Best of LinkedIn: Next-Gen Vehicle Intelligence CW 39/ 40

Show notes

We curate most relevant posts about Next-Gen Vehicle Intelligence on LinkedIn and regularly share key takeaways.

This edition offers an extensive overview of the Software-Defined Vehicle (SDV) paradigm, confirming that software, rather than hardware, is becoming the key product in the automotive industry. A central theme is the strategic and organizational transformation required by European automakers to adopt SDV and AI, including embracing a "code-first" development approach, mastering software layers, and adopting DevSecOps practices for continuous improvement and lifecycle management. The sources also highlight the critical importance of cybersecurity in these highly connected vehicles, the role of middleware and open-source platforms like Eclipse S-CORE in managing complexity, and the fierce competition among silicon players (Qualcomm, NVIDIA, Mobileye) to provide the foundational compute platforms for autonomous driving and ADAS features. Furthermore, the content discusses key industry events like IAA Mobility, noting China’s growing market dominance in EVs and the necessity of focusing on customer-defined vehicles and overcoming massive financial and logistical challenges associated with in-house software development.

This podcast was created via Google NotebookLM.

Show transcript

00:00:00: brought to you by Thomas Allguyer and Frenness.

00:00:02: This edition highlights key LinkedIn posts on next-gen vehicle intelligence in weeks thirty-nine and forty.

00:00:08: Frensa 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: Okay, let's unpack this.

00:00:27: Welcome back to the deep dive.

00:00:28: Hey there.

00:00:28: Yeah, we're diving into a stack of insights from calendar.

00:00:31: weeks thirty nine and forty focused pretty squarely on next-gen vehicle intelligence.

00:00:36: And what really jumped out?

00:00:38: Well, the clear, I'd say, overwhelming focus across all the LinkedIn conversations we curated was this ongoing, really complex, and frankly, hugely expensive transformation toward the software-defined vehicle, the SDV.

00:00:51: Right.

00:00:51: It feels like it's moved beyond just a technical transformation now.

00:00:55: It's become a strategic reckoning, hasn't it?

00:00:57: We saw the discussions kind of cluster into three major themes.

00:01:00: First, we probably need to look at the core architecture and what the SDV really means, especially when the financial stakes are getting so high.

00:01:08: Second, we should examine the evolution of ADS, autonomy, and that fierce fight for compute sovereignty that's going on.

00:01:17: Yeah,

00:01:18: that's a big one.

00:01:19: And finally, let's look at the crucial operational challenge.

00:01:22: How the industry is trying to master development speed and security through new ecosystems and validation methods.

00:01:28: It's a lot to juggle.

00:01:30: Definitely.

00:01:31: Our mission for you today is basically to pull out the most important nuggets from these industry experts, give you a shortcut to understanding where the mobility sector is facing its biggest challenges and maybe making its most critical strategic pivots.

00:01:45: Okay, sounds good.

00:01:46: Where do we start?

00:01:46: Let's

00:01:47: start right at the foundation, how the industry is even defining the car now, the whole SDV definition itself.

00:01:53: Okay, here's where the definition gets, well, a bit contentious, right?

00:01:57: The industry seems to be arguing about the fundamental nature of the product itself.

00:02:01: Yeah, you saw that with Robert Faye, for instance.

00:02:03: He gave a pretty hard reality check.

00:02:05: What did he say?

00:02:06: He argued that treating the EE architecture as the product is essentially denial.

00:02:11: Strong word.

00:02:11: Denial.

00:02:12: Wow.

00:02:12: Okay.

00:02:13: Unpack that.

00:02:13: What's fascinating is how clear his pushback was.

00:02:17: Robert Faye states explicitly that EE is merely infrastructure or, you know, plumbing, as he put it.

00:02:23: Just

00:02:23: the plumbing.

00:02:24: Right.

00:02:24: The product is the software, but, and this is key, only if you achieve true abstraction.

00:02:30: Abstraction.

00:02:31: He segments this into the north side, which is the reusable functional platform, and the south side, the hardware abstraction layer.

00:02:39: Basically separating the software functions from the physical hardware.

00:02:41: Okay.

00:02:42: If you don't get that separation, you don't get reuse.

00:02:44: And ultimately, you don't get a true SDV.

00:02:47: No extraction, no SDV.

00:02:48: Simple as that, in his view.

00:02:49: Right,

00:02:50: okay.

00:02:50: And that requirement for abstraction, it ties directly into those radical architectural shifts Garov-Dingrolia highlighted, doesn't it?

00:02:57: Exactly.

00:02:58: Moving from like hundreds of individual electronic control units, ECUs, down to centralized compute zones.

00:03:05: Which we hear about all the time.

00:03:07: And adopting service-oriented architectures, leveraging over-the-air or OTA updates.

00:03:12: Yeah.

00:03:12: That centralization.

00:03:14: That's really the only way to enable continuous monetization to get features out in weeks, not years.

00:03:18: Makes sense.

00:03:19: But we also saw a really necessary counterpoint from Dr.

00:03:22: Gabriel Cyberth.

00:03:24: Oh, yeah.

00:03:25: What was his take?

00:03:26: Well,

00:03:27: he agrees, software central, obviously.

00:03:29: But he argues pretty forcefully that the V, the vehicle itself, that's what still drives sales.

00:03:35: Ah, don't forget the car part of the software-defined car.

00:03:39: Precisely.

00:03:40: He emphasizes that mastering the complexity of a modern car requires excelling in both software and classical vehicle engineering.

00:03:48: Because you know at the end of the day all that revolutionary software still runs on metal.

00:03:51: It's still a car.

00:03:52: The car remains the ultimate complex mobile device.

00:03:56: You can't just treat it like a server rack on wheels.

00:03:58: Okay, so the complexity is immense, but what does that complexity cost when you try to bring it all in-house?

00:04:03: That kind of brings us to the financial reality check.

00:04:06: Why?

00:04:06: Absolutely.

00:04:07: We saw that massive cautionary tale from Matt Damascino regarding Volkswagen's carry-at unit.

00:04:12: Yeah, the numbers there are staggering.

00:04:14: Delays and operating losses estimated to cost VW more than, what was it, sixteen billion dollars?

00:