AD #4338 - U.S. Could Ban Mercedes; GM Spends $11 Billion on EV Write-Offs; BMW Developing Robot Teaching Software
About this episode
Hosts connect manufacturing cost targets and software strategy to the growing role of cloud-connected systems. They explain an ADAS setup that uses a forward-facing camera, sends data to the cloud for processing, then returns it to support lane-keeping up to “Level two plus hands free driving.” BMW’s German plant is also building a modular, open AI platform for humanoid robots’ motor skills and spatial awareness. The discussion adds a compliance workflow that uses software versions, analytics, and diagnostic trouble codes to isolate non-compliant vehicles in real time.
per vehicle production cost
"The company is reportedly targeting a per vehicle production cost of no more than three thousand euros. However, right now its factory costs are about four thousand euros per vehicle."
It means how much it costs, on average, to build one car. Automakers use this to judge whether their factories are getting cheaper and more efficient.
This is the average manufacturing cost allocated to each vehicle coming off the line. When automakers talk about targeting a specific per-vehicle number, they’re trying to improve unit economics even if total factory spending stays high.
JSW Group
"Bloomberg reports that Volkswagen is in advanced talks with Indian conglomerate JSW Group to sell part of its India unit. This is not VW trying to exit India, rather form a partnership so it doesn't have to pay for everything on its own."
JSW Group is a large Indian company with many businesses. The idea here is that Volkswagen could partner with them so it doesn’t have to pay for everything by itself.
JSW Group is an Indian conglomerate that Bloomberg says Volkswagen is in advanced talks with. In this context, it’s relevant because the deal is framed as a way to share costs and reduce what Volkswagen must fund on its own in India.
advanced driver assistance
"Stillantis is turning to mobilize to help develop its advanced driver assistance tech. Next year, selects Stalanta's vehicles will be equipped with mobilized hardware and road management technology."
ADAS is the set of systems that help the car drive—like keeping you in your lane or helping manage speed. It’s not fully autonomous, so the driver usually still has to pay attention.
Advanced driver assistance tech (ADAS) is the suite of sensors and software that helps with driving tasks like lane keeping, adaptive cruise, and automated steering/braking. It’s a step toward higher automation, but it still depends on the driver to supervise.
Level two plus hands free driving
"The system uses a forward facing camera to capture data about the vehicle's environment, which is sent to the cloud, processed, and then sent back to help support driving features from lane keep Assist to Level two plus hands free driving."
This describes a car that can do more of the driving for you—like steering and speed assistance—while you still have to watch the road. You’re not supposed to fully disengage.
“Level two plus” is a marketing-style way to describe advanced driver assistance that can handle more driving tasks than basic lane/acceleration assist, while still requiring the driver to supervise. It’s not full autonomy; the driver must remain ready to take over immediately.
lane keep assist
"The system uses a forward facing camera to capture data about the vehicle's environment, which is sent to the cloud, processed, and then sent back to help support driving features from lane keep Assist to Level two plus hands free driving."
Lane keep Assist is the feature that helps keep you from drifting out of your lane. It uses the car’s sensors to nudge steering when it detects you’re getting off-center.
Lane keep Assist is an ADAS feature that helps keep the vehicle centered in its lane by monitoring lane markings and applying steering corrections. It typically works best on clearly marked roads and within certain speed/conditions.
forward-facing camera
"The system uses a forward facing camera to capture data about the vehicle's environment, which is sent to the cloud, processed, and then sent back to help support driving features from lane keep Assist to Level two plus hands free driving."
This is the camera that looks out the front of the car. It helps the car “see” what’s happening ahead so it can support features like lane keeping.
A forward-facing camera is a sensor mounted near the windshield that looks ahead to detect lane markings, vehicles, pedestrians, and road conditions. In ADAS systems, its data is often combined with other sensors and processed to support driver-assist functions.
Mobile Eye
"The first applications will be in the US, but Mobile Eye also operates in the rest of North America, Europe, South America, and parts of Asia."
Mobileye is a company that makes vision-based driving-assist technology. They’re mentioned here because they help support the kinds of systems cars use to “see” the road.
Mobileye is a technology company known for driver-assistance and computer-vision systems used in many vehicles. The segment notes it operates across multiple regions, tying it to the broader ADAS ecosystem being discussed.
Matt Van Dijk
"Speaking of Stilantis, it just hired new executives to lead two of its most important brands. Matt Van Dijk has been named the new CEO of the RAM brand."
Matt Van Dijk is the person the company picked to lead the RAM brand. A CEO like this can shape what models get prioritized and how the brand is run.
Matt Van Dijk is named CEO of the RAM brand in the segment. This is notable because brand leadership can influence product strategy, marketing, and how quickly new tech and platforms roll out.
