March 23, 2026 | VW cuts capacity despite backlog; Nvidia's Ali Kani
About this episode
Volkswagen is pressing ahead with restructuring, cutting factory capacity and planning job reductions in Germany even as its order backlog grows, aiming to avoid overcapacity amid margin pressure from tariffs and EV spending. ZF is also under strain, exploring asset sales while dealing with more than $11B in debt and further job cuts after scaling back EV projects. Jaguar Land Rover confirms design chief Jerry McGovern’s departure. Mercedes doubles down on EVs via an expanded hybrid/ICE/EV lineup and affordability-focused marketing. NVIDIA’s Ali Kani explains its modular “car-to-cloud” compute stack and why GPU-based autonomy can be efficient and safety-ready.
Volkswagen pushes ahead with factory cuts and job reductions despite a backlog of orders, while ZF hunts for asset buyers as debt mounts. Automotive News’ Urvaksh Karkaria breaks down Mercedes-Benz’s product push. Plus, Nvidia’s Ali Kani talks about the company’s modular automotive approach.
Experian Automotive
"Experian Automotive gives dealerships accurate, actionable data insights into their prospects and customers so you can reach the right consumer with the right message at exactly the right moment."
Experian Automotive is a company that helps car dealers use data to find the right customers. Instead of guessing who to advertise to, it helps dealers target people at the right time with the right message.
Experian Automotive is a data and analytics company that helps dealerships understand and target potential buyers. In practice, it uses customer and prospect data to improve marketing timing and messaging so dealers can reach the right people more efficiently.
Volkswagen
"Today on the show, Volkswagen doubles down on capacity cuts despite an order backlog... Volkswagen's restructuring isn't slowing down, even as orders pile up."
Volkswagen is the company being discussed here. The segment is about VW changing how much it produces to avoid building too many cars for the market.
Volkswagen (VW) is the automaker at the center of this segment, and the discussion focuses on its restructuring and production strategy. The key point is how VW is managing factory output worldwide to avoid overcapacity while dealing with backlog and cost pressures.
custom tech stacks
"...Ali Khani talks about how automakers can build custom tech stacks using the chip giant's tools."
A “tech stack” is the combination of hardware, operating systems, middleware, and applications that work together to deliver a product. In automotive, “custom tech stacks” means automakers tailor the software and compute layers to their specific vehicle features and architectures.
restructuring
"Volkswagen's restructuring isn't slowing down, even as orders pile up."
Restructuring is when a company changes how it operates to get back on track financially. For carmakers, it often means cutting costs and adjusting production plans.
Restructuring is when a company reorganizes its operations—often involving production changes, workforce reductions, and cost-cutting—to improve financial performance. In automaking, restructuring is commonly tied to demand shifts, EV investment needs, and supply-chain pressures.
overcapacity
"...to avoid overcapacity."
Overcapacity means a company can build more cars than people are willing to buy. When that happens, it can hurt profits and force companies to cut jobs or production.
Overcapacity happens when a manufacturer’s production ability exceeds what the market can absorb profitably. It often leads to price pressure, higher costs per vehicle, and tougher restructuring decisions.
EV spending
"VW expects operating returns as low as 4% this year, as tariffs and EV spending squeeze"
EV spending refers to the investment automakers make to develop and produce electric vehicles—such as new platforms, batteries, factories, and software. The segment frames it as a cost pressure that can reduce near-term profitability.
tariffs
"VW expects operating returns as low as 4% this year, as tariffs and EV spending squeeze"
Tariffs are taxes imposed on imported goods, which can raise the cost of vehicles or parts crossing borders. For automakers, tariffs can pressure pricing, reduce demand, and complicate supply-chain planning.
ZF
"Staying in Germany, ZF is looking to shed more assets as it wrestles with over $11 billion in debt."
ZF is a big parts company that makes important stuff for many car brands. Here, they’re dealing with financial trouble and reorganizing parts of their business.
ZF is a major automotive supplier best known for transmissions, driveline components, and advanced driver-assistance systems. In this segment, it’s described as restructuring due to large debt and shifting strategy around EVs and other business units.
driver assistance unit
"That's on top of selling its driver assistance unit to Samsung's Harman for $1.7 billion last December."
A driver assistance unit refers to the hardware and software that enable features like adaptive cruise control, lane keeping, and automated safety functions. Selling this unit indicates ZF is refocusing its resources away from that segment.
EV projects
"It took a $1.8 billion hit last year after scaling back EV projects when demand didn't materialize as expected."
EV projects are plans and spending to build electric-car technology. The story says they cut back because fewer people were buying EVs than expected.
EV projects are a company’s investments in electric-vehicle technology and programs, such as powertrain development, components, and manufacturing plans. The segment says ZF scaled back EV projects when demand didn’t meet expectations, which is a common industry response to shifting EV sales forecasts.
Land Rover Range Rover
"...starting in 2006. He's the guy behind the modern Range Rover Defender and models like the Evoke. His exit com..."
