Great article, somewhat fuzzy on who is responsible for failure, bosses of carmakers?
Maybe future Europe will be just tourism states. Have you seen the conveyor belt of buses from China unloading and loading crowds who came just to take the same picture in front of the Eiffel Tower and go to next attraction. There’s billions more of them to come.
Thanks! Reasons for failure range from systemic to self inflicted. High energy prices, high cost of employment, and strict labor laws that make invention and volatility costly deter from true invention in industry. That said, no one forced BMW to go after hydrogen-powered cars or all of them waiting for state-funded EV infrastructure like charging stations when Tesla just built their own or general unwillingness to invest profits in inventions (EVs, AVs, batteries) and controlling the full stack over improvements of the combustion engine. Also, the sheer ignorance and arrogance to move production and IP to China expecting them to sit idle.
The carrier layer / platform layer distinction is the clearest framing of this I’ve come across - it names what most analyses leave vague.
The severance point is underappreciated: 31 months vs. 7 months means risk avoidance isn’t cultural temperament, it’s architectural. And the CARIAD story is telling - VW’s response to failing at software was to license more, not invest more. The system defaults to protecting what it already knows how to do.
What strikes me most is that the pattern is self-reinforcing. A system optimized for the carrier layer literally cannot see that value has migrated to the platform layer - because everything it measures confirms that it’s winning. The aperture is set by the construction. This extends well beyond automotive - and the fact that the system cannot see the migration is itself part of the problem.
What an awful article. Either you are outright lying or did no research at all.
VW is currently mass producing and operating robo taxis.
I am sure you will not believe this, but it is 100% true and you will find out that it is, if you do just a little bit of research. Writing an entire article about German automotive driving endeavours without even *mentioning* the robo taxis operator owned by Germanys largest automotive company is so absurd, that it is hard to comprehend how even incompetency could have led you there. *Surely* ChatGPT would have known about MOIA. Surely an article about this topic would have deserved a mention of ADMT.
By all objective evidence VW and Tesla are peer competitors, currently scaling their robo taxis operations and manufacturing capabilities. With both having services available on the road right now.
Our argument is not that German automakers aren’t preparing driving-automation offerings. We mention that Mercedes, BMW, and Audi (a VW brand) are all doing so! To quote ourselves:
“These partnerships are impressive, but point to a common frailty. All three German luxury brands have licensed core driving intelligence from the same Chinese startup for the same market for the same reason: they cannot match what domestic Chinese competitors are shipping.”
VW is equally convicted, as it’s relying on Mobileye, a USA/Israel firm for its automated-driving stack. As well it should! VW’s internal software team, CARIAD, was so bad that in 2023 the CEO fired the CEO and the entire board.
The focus of the article is on German luxury brands, but expand our scope to include the mass market, and our point stands.
Your factual claims are mostly just plain false. It is a farce to say you focus on luxury brands, as if „brand“ matters for a project like autonomous driving. It is a VW *group* project.
It is just plainly ridiculous to talk about autonomous vehicles by German OEMs and not mention that VW is operating and mass producing robo taxis. If your claim is that the German OEMs have suppliers, which cooperate with them on key technologies, then this is the least shocking development in automotive history.
Frankly, I find it bizarre that you even try to defend an article not mentioning MOIA or ADMT while talking about autonomous vehicle development by German OEMs. If you had known about them you would have realised that in fact VW Group is competitive with Tesla when it comes to Robotaxi offerings. That VW, through ADMT has global partners which help VW set up *their* platforms (MOIA is one of those platforms) is not an indictment of a failing strategy.
"Volkswagen's experience makes the structural point most bluntly. Its in-house software unit, CARIAD, spent years attempting to build an integrated software stack and largely failed, prompting what the Financial Times described as a complete reset, a pivot from indigenous development to integrating partner technology. VW had also co-funded Argo AI, a full-stack autonomy venture with Ford; when Argo shut down in 2022, VW's response was to partner more, not invest more."
The letters MOIA and ADMT don't appear here, but that's what we're talking about.
