, who are Vay employees currently based in Germany, to pilot vehicles to customers. If the launch goes according to plan, Vay might be one of the first to offer a “driverless” commercial service in Europe, but it’s not without serious competition.

Mobileye, an Intel company that is expected to go public next year, recently announced a partnership with rental car giant Sixt SE to launch a robotaxi service in Munich with Moovit, an Israeli transit tech company. While the Mobileye service won’t be commercial at first, the plan is to scale it across Germany and other European countries throughout the decade. Germany is quickly becoming the place to start European autonomous operations because the country adopted legislation in May that would allow driverless vehicles on public roads by 2022 without a human safety operator behind the wheel.

Vay’s service will launch initially as a teledriven one that gradually introduces autonomous features, an approach the company said will create real revenues faster than other industry strategies while ensuring high levels of safety. While most other AV companies are aiming to launch fully autonomous commercial services from the get-go, Vay’s strategy is already quite common in the robotic sidewalk delivery space, where companies like Serve Robotics, Coco and Tortoise are using teleoperations to establish partnerships and enter markets faster than waiting for fully autonomous tech and favorable regulations to settle in.

“We think this is the most pragmatic solution as we are able to put an autonomous driving experience in the hands of customers in a matter of months, allowing us to accumulate highly valuable data that will be utilized in building the most informed and safe autonomous driving service in the market,” a spokesperson for Vay told TechCrunch. “We will be able to start automating parts of our mobility stack as soon as our initial fleet of cars are on public streets.”

Remote operated systems often have trouble with cell latency, which could cause a delay in driver reaction time. Vay said it uses technology that minimizes the chance of latency events, but on the off chance one occurs, the car will have enough autonomy to pull itself over. Other autonomous features, which are being trained on teledriving data, will be launched in a step-by-step approach, the company says.

“For example, we can start with solving simpler features such as driving straight on highways autonomously,” said the spokesperson. “Whereas more complex driving tasks such as intersections, off-ramps of highways and other complex maneuvers would still be done by a teledriver.”

Vay declined to share how large its initial fleet in Hamburg would be, but did say it would be a small fleet accessible to a restricted group of customers. Over time, the company plans to sequentially expand its geofenced operating design domain, add more vehicles to its fleet and offer the service to more customers.

“Our vision is to become as present and available on public streets as popular ride-hailing services, which is why we are already working very closely with the city of Hamburg and other cities in Europe and the U.S. to ensure the best path to public transport integration and endorsement,” said the Vay spokesperson.

Once the service has launched, customers will be able to book a VayCar via the Vay app, which will result in the arrival of a branded electric Kia e-Niros. Vay plans to establish partnerships with others in the auto industry ahead of its expansion into new cities in the coming years, the company said.

Vay still needs to acquire regulatory approval in order to launch, which might be why the company declined to provide a more specific launch date next year. The company is still in its final testing and validation stages with a technical certification authority, said the spokesperson, which is the basis for regulatory approval to get its fleet on roads.

Vay’s vehicles will rely primarily on cameras that enable the teledriver to have a 360-degree view of the environment, but the cars will also have ultrasonic and radar sensors for assisted safety features, like emergency brake systems. The teledrivers will have access to “augmented skills,” which means other road users, like pedestrians or cyclists, will be highlighted in the video stream for the drivers to improve safety.

Vay doesn’t use expensive lidar sensors, which it says allows it to “bring down costs into the thousands rather than tens or even hundreds of thousands for Level 4 autonomous driving companies,” and makes for a highly scalable solution.

Whether or not the unit economics of this will pan out is hard to say. Without lidar sensors, it’s debatable that Vay will even be able to achieve Level 4 autonomy, which SAE describes as a system that can drive itself without needing a human to jump in. And by using teledrivers, each of which will be limited to driving one vehicle at a time, it’s not clear how Vay is bringing down the cost of paying a human to drive.

A Vay spokesperson said that teledrivers are only present for about 33% of a trip, using the example of it taking five minutes to bring the car to the customer, 20 minutes for the customer to drive themselves and another five minutes to telepark the car or bring it to the next customer.

“Hence, we can make this much more affordable than ride hailing where a driver is physically bound to the car before, during and after the ride,” said the spokesperson.

Vay did not respond in time to further requests for clarification as to how this makes for a cheaper service, especially given the 1:1 relationship between teledriver and vehicle.

Vay also says its service will “cost a fraction of existing ride-hailing services.” Given the above example, Vay even estimates its service can be 50% lower than the price of a ride-hailing service, a number that will only further reduce as the company reaches competitiveness with the cost of urban car ownership. It will be interesting to see if scaling this rather unique model can actually lead to the operating cost savings Vay is hoping for.