It’s approaching five years since former Ford CEO Mark Fields announced plans to build automated ride-hailing vehicles (aka robotaxis) by 2021. Since then, Ford has made substantial investments in Argo AI and built up its mobility business. Meanwhile, in that interim, Lyft launched and eventually abandoned its own development program for an automated driving system (ADS). Lyft has now joined Ford and Volkswagen as the third outside investor into Argo AI and Ford automated vehicles will be deployed to carry Lyft passengers later this year in Austin, Texas and Miami.
Argo AI was founded in late 2016 by Bryan Salesky, a veteran of the Google self-driving program and the software lead on the Carnegie Mellon team that won the DARPA Urban Challenge and Pete Rander, formerly of Uber and Carnegie Mellon. A few months later Ford invested $1 billion for a majority stake in Argo and the team based in Pittsburgh took over responsibility for developing the ADS that Ford would use on its future production vehicles.
In mid-2019, Volkswagen joined in, making its own investment in Argo including the contribution of its Automated Intelligent Drive group based in Munich. VW plans to launch its own automated taxi and delivery vehicles based on its electric ID Buzz van in 2025 although early prototypes recently started testing in Germany. Argo is also currently testing Ford vehicles in Miami, Austin, Washington DC, Pittsburgh, Palo Alto and Detroit.
Lyft pursues robotaxis
In 2017, Lyft followed its main competitor Uber in launching development of its own ADS, but it had limited success with the program. Earlier this year, it sold the Level 5 group to Toyota’s Woven Planet unit for $550 million. Fortunately for Lyft, it has been pursuing parallel trajectories on its way to eventually eliminating many if not all of the human drivers from its platform. In addition to developing its own ADS, Lyft has partnered with other developers including Motional (a joint venture between Aptiv and Hyundai Motor Group) and Waymo.
Some of Waymo’s automated Chrysler Pacificas are available through the Lyft app for customers in Chandler, Arizona while Motional and Lyft have been operating paid robotaxi rides in Las Vegas since 2018. Motional and Lyft plan to launch multi-city robotaxi services beginning in 2023 using a vehicle based on the Hyundai Ioniq 5.
The partnership with Argo is the first where Lyft has taken an equity stake in another ADS developer. Argo and Ford have previously announced plans to launch commercial AV services in Miami, Austin and Washington in 2022. As part of the new deal with Lyft, initial robotaxi services are planned to launch in Austin and Miami later in 2021.
Argo takes on more responsibility
According to Argo CEO Bryan Salesky, there will probably be “several dozen” vehicles at first, eventually growing to a larger fleet of about 1,000 vehicles as they expand the operating domain and add more cities . During the development process over the past several years, Argo has worked closely with Ford and now Volkswagen on integrating its ADS into the vehicles. The fourth generation of Ford test vehicles based on the Escape hybrid have been testing in multiple cities since late 2020 and Argo is now preparing its second generation of vehicles in Germany based on the latest prototypes of the VW ID Buzz electric van.
During the testing and development programs Argo crews have been responsible for operating and servicing both the earlier Ford Fusions and now the Escapes. This will continue with the launch of robotaxi services on the Lyft platform. Until recently, the expectation was that Ford would be launching a new transportation service business to operate its AVs for both passenger carrying and deliveries.
However, during a recent Deutsche Bank conference, Ford CEO Jim Farley briefly referenced Argo, saying “My personal opinion is that I think they deserve the opportunity to be a one-stop-shop company, and that they will take on more of the go-to market responsibilities for our AVF (automated vehicle fleet).” Ford and Argo have since declined to comment on what this remark means for Ford’s mobility service ambitions.
However, it does appear that just as Cruise has supplanted Maven as GM’s brand for mobility services, Argo AI may be taking on a similar role for Ford. Given that Argo is currently pursuing a new private funding round ahead of a potential IPO in 2022, it may be that investors are more inclined to put money into a dedicated AV business like Argo than to provide Ford with the capital required to launch an entirely new line of business. This is likely part of the reason why Ford delayed its originally planned 2021 launch into 2022.
With Ford’s recent moves to launch its Ford Pro business unit to support commercial vehicle customers with a range of services, this may well work to the advantage of both companies. Through Ford Pro, the automaker can sell vehicles and provide service and maintenance operations for the Argo fleet while the startup actually handles operations and deployment.
Why Lyft and how will the service work?
Before Lyft divested its Level 5 group, one its claimed advantages was the data that it already had from current human drivers about the usage of ride-hailing. While it was always a dubious proposition about whether data from driver’s smartphones could actually be directly helpful in developing an ADS, that data does have some value. Lyft has in-depth knowledge of where there is demand for ride-hailing services, when people need rides and how much they are willing to pay. Lyft will be sharing this data directly with Argo, something they previously haven’t done with Motional or Waymo. This will help Argo optimize its own algorithms for where to deploy AVs, pickup and drop-off locations and what routes to take in order to optimize utilization of the vehicles.
Perhaps more importantly, Lyft has a lot of safety data from its operations about where crashes have occured, where there have been near misses, average speeds and more. This data that Lyft has been collecting over the past decade of operations will provide a localized safety baseline for Argo. Surprisingly little data is actually collected about the specifics of where crashes happen or nearly happen with very specific location data. Argo will be able directly compare its own safety record against that of Lyft drivers in the areas where they actually operate rather than relying on national aggregate data that may not be specifically relevant. Argo will have a much better understanding of how safe its system actually is than most other ADS companies.
To Argo, this data is so valuable that it is willing to trade a 2.5% stake in the company to Lyft in exchange. Lyft won’t be making a direct monetary contribution to Argo.
While Argo hasn’t been conducting any public pilot programs up until now, it has conducted a broad range of tests with Ford and other partners, especially around delivery services. Argo has also been operating its own in-house ride-hailing service in its test cities for employees, much as Cruise has done in San Francisco for several years and Waymo has done in Mountainview.
When the Argo Escapes start carrying public passengers, they will still have safety operators on board just as Motional and most of Waymos vehicles do. Salesky declines to give a timeline for when they might go driverless, instead insisting that it will happen when the system is ready and deemed safe enough. Argo has also been progressing on its overall system development. Its vehicles now have the ability to detect construction zones and respond accordingly. “I always make the mistake of talking publicly about the AV stuff and I don’t mention enough, the other several hundred people that are building out the web stack,” adds Salesky. “We have a really service-ready, set of API’s that can integrate with any business’s logistics systems.”
Since the partnership with Lyft is not exclusive for either party, it will be necessary for Argo to integrate with multiple platforms as it expands the deployments of its AVs. Lyft is the current partner for ride-hailing services, but Uber has previously made it clear that its platform is open to anyone that wants to deploy AVs. A lot of the near term AV business is also expected to be focused on goods delivery rather than passengers and Salesky says additional partners will be announced for this application.
If Argo is to become the go-to-market partner for Ford and potential Volkswagen automated vehicles, it will have to make substantial investments in its infrastructure and purchasing or leasing vehicles in the coming years. That means it will have to raise substantially more money to be competitive with the likes of Waymo and Cruise which have both had multi-billion dollar funding rounds this year. This partnership with Lyft is just the first of many steps going forward.