Further, Tesla and VW aren't peers in the sense that matters. Tesla owns its autonomy stack and trains it on data from millions of vehicles daily. VW runs Mobileye's stack. Calling them peers is like calling a record label and a recording studio peers because they're both in the music industry. Our whole point is that what matters is who owns the learning system, and on that question, VW and Tesla are not remotely comparable.
Finally, we state explicitly what would falsify our argument: a German OEM with a genuinely owned full-stack autonomy subsidiary. MOIA running on Mobileye's stack is NOT THAT. It's the very pattern we describe!
You did not mention them at all. MOIA is not CARIAD, neither is ADMT. Your implication that CARIAD failed might be true (why did you not mention RVTech, another bizarre omission), but it is an unrelated failure of a different part of the company.
Do you not think it is preposterous to write an article about German OEMs and automotive driving and not mention the flagship project of Germanys largest Automotive group? And als not mention their autonomous driving subsidiary?
Whatever your analysis is, it pretty much does not matter, when your article does not even mention *the most basic* facts about the subject.
Take a step back and imagine reading the article, while knowing very little about Germanys automotive industry. Is there a remote possibility that you would have come away with an accurate picture of the state of the industry, e.g. that VW robo taxis are currently operating?
With "mass-producing robo taxis", I assume you mean the VW ID. Buzz AD where VW plans to produce roughly 500 level 4 capable vehicles this year but hasn't yet. VW's own timeline, presented in September 2025 is Q3 2026 for initial commercial launch with safety drivers in the US; Q4 2026 for a closed-user-group launch with Uber in LA; and Q3 2027 for commercial launch of more than 500 autonomous vehicles. Today they have exactly ZERO in operation. Waymo already operates over 3.000 vehicles today.
MOIA is a VW Group subsidiary that has been operating an electric ridepooling service with human drivers in Hamburg (and Hannover) since 2019. The 450 MOIA vehicles in Hamburg are all-electric minibuses with HUMAN DRIVERS as of today. It is not yet a robotaxi service. They want to use the VW ID. Buzz AD in the future.
RV Tech is the Volkswagen-Rivian joint venture building a shared electrical/electronic architecture for future EVs like the wiring, compute hardware, and OTA update infrastructure. The first VW car to use it will be the ID.Every1 in 2027. It has zero to do with autonomous driving. You seem to be confusing a vehicle's electrical backbone with a self-driving stack.
What is the point of this comment? You are nitpicking irrelevant details and are even wrong about most things. E.g. MOIA is already doing autonomous driving, like Tesla did early on in Texas, with a supervising driver. That Waymo is the tech leader is indisputable, why you are bringing them up is inexplainable to me.
What is also inexplainable is why you claim that I think RVTech is developing autonomous driving. I brought them up, because they are an incredibly strong argument in OPs thesis (sadly, he is not particularly informed about the German automotive industry and thus couldn’t deliver that point) and *he* brought up CARIAD and their failure.
It is just so incredible to me that you are defending an article where OP did not know the single most important autonomous vehicle project of the single biggest automotive OEM in Germany. Do you seriously believe that ignoring MOIA, the largest project, by the largest OEM while taking about autonomous driving in Germany is would lead to a good article?
My objections to this article are not on personal grounds. They are on the basis of factual errors and ignorance of the most important evidence.
(In case you think that this is about telling people that VW is doing awesome and will outdo everyone. This is not so. The qualities OP ascribes to German engineering are not real, at least not in Comparison to China, which does these things far better. Software failures are common to the entire German industry and are structural and essentially unfixable, bad education system, poor pay for local software developers, complete reliance on outsourcing, importing of low skill but high credential labour from India and the list goes on. VW will not fail because of autonomous driving, but because making worse products for 30% more costs, in an economic structure were innovation is punished and governments desperately need to extract money from corporations and their employees to fund large welfare programs, does not lead to a globally competitive business model.)
Great article, somewhat fuzzy on who is responsible for failure, bosses of carmakers?