The Range Rover is a luxury SUV from Land Rover. It’s designed to be comfortable for everyday driving but still capable off-road. The podcast brings it up because it’s part of the newer lineup shaped by the same leadership.
The Land Rover Range Rover is the brand’s flagship luxury SUV, combining high-end comfort with serious capability. The podcast links it to the same leadership behind the modern Defender and other newer models, highlighting how product direction can shape multiple vehicles. That’s why it comes up when discussing company changes and future plans.
Land Rover Defender
"...2006. He's the guy behind the modern Range Rover Defender and models like the Evoke. His exit comes before..."
The Defender is a tough SUV made for off-road driving. It’s known for being durable and capable on rough roads. The podcast mentions it because it connects to the newer version of the Defender and the team behind it.
The Land Rover Defender is a rugged, off-road-focused SUV known for its long history and go-anywhere reputation. In the podcast context, it’s tied to the development of the modern “Defender” line and the people behind newer Land Rover designs. That makes it a natural topic when discussing brand strategy and product leadership changes.
battery-electric models
"Why is Mercedes betting on battery-electric models while many other brands retreat from the segment?"
Battery-electric cars are fully electric—no gas engine. They get their power from a battery and an electric motor.
“Battery-electric” (BEV) models run on electricity stored in a high-voltage battery and use an electric motor instead of an internal-combustion engine. The segment frames BEVs as a strategic bet for Mercedes even as the broader EV market faces headwinds.
Mercedes
"Beyond these new AMG models, how does Mercedes plan to reach its goal of 400,000 annual US sales?"
Mercedes is trying to sell more cars in the U.S. by making the lineup easier to choose from and by pushing deals that fit more budgets.
Mercedes-Benz is using a mix of product planning and sales tactics to grow U.S. volume. In this segment, the focus is on hitting a high annual sales target by simplifying choices and improving affordability messaging.
marketing is going to now focus on affordability
"They're also going to get more tactical with their marketing. And the marketing is going to now focus on affordability."
Instead of selling the car as “the best,” the ads will highlight how much it costs to own each month. That can attract people who were priced out before.
Shifting marketing toward affordability typically means emphasizing lower monthly payments, incentives, and value positioning rather than only performance or luxury. This can help a brand regain shoppers who are comparing mainstream competitors.
EV only models
"In terms of the affordability angle, Mercedes also is by essentially developing a comprehensive lineup of ICE, EV and hybrid models rather than EV only models."
“EV only” would mean the company only sells electric cars. Mercedes is saying it doesn’t want to limit itself that way because not everyone is ready for EVs yet.
“EV only models” would mean the brand’s lineup is exclusively battery-electric, with no gas or hybrid options. The segment says Mercedes is moving away from that approach, likely to broaden appeal and reduce the sticker shock of early EV pricing.
ICE, EV and hybrid models
"In terms of the affordability angle, Mercedes also is by essentially developing a comprehensive lineup of ICE, EV and hybrid models rather than EV only models."
ICE is the regular gas-engine car. EV is fully electric, and a hybrid uses both gas and electricity. Mercedes is saying it wants to sell all types, not only electric.
ICE means internal combustion engine (gas/diesel), EV means battery-electric, and hybrids combine an engine with an electric motor/battery. Offering all three helps automakers cover different consumer preferences and infrastructure realities while transitioning toward electrification.
lower the price of these electric variants
"They hope to lower the price of these electric variants."
Mercedes wants its electric versions to cost less. That usually means making the batteries and the car cheaper to build, so more people can afford them.
Lowering the price of electric variants typically involves changes to battery cost, packaging, manufacturing scale, and sometimes incentive strategy. In practice, it’s about making EV trims less expensive so they can compete with mainstream luxury alternatives.
vehicle, consumer and credit insights
"By leveraging powerful vehicle, consumer and credit insights, Experian enables you to identify who is actively in the market before your competitors do."
It means using different kinds of information—about cars people have, who they are, and what financing they might qualify for. Dealers use it to find buyers who are ready to shop.
“Vehicle, consumer and credit insights” describes combining multiple data sources to understand who owns what, who is likely to buy, and what financing/credit profile they may have. For dealers, this can improve targeting and conversion by focusing on shoppers who are actively in the market.
NVIDIA
"NVIDIA has become one of the biggest names in automotive technology, offering everything from AI training tools to in-vehicle computing platforms."
NVIDIA makes computer chips and AI software. Automakers use NVIDIA tech to power the computers in cars that help with advanced driver features like safety systems and navigation.
NVIDIA is a major semiconductor and AI computing company, and it has become influential in automotive through its hardware and software for vehicle computing. In this segment, they’re discussed as providing the “full stack” needed for things like perception, safety features, and in-car AI.
trained the models and we simulated the models
"And we trained the models and we simulated the models. And then the point is, is now we go to a customer and we say, do you need help building your version of this?"
They teach the AI using data, then test it in a virtual environment. That way, the car can handle more situations safely before it’s used on real roads.
