https://arstechnica.com/cars/2023/04/whats-going-on-with-volkswagens-software-division/ Skip to main content * Biz & IT * Tech * Science * Policy * Cars * Gaming & Culture * Store * Forums Subscribe [ ] Close Navigate * Store * Subscribe * Videos * Features * Reviews * RSS Feeds * Mobile Site * About Ars * Staff Directory * Contact Us * Advertise with Ars * Reprints Filter by topic * Biz & IT * Tech * Science * Policy * Cars * Gaming & Culture * Store * Forums Settings Front page layout Grid List Site theme light dark Sign in vw.os -- In 2019, Volkswagen decided to create a car OS--how's that going? The CEO of CARIAD, VW's software arm, tells Ars what's been going on at work. Jonathan M. Gitlin - Apr 6, 2023 6:50 pm UTC In 2019, Volkswagen decided to create a car OS--how's that going? Enlarge Aurich Lawson reader comments 139 with Share this story * Share on Facebook * Share on Twitter * Share on Reddit In 2019, Volkswagen Group had a bold plan. After proving that it made sense to use a few common architectures to design a varied range of vehicles across multiple different brands, it decided to apply that same approach to software. It set up a new division and moved the entire VW Group's software development under that roof, with a mandate to create a new unified automotive operating system for future VW Group EVs. In fact, the division had to work on three different systems simultaneously. Called E^3 for end-to-end architecture, E^3 1.1 would be the software to run on VW Group's MEB platform for mass-market EVs. Cars using this software are now on the road, including the VW ID.4, Audi Q4 e-tron, and of course, everyone's favorite, ID. Buzz. E ^3 1.2 is destined for more upmarket EVs from Audi and Porsche, using the upcoming PPE platform. And that unified OS, called E^3 2.0, would show up mid-decade in a new, unified platform across the entire VW Group. It hasn't exactly gone smoothly. In 2020 VW replaced Christian Senger as the head of the division--called Car.Software.Org, now called CARIAD--with Dirk Hilgenberg. By 2022, problems with CARIAD's development and buggy software for the launch of the ID.3 and ID.4 EVs saw VW Group fire its chairman, Herbert Diess, along with multiple reports of delays to future group vehicles, including the electric Porsche Macan. The division cost VW Group more than $2 billion last year in the process. Advertisement Still the right move "The idea is still right, but of course, as you have seen, we've gone through the storming phase, a norming phase, now you know, having to deliver in the performing phase, but it's what you would expect," said Hilgenberg, CARIAD's CEO. "So, yes, that was the right decision, taken on very bold and very decisive, while also bearing in mind the big backpack on our shoulders. The task was at a time on the MEB, what we call now the 1.1 software on the ID, and we have gone through those teething problems--the platform stability, with the rollouts, with OTA working fine now with the new updates, pulling vehicles into the workshops, making sure they get upgraded with minimal invasive impact to the customer. So, lessons learned about the brand side as well as CARIAD," Hilgenberg told me. Then there's E^3 1.2. "The other big package in our backpack is the Premium Platform Electronic [PPE], which is being developed as we speak, and we are, I would say, in the final stretch; we can see the finish line," he told me. E^3 1.2 sounds like quite a step up from E^3 1.1. "There's much more complexity in there, of course, for premium vehicles with more functionality, but again, also looking into over-the-air updates, sensor belts massively enhanced--which will be important to deploy the driver assistance function," Hilgenberg explained. "More safety for the customer, more catering for data, which we can use for not only crowd data but also for services, which are anticipating--you know, the road has a curve, there is fog there coming up, there's a big accident--so with connected vehicles you bring information back into the vehicle. The vehicle, the companion, tells you, 'Hey, you got to be careful,' the driver assistance system says, 'You got to take over you have to be alert.' So, it is the evolutionary roadmap on the ADAS stack, really materializing through 1.2 in a different way," he told me. Page: 1 2 Next - reader comments 139 with Share this story * Share on Facebook * Share on Twitter * Share on Reddit Jonathan M. Gitlin Jonathan is the automotive editor at Ars Technica, covering all things car-related. Jonathan lives and works in Washington, D.C. Email jonathan.gitlin@arstechnica.com // Twitter @drgitlin Advertisement Promoted Comments [avatar] spryde Disclosure: I work for Stellantis NA but do not speak for them. What the automotive world really needs is a Windows - an OS that doesn't force every automaker to reinvent the