Best of LinkedIn: Next-Gen Vehicle Intelligence CW 47/ 48

00:00:00: brought to you by Thomas Alguyer and Frennus.

00:00:02: This edition highlights key LinkedIn posts on Next Gen Vehicle Intelligence in weeks forty-seven and forty-eight.

00:00:08: Frennus supports automotive enterprises and consultancies with market and competitive intelligence, decoding disruptive technologies, customer needs, regulatory change, and competitive moves.

00:00:19: so product teams and strategy leaders don't just react but shape the future of mobility.

00:00:25: Welcome to the deep dive.

00:00:27: Today, we're digging into next-gen vehicle intelligence, looking at the key trends and some pretty surprising insights that leaders were sharing on LinkedIn over calendar weeks, forty-seven, forty-eight.

00:00:37: It was really fascinating couple of weeks, and what we saw wasn't, you know, just vague, future-gazing.

00:00:42: It was more of a confirmation that this shift to software-defined platforms.

00:00:46: It's really moved from theory into concrete execution.

00:00:49: Though we're past the what-if stage.

00:00:51: We are.

00:00:51: We're seeing real maturity in things like validation practices and these ecosystems are forming across the whole supply chain.

00:00:59: Okay, so if the vision is settled, where's the action right now?

00:01:01: Well the focus seems to be on steady progress in four key areas.

00:01:06: You have the foundational SDV strategy, then the really practical needs of validation, the role of ecosystems, and then the core enablers.

00:01:16: Things like energy and AI that make it all work.

00:01:18: We're sort of moving beyond the hype.

00:01:20: Into the hard work.

00:01:21: Exactly.

00:01:22: Into the gritty work of actually making this happen.

00:01:24: Okay, let's unpack that.

00:01:25: Let's start with what so many sources pointed to as the biggest bottleneck.

00:01:29: And it isn't Silicon.

00:01:32: It's organizational.

00:01:33: It's culture and leadership.

00:01:36: There's this... global realization sinking in that you have to start acting like a software company, not just a car company that happens to use software.

00:01:44: But

00:01:44: that's a huge shift.

00:01:46: It is, and the difficulty is in changing these really deeply entrenched incentives.

00:01:50: This is where Philip Rash talked about what he calls the fifty cent dilemma, and it just perfectly illustrates this conflict.

00:01:56: Fifty cent dilemma, remind us.

00:01:57: Well

00:01:57: Rash points out that if you're a department head and your bonus is tied to immediate cost cutting, you're incentivized to save, say, fifty cents per vehicle by choosing a smaller memory.

00:02:07: And if you make two million cars, you've just saved a million euros.

00:02:10: You hit your targets.

00:02:12: The whole organization is built to reward that completion of the hardware.

00:02:15: Right, which sounds great.

00:02:17: Yeah.

00:02:17: Until three years down the line, you want to sell a high margin over the air update.

00:02:22: And you can.

00:02:22: The update fails because that tiny memory chip is full, costing the company hundreds of millions in potential revenue.

00:02:29: The organization is built for a product that's finished, but software is fundamentally never finished.

00:02:34: So the whole culture is misaligned.

00:02:36: aligned with where the money is actually going to be made in the future?

00:02:39: Precisely.

00:02:40: And that forces a really deep strategic pivot, especially for the legacy OEMs in Europe.

00:02:45: Vosheik Dimbara and Konstantin M. Gaul both talked about this.

00:02:49: This need to shift from project thinking to product thinking.

00:02:53: And that's not just jargon, is it?

00:02:54: No, not at all.

00:02:55: A project has a fixed scope, a finite budget, and it ends.

00:02:59: A product.

00:03:00: A product evolves.

00:03:01: It needs continuous funding, continuous ownership for its entire ten-year life on the

00:03:06: road.

00:03:06: And this pivot also changes how Europe competes.

00:03:10: I mean, you look at Chinese OEMs, they have these massive cost advantages and development cycles as fast as twenty months.

00:03:17: You can't win that race on speed or cost if you're a legacy player.

00:03:21: So what's the move?

00:03:22: Well, that's Diem Boucher's argument.

00:03:23: If you can't win the price war, you have to win the desirability war.

00:03:27: Focus on European strengths, craftsmanship, brand heritage, that premium feel that lets you command higher margins.

00:03:34: That's how you afford the software investment.

00:03:36: We saw a pretty high-profile example of this organizational struggle with Cariad, didn't we?

00:03:40: Lucas Tim's analysis was interesting.

00:03:42: It was.

00:03:43: He pointed out that, you know, they did build real things.

00:03:45: They have OTA systems in forty-five million vehicles.

00:03:48: That's a huge achievement.

00:03:50: But they failed at that big goal of vertical integration.

00:03:53: They only hit about thirty-five percent in-house software, not the sixty percent they were aiming for.

00:03:59: So what's the lesson there?

00:04:00: It seems to be an expensive lesson in pragmatism.

00:04:03: Right.

00:04:03: The new leadership under Peter Bosch is refocusing on just delivering and crucially embracing partnerships.

00:04:08: Like the Rivian Joint Venture or the X-Peng deal in China?