Maybe future Europe will be just tourism states. Have you seen the conveyor belt of buses from China unloading and loading crowds who came just to take the same picture in front of the Eiffel Tower and go to next attraction. There’s billions more of them to come.
Thanks! Reasons for failure range from systemic to self inflicted. High energy prices, high cost of employment, and strict labor laws that make invention and volatility costly deter from true invention in industry. That said, no one forced BMW to go after hydrogen-powered cars or all of them waiting for state-funded EV infrastructure like charging stations when Tesla just built their own or general unwillingness to invest profits in inventions (EVs, AVs, batteries) and controlling the full stack over improvements of the combustion engine. Also, the sheer ignorance and arrogance to move production and IP to China expecting them to sit idle.
The carrier layer / platform layer distinction is the clearest framing of this I’ve come across - it names what most analyses leave vague.
The severance point is underappreciated: 31 months vs. 7 months means risk avoidance isn’t cultural temperament, it’s architectural. And the CARIAD story is telling - VW’s response to failing at software was to license more, not invest more. The system defaults to protecting what it already knows how to do.
What strikes me most is that the pattern is self-reinforcing. A system optimized for the carrier layer literally cannot see that value has migrated to the platform layer - because everything it measures confirms that it’s winning. The aperture is set by the construction. This extends well beyond automotive - and the fact that the system cannot see the migration is itself part of the problem.
What an awful article. Either you are outright lying or did no research at all.
VW is currently mass producing and operating robo taxis.
I am sure you will not believe this, but it is 100% true and you will find out that it is, if you do just a little bit of research. Writing an entire article about German automotive driving endeavours without even *mentioning* the robo taxis operator owned by Germanys largest automotive company is so absurd, that it is hard to comprehend how even incompetency could have led you there. *Surely* ChatGPT would have known about MOIA. Surely an article about this topic would have deserved a mention of ADMT.
By all objective evidence VW and Tesla are peer competitors, currently scaling their robo taxis operations and manufacturing capabilities. With both having services available on the road right now.
I don’t think you read our article closely.
Our argument is not that German automakers aren’t preparing driving-automation offerings. We mention that Mercedes, BMW, and Audi (a VW brand) are all doing so! To quote ourselves:
“These partnerships are impressive, but point to a common frailty. All three German luxury brands have licensed core driving intelligence from the same Chinese startup for the same market for the same reason: they cannot match what domestic Chinese competitors are shipping.”
VW is equally convicted, as it’s relying on Mobileye, a USA/Israel firm for its automated-driving stack. As well it should! VW’s internal software team, CARIAD, was so bad that in 2023 the CEO fired the CEO and the entire board.
The focus of the article is on German luxury brands, but expand our scope to include the mass market, and our point stands.
Your factual claims are mostly just plain false. It is a farce to say you focus on luxury brands, as if „brand“ matters for a project like autonomous driving. It is a VW *group* project.
It is just plainly ridiculous to talk about autonomous vehicles by German OEMs and not mention that VW is operating and mass producing robo taxis. If your claim is that the German OEMs have suppliers, which cooperate with them on key technologies, then this is the least shocking development in automotive history.
Frankly, I find it bizarre that you even try to defend an article not mentioning MOIA or ADMT while talking about autonomous vehicle development by German OEMs. If you had known about them you would have realised that in fact VW Group is competitive with Tesla when it comes to Robotaxi offerings. That VW, through ADMT has global partners which help VW set up *their* platforms (MOIA is one of those platforms) is not an indictment of a failing strategy.
We certainly DO mention them:
"Volkswagen's experience makes the structural point most bluntly. Its in-house software unit, CARIAD, spent years attempting to build an integrated software stack and largely failed, prompting what the Financial Times described as a complete reset, a pivot from indigenous development to integrating partner technology. VW had also co-funded Argo AI, a full-stack autonomy venture with Ford; when Argo shut down in 2022, VW's response was to partner more, not invest more."
The letters MOIA and ADMT don't appear here, but that's what we're talking about.