This describes the development workflow where AI models are trained on data and then validated in simulation before real-world use. Simulation helps test edge cases and reduce risk before deploying to production vehicles.
full stack
"But then here is the case, a reference partner who obviously is asking us to do the full, like all three computers and full stack."
“Full stack” means having the whole toolkit—from the car-side computing to training and testing. Some customers only use parts of it, depending on what they already have.
“Full stack” here means providing the complete end-to-end set of tools and compute components needed for autonomy development and deployment. The segment contrasts a “full, like all three computers and full stack” approach with customers who only take parts of the pipeline.
end to end pipeline
"So it's a three computer customer, though they didn't take the full stack. They took the portions of that end to end pipeline that made sense for them and they augmented."
An end-to-end pipeline is the full process from learning (training) to testing and then using the software. Wave didn’t take everything—they picked the parts that worked for them and customized the rest.
An “end to end pipeline” refers to the complete workflow from training AI models to running them in the vehicle and validating them through simulation/testing. The segment says Wave used portions of this pipeline that fit their needs and then augmented them.
Waymo
"[851.0s] And so you're kind of seeing just recent news of, you know, different kinds of companies. [856.8s] Waymo is another one they use us in the car. [858.8s] And of course, they use our models to trade."
Waymo is a self-driving company. They use computer systems in their cars to run AI that helps the vehicle understand the road.
Waymo is an autonomous driving company referenced as using the discussed compute/models “in the car.” The transcript implies an end-to-end pipeline: training models (in the cloud) and deploying them on vehicle hardware for real-time driving.
Uber
"[920.1s] Actually, I don't know if you know that there was a good session yesterday with Uber. [925.5s] And I was hosting it, but Sophros, who's the president of their mobility solution, was talking about, hey, when we're building a level four car, [932.6s] if you needed a few extra sensors or a little bit more compute, we would invest that because it's worth it."
Uber is being mentioned as a company working on self-driving or automated ride services. The key idea is they’re willing to add hardware if it makes the system safer.
Uber is discussed here as a partner building or evaluating Level 4 mobility solutions. The transcript frames Uber’s approach as investing in extra sensors or compute when it improves safety and capability.
sensors
"[925.5s] And I was hosting it, but Sophros, who's the president of their mobility solution, was talking about, hey, when we're building a level four car, [932.6s] if you needed a few extra sensors or a little bit more compute, we would invest that because it's worth it. [940.1s] Safety is the most critical thing."
Sensors are the car’s “eyes and ears.” They help the vehicle understand what’s around it, and adding more sensors can make the system more reliable.
In autonomy, sensors (e.g., cameras, radar, lidar) provide the raw perception inputs used to detect objects, estimate motion, and understand the driving environment. Adding sensors can improve coverage and redundancy, which is especially valuable for safety-critical Level 4 systems.
active safety
"But if somebody said, I need active safety, I would say you could probably do that in 50 TOCs."
Active safety is technology that helps prevent accidents or reduce their severity while you’re driving. The speaker is saying it may not require as much compute as full autonomy.
Active safety refers to driver-assistance and automated safety systems that intervene to prevent or mitigate crashes (e.g., braking/steering assist). The speaker contrasts compute needs for active safety versus autonomy, implying different hardware requirements.
performance peaks and drops
"Another perspective I've heard related to the GPU based approach versus NPU based approach is that GPU based performance peaks and drops."
This means the chip might feel fast at first, then slow down later. The question is whether that kind of change is a problem for a car that must drive safely all the time.
“Performance peaks and drops” describes a pattern where compute performance is high briefly, then reduces due to power/thermal throttling or scheduling. The speaker is discussing whether that behavior is acceptable for safety-critical driving.
sustained performance
"how do you address concerns about the sustained performance of GPU based systems given the safety critical nature of driving?"
Sustained performance means the system keeps working at a high level for a long time, not just for a short burst. In a car, heat and power limits can make performance drop if the system isn’t designed for it.
Sustained performance is how well a compute system maintains speed over time, rather than only at peak bursts. In cars, sustained performance is crucial because thermal limits and power management can cause throttling that affects perception and control.
algorithms
"But this application is not like that because the algorithms are running to right now are different than if you weren't engaging autopilot and you just want an active safety."
Algorithms are the car’s software “rules” for interpreting what it sees and deciding what to do. Different driving modes make the car run different software tasks.
Algorithms are the software logic that processes sensor data (cameras/radar/ultrasonics) to make decisions like lane keeping, speed control, and collision avoidance. The speaker is arguing that different functions (autopilot vs active safety vs parked monitoring) require different algorithm workloads.
peak power
"So if I hear you correctly, you're saying that the peaks and drops are... valuable."
Peak power is the highest “energy use” the car’s computer hits when it’s working hardest. The point here is that some designs waste power by staying at high demand all the time.
Peak power is the maximum electrical power the system draws during the most demanding moments. The speaker contrasts architectures that run at a constant peak versus one that scales compute/power to match the current function being performed.
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.