Brendan Cody
"The company also named Brendan Cody as the new CEO of Jeep. He was most recently brand president of AutoNation and also held leadership roles at Aston, Martin Canoe and Mercedes Benz USA."
Brendan Cody is the new leader for the Jeep brand. That kind of role can influence which vehicles and technologies the brand focuses on next.
Brendan Cody is named CEO of the Jeep brand in the segment. Like other brand CEOs, he can affect product planning and the pace of adopting new technologies across the brand’s lineup.
modular, open platform AI ecosystem
"On the software side. The factory is developing a modular, open platform AI ecosystem designed to give humanoids the motor skills and spatial awareness needed for complex parts manufacturing."
It means the AI system is built like LEGO blocks—separate parts that can be swapped or reused. The goal is to make it easier to apply the same robot “brain” across different factories and tasks.
This describes an AI software setup built from interchangeable modules that can work together, rather than one tightly coupled system. “Open platform” implies the architecture is designed to be extensible and reusable across different robots, tools, or manufacturing lines.
spatial awareness
"The factory is developing a modular, open platform AI ecosystem designed to give humanoids the motor skills and spatial awareness needed for complex parts manufacturing."
Spatial awareness is how well a robot understands its position and the position of objects around it. It helps the robot pick up and place parts accurately.
Spatial awareness is a robot’s ability to understand where objects and surfaces are in 3D space relative to itself. For manufacturing, it’s crucial for tasks like reaching, aligning parts, and operating tools without collisions or misplacement.
visual language action models
"By blending traditional programming with vision language action AI models, BMW aims to help robots perceive environments and make decisions on their own, complementing ongoing Humanoid Robot testset two of BMW's other plants."
It’s a type of AI that can use what it sees and also use language-like understanding to decide on an action. For robots, that means turning perception into “do this next” behavior.
These are AI models that combine visual understanding (from cameras or other sensors) with language-based reasoning and action selection. In a robotics context, the goal is to translate what the robot “sees” into decisions about what to do next.
diagnostic trouble codes
"identify diagnostic trouble codes and defects by pinpointing which vehicles have specific software versions. This allows them to isolate non"
A diagnostic trouble code is like an error message stored by the car. When something goes wrong, the car records a specific code so you (or a shop) can figure out what system needs attention.
Diagnostic trouble codes (DTCs) are standardized identifiers stored by a vehicle’s onboard computer when it detects a fault. They help technicians and software systems pinpoint what system is malfunctioning (like emissions, sensors, or powertrain components).
DTCs
"identify diagnostic trouble codes and defects by pinpointing which vehicles have specific software versions. This allows them to isolate non [478.0s] compliant vehicles in real time using analytics, part numbers, DTCs, and organized secure, cloud based data."
DTCs are short for diagnostic trouble codes. They’re the specific “error codes” the car stores that can be checked to see whether the vehicle has a problem.
DTCs is the common abbreviation for diagnostic trouble codes. In fleet or compliance workflows, DTCs are used as data points to filter out vehicles that have active faults and may not meet production or regulatory requirements.
non-compliant vehicles
"[478.0s] compliant vehicles in real time using analytics, part numbers, DTCs, and organized secure, cloud based data. [486.8s] Speaker 6: All of these steps ensure vehicles are cleared for production and ready for the road, taking your fleet testing into the future of mobility intrepid control systems driven by your data."
Non-compliant vehicles are vehicles that don’t meet required rules or standards. Here, they’re using software and data to spot which cars aren’t passing the required checks yet.
Non-compliant vehicles are cars that fail to meet required standards—often regulatory, emissions, safety, or manufacturer production requirements. In this context, the workflow uses software versions and analytics to identify which vehicles are not meeting the rules in real time.
cloud based data
"[478.0s] compliant vehicles in real time using analytics, part numbers, DTCs, and organized secure, cloud based data. [486.8s] Speaker 6: All of these steps ensure vehicles are cleared for production and ready for the road, taking your fleet testing into the future of mobility intrepid control systems driven by your data."
Cloud-based data means the information is stored and analyzed on internet-connected servers. That makes it easier to collect results from lots of vehicles and compare them in one place.
Cloud-based data refers to storing and processing information on remote servers instead of only on a local computer. For vehicle compliance and fleet testing, it enables centralized analytics across many vehicles and software versions.
Request an Explanation
Heard something you'd like explained? We'll add it to this episode.
Sign in to request explanations for terms you heard.
Want to learn more?
Browse our glossary for plain-English explanations of automotive terms, jargon, and concepts.
Help improve this episode
See something that's not quite right? Our annotations are AI-generated and can sometimes miss the mark. Click the flag icon on any annotation to suggest a correction.