Further, Tesla and VW aren't peers in the sense that matters. Tesla owns its autonomy stack and trains it on data from millions of vehicles daily. VW runs Mobileye's stack. Calling them peers is like calling a record label and a recording studio peers because they're both in the music industry. Our whole point is that what matters is who owns the learning system, and on that question, VW and Tesla are not remotely comparable.
Finally, we state explicitly what would falsify our argument: a German OEM with a genuinely owned full-stack autonomy subsidiary. MOIA running on Mobileye's stack is NOT THAT. It's the very pattern we describe!
You did not mention them at all. MOIA is not CARIAD, neither is ADMT. Your implication that CARIAD failed might be true (why did you not mention RVTech, another bizarre omission), but it is an unrelated failure of a different part of the company.
Do you not think it is preposterous to write an article about German OEMs and automotive driving and not mention the flagship project of Germanys largest Automotive group? And als not mention their autonomous driving subsidiary?
Whatever your analysis is, it pretty much does not matter, when your article does not even mention *the most basic* facts about the subject.
Take a step back and imagine reading the article, while knowing very little about Germanys automotive industry. Is there a remote possibility that you would have come away with an accurate picture of the state of the industry, e.g. that VW robo taxis are currently operating?
With "mass-producing robo taxis", I assume you mean the VW ID. Buzz AD where VW plans to produce roughly 500 level 4 capable vehicles this year but hasn't yet. VW's own timeline, presented in September 2025 is Q3 2026 for initial commercial launch with safety drivers in the US; Q4 2026 for a closed-user-group launch with Uber in LA; and Q3 2027 for commercial launch of more than 500 autonomous vehicles. Today they have exactly ZERO in operation. Waymo already operates over 3.000 vehicles today.
https://www.electrive.com/2026/03/06/vw-commercial-vehicles-ramps-up-pre-series-production-of-the-autonomous-id-buzz-ad/
MOIA is a VW Group subsidiary that has been operating an electric ridepooling service with human drivers in Hamburg (and Hannover) since 2019. The 450 MOIA vehicles in Hamburg are all-electric minibuses with HUMAN DRIVERS as of today. It is not yet a robotaxi service. They want to use the VW ID. Buzz AD in the future.
RV Tech is the Volkswagen-Rivian joint venture building a shared electrical/electronic architecture for future EVs like the wiring, compute hardware, and OTA update infrastructure. The first VW car to use it will be the ID.Every1 in 2027. It has zero to do with autonomous driving. You seem to be confusing a vehicle's electrical backbone with a self-driving stack.
So how about scaling the tone down a bit?
What is the point of this comment? You are nitpicking irrelevant details and are even wrong about most things. E.g. MOIA is already doing autonomous driving, like Tesla did early on in Texas, with a supervising driver. That Waymo is the tech leader is indisputable, why you are bringing them up is inexplainable to me.
What is also inexplainable is why you claim that I think RVTech is developing autonomous driving. I brought them up, because they are an incredibly strong argument in OPs thesis (sadly, he is not particularly informed about the German automotive industry and thus couldn’t deliver that point) and *he* brought up CARIAD and their failure.
It is just so incredible to me that you are defending an article where OP did not know the single most important autonomous vehicle project of the single biggest automotive OEM in Germany. Do you seriously believe that ignoring MOIA, the largest project, by the largest OEM while taking about autonomous driving in Germany is would lead to a good article?
My objections to this article are not on personal grounds. They are on the basis of factual errors and ignorance of the most important evidence.
(In case you think that this is about telling people that VW is doing awesome and will outdo everyone. This is not so. The qualities OP ascribes to German engineering are not real, at least not in Comparison to China, which does these things far better. Software failures are common to the entire German industry and are structural and essentially unfixable, bad education system, poor pay for local software developers, complete reliance on outsourcing, importing of low skill but high credential labour from India and the list goes on. VW will not fail because of autonomous driving, but because making worse products for 30% more costs, in an economic structure were innovation is punished and governments desperately need to extract money from corporations and their employees to fund large welfare programs, does not lead to a globally competitive business model.)