00:04:12: Exactly.

00:04:12: It's a clear admission that you can't out-engineer everyone in every single vertical.

00:04:17: Smart partnering is the only way to go.

00:04:19: And this all comes back to a foundational mindset.

00:04:22: Andres Leviton made a really sharp point about this.

00:04:24: He

00:04:25: was very direct.

00:04:26: He basically said, if you don't think software first and data first from the very beginning, you can just forget about data monetization.

00:04:32: Because

00:04:32: monetization is the result, not the strategy itself.

00:04:36: Right.

00:04:36: It's the consequence.

00:04:37: You need the clean architecture.

00:04:39: You need good data hygiene.

00:04:40: You need those CICD habits.

00:04:42: Without clean data flowing constantly, the data is basically useless.

00:04:46: OK.

00:04:47: So let's assume the industry is working on that cultural bottleneck.

00:04:51: Painfully, maybe, but they're working on it.

00:04:53: The next challenge is just pure speed and safety.

00:04:57: So let's move to our second thing.

00:04:59: Validation, testing, and the role of virtualization.

00:05:02: Yeah, this area is doing so quickly because it has to.

00:05:05: The old methods are just breaking under the complexity.

00:05:08: Andrea Leitner really zeroed in on the core challenge, which is that you just can't physically test every possible real-world scenario.

00:05:15: The variety is infinite.

00:05:17: You can't.

00:05:18: Which is why... Scenario-based testing is, you know, it's non-negotiable now.

00:05:23: It's the only way to manage that complexity.

00:05:25: It lets you link requirements to realistic situations and, really importantly, prioritize the critical edge cases for safety.

00:05:33: And it's not just about good engineering practice.

00:05:35: Yeah.

00:05:35: It's becoming mandatory.

00:05:37: It is.

00:05:37: Dr.

00:05:38: Wilhelm Grappner, map this out.

00:05:39: To get continuous compliance with the new UN regulations, R-one fifty-five and R-one fifty-six, you have to completely overhaul your approach.

00:05:47: Who does that look like?

00:05:48: It forces you into what he calls a multi-pillar validation approach.

00:05:52: You need software in the loop, or SIL, hardware in the loop.

00:05:56: hill, vehicle in the loop, and of course the physical proving grounds, all of them have to work together.

00:06:01: The

00:06:01: goal isn't just a one-time pass at launch.

00:06:03: No, the goal is to generate continuous safety evidence for the entire life of the car.

00:06:08: But even with all these amazing virtual tools, Stefan Terlier pointed out that the biggest practical slowdown is still what he calls integration hell.

00:06:17: Ah yes, the reality of hitting the hardware.

00:06:21: The bottleneck isn't the code.

00:06:23: It's trying to merge code from five different tier ones onto three different ships all with their own protocols.

00:06:29: It works fine in simulation.

00:06:31: And

00:06:31: then you put it on the metal and it's chaos.

00:06:32: It's chaos.

00:06:33: Interfaces collide, things break, and you get long, long delays.

00:06:38: So how do you break free from being chained to that slow hardware cycle?

00:06:42: Virtualization.

00:06:43: It's the only way.

00:06:44: Robert Faye called virtual vehicles and digital twins the last escape for legacy OEMs.

00:06:49: It gives you instant iteration, independence from the hardware.

00:06:52: And the data on the efficiency gains is just, it's staggering.

00:06:57: Stefan Inter had some numbers on this.

00:06:59: He did.

00:06:59: He showed that sill testing can take the pre-production evaluation loop.

00:07:03: So the time to get feedback on your code from five months down to just three days.

00:07:07: Five months to three days.

00:07:09: It's a speed multiplier the industry has never even imagined.

00:07:12: It completely changes who you can hire and how you can work.

00:07:15: Let's shift gears from the process to the plumbing, the deep stack, ecosystems, standardization.

00:07:21: We often talk about the flashy interface, but the real complexity is way down in the base layers.

00:07:27: The hidden champions.

00:07:28: Exactly.

00:07:30: Sean Sicki and Augustine Friedle were really clear on this.

00:07:33: The most critical work is being done by companies that many people, even executives, might overlook.

00:07:38: Companies like Green Hill Sockware or QNX.

00:07:43: deep in middleware, what makes them so critical?

00:07:45: Well, take Green Hills.

00:07:46: They provide certified separation kernels.

00:07:49: Think of it as an ironclad security wall inside the computer.

00:07:52: If your infotainment system gets hacked,

00:07:54: that hat can't cross the wall to mess with your brakes or your steering.

00:07:57: Precisely.

00:07:58: It guarantees functional safety at the OS level.

00:08:01: And then you have companies like QNX providing deterministic schedulers, which ensure that critical things like an airbag deploying happen in a guaranteed time window, no matter what else the processor is doing.

00:08:12: If you get that foundational layer wrong, an OTA update goes from a convenience to a massive global recall.

00:08:18: So on top of that solid foundation, there's this big push for API-first design, which Bartow's Bertha highlighted.

00:08:25: What's the benefit of treating the car like a set of APIs?

00:08:28: It's all about decoupling.