wheel and which can be run on a wide range of hardware. This doesn't really exist across the industry as there are too many variables on ECUs, features, etc. to standardize across the entire industry. The closest we have for industry wide OSes are Yocto Linux and a number of RTOS like VxWorks for non-head unit ECUs. Even then, the software is heavily customized on top to support specific functions. For the head units (read: Infotainment and that big ass screen in vehicles you now see), there is Automotive Grade Linux (supported by Toyota, Subaru, and Suzuki), QNX (use to be supported by nearly everyone), and Android Automotive (too many to list but Stellantis, Ford, Volvo/Geely/Polestar, GM, Honda, and others) have all adopted it. Recently, as this article shows, OEMs have gone out and made their own full vehicle OS that works across ECUs and spaces to work in harmony for all their platforms. That's the theory at least. CARIAD by VW Group is the one covered here. MB.OS by Mercedes is another. ccOS by Hyundai/Kia/Genesis is another I am aware of. Even then, they may adopt Android Auto for the head unit and make their OS interoperable depending on what level of effort they want to go through vs buying off the shelf. Making a generic OS that works with every ECU by every Tier 1 supplier and fit into automotive grade hardware at a cheap per unit price is very, very hard. Limiting your OS to known ECUs is just very hard. As a customer, I'd be weighing Android Automotive tracking my car vs. VW tracking my car. As a auto mfr, I'd be weighing the cost of developing our own OS against that of licensing from google. And I'd be looking at the opportunity cost of letting google monetize location info vs. the huge revenue available to VW from tracking driver locations and building a targeted advertising and marketing platform to take advantage of that finegrained info we've collected. I suspect powerpoints of how great the opportunity is are probably well received. Needs more bitcoin though. The actual strategic risk seems to be relying on a vendor like google for the core software guts of the car. That's worth thinking thru as a VW strategist. As a customer, google already knows where I am thanks to me conveniently carrying a mobile tracking device around. And they're a known quantity as far as giving others access to my attention. VW is a different question: what will they do w/ tracking info they gather? Who will they sell it to? Will employees internally pass around maps of where I've been today and laugh at my grocery choices and amorous misadventures? Every OEM has the capability of pulling massive amounts of data and most if not all do. We do it to diagnose issues on individual vehicles as well as discover fleet wide trends. That data is protected like gold because of the implications of it getting out. But it can be and has been monetized with third parties either on an anonymized and aggregated level or on an individual level with explicit opt-in (think parking services, paying for gas, etc). Which leads to the more interesting bit... Anyone can use Android Automotive and not provide data back to Google if they choose. However, if they choose to use Google Automotive Services (GAS), then you do have to have a data sharing agreement. GAS include GMaps, Navigation, POI, UX/UI, and a number of other things you don't need to roll yourself as the GAS apps already take care of that for you. We decided not to sign the GAS agreement but Ford, GM, and a few others did. That means we had to roll our own UI/ UX, contract maps out with TomTom, build our own CANBUS interfaces, etc. to make UConnect 5 work. April 6, 2023 at 9:05 pm [avatar] Exci Not a big deal, but pretty sure there are more folks working on it in Redmond than Santa Clara. Pretty sure = forwarding this article to a friend working on this and getting that comment in reply. April 6, 2023 at 10:57 pm Channel Ars Technica - Previous story Next story - Related Stories Today on Ars * Store * Subscribe * About Us * RSS Feeds * View Mobile Site * Contact Us * Staff * Advertise with us * Reprints Newsletter Signup Join the Ars Orbital Transmission mailing list to get weekly updates delivered to your inbox. Sign me up - CNMN Collection WIRED Media Group (c) 2023 Conde Nast. All rights reserved. Use of and/or registration on any portion of this site constitutes acceptance of our User Agreement (updated 1/1/20) and Privacy Policy and Cookie Statement (updated 1/1 /20) and Ars Technica Addendum (effective 8/21/2018). Ars may earn compensation on sales from links on this site. Read our affiliate link policy. Your California Privacy Rights | Do Not Sell My Personal Information The material on this site may not be reproduced, distributed, transmitted, cached or otherwise used, except with the prior written permission of Conde Nast. Ad Choices