00:08:29: Instead of every single team trying to reverse engineer CAN bus traffic to figure out the vehicle's speed, they just plug into a single standard contract, something like vehicle dot speed.

00:08:40: Using standards like Covisa VSS for the data model and ASAM SOVD for diagnostics.

00:08:46: Correct.

00:08:47: It means your AI agent, a web dashboard, a fleet tool.

00:08:50: They all just plug into the same contract.

00:08:52: It slashes integration time and future proofs the whole system.

00:08:56: And we're seeing this software approach totally redefine hardware domains.

00:09:00: James A had a great example with QNX sound.

00:09:02: Yeah, that was fascinating.

00:09:03: They shifted the entire audio processing from hardware into software.

00:09:07: And

00:09:07: the impact.

00:09:08: It was huge.

00:09:08: Cost savings up to ninety eight dollars per vehicle.

00:09:10: Weight reduction of twenty eight percent.

00:09:12: And most importantly, speakers become a commodity and you can push new audio features like noise cancellation over the air years after the car is sold.

00:09:20: That's a perfect example.

00:09:21: It's not just new features.

00:09:22: It's actively reducing cost and weight.

00:09:25: Okay, so let's move to our final theme, the core enablers.

00:09:29: Energy, AI, and global reality.

00:09:32: And let's start at the physics level.

00:09:34: We have to.

00:09:34: We have to talk about energy.

00:09:36: Dr.

00:09:36: Gabriel Cyberth argued that while we all focus on code, the real dark matter shaping the architecture is just electrons.

00:09:44: He said neglecting it compromises the whole stack.

00:09:46: How so?

00:09:47: While he points to the failure of continuing to rely on the old, twelve-volt system while piling on massive compute loads, that lawnmower skeleton, as he called it, just can't handle the power delivery or the heat.

00:09:59: Energy has to be a first-class citizen in every single architectural decision.

00:10:03: And that need for compute is only going up, mostly because of AI.

00:10:07: Matt Demiseno's analysis found that AI is now seen by over a third of experts as the single-top SDV enabler, not just for ADES, but across the whole vehicle.

00:10:16: Which leads to what Justine Litter-Cumthenum called the agent-driven future.

00:10:20: We're moving towards AI agents that orchestrate the whole mobility ecosystem.

00:10:24: They'll handle prediction, optimize energy, maybe even self-heal the software stack.

00:10:29: That's the utopian vision of the seamless global car.

00:10:33: But then Henrique Manguino brings us crashing back to reality.

00:10:37: He argues the global platform dream is basically collapsing.

00:10:40: And his point is critical.

00:10:42: Your SADV strategy can't be one-size-fits-all anymore.

00:10:45: It has to be global vision, regional stacks.

00:10:49: He outlined three completely different realities that you have to engineer for.

00:10:53: Okay, let's break those down.

00:10:54: What are the three zones?

00:10:55: So first, the EU is permission first.

00:10:57: It's all about compliance, R-one-fifty-five, R-one-fifty-six.

00:11:00: You have to prove it safe and secure before you can deploy it.

00:11:03: What about

00:11:03: the US?

00:11:04: The US is liability first.

00:11:06: It's mostly self-certification, which means means you're really designing for legal defense.

00:11:10: You're focused on documenting your safety case to protect yourself in court.

00:11:14: And finally, China is all about data control.

00:11:16: Exactly.

00:11:17: Sovereignty first.

00:11:18: Data has to stay on shore.

00:11:19: You have to use local crypto standards.

00:11:21: You're tailoring everything to their domestic data laws.

00:11:25: Pushing one global code base through those three worlds is just strategically impossible.

00:11:29: Wow.

00:11:30: So the challenge is truly multi-dimensional.

00:11:33: You have to align the culture, the speed of virtualization, the deep stack stability, the physics.

00:11:38: of energy and manage it all across these fragmented global regulations.

00:11:43: So wrapping up, the consensus seems to be that SDV success isn't about writing more code.

00:11:48: It's about getting better at collaboration and creating these incredibly fast loaning loops, almost like a Formula One team.

00:11:55: Right.

00:11:56: And when you combine that cultural resistance, that fifty cent dilemma with the insane pressure from AI and global fragmentation, it's clear the winner will be the company that masters the cultural shift first.

00:12:06: Robert, no.

00:12:07: We'll put it perfectly.

00:12:08: The job isn't managing change anymore.

00:12:10: It's managing continuous reinvention and making that feel energizing, not exhausting.

00:12:15: So I guess the provocative thought to leave everyone with is this.

00:12:19: If your competition can develop a software car in twenty months and AI is accelerating everything, How long can you afford to let organizational inertia slow you down before your hardware excellence just becomes irrelevant?

00:12:33: If you enjoyed this deep dive, new additions drop every two weeks.

00:12:36: Also check out our other additions on electrification and battery technology, future mobility and market evolution and commercial fleet insights.

00:12:42: Thank you for tuning in to the deep dive.

00:12:45: If you want a shortcut to being well informed on vehicle intelligence, go ahead and hit that subscribe button.

00:12:49: We'll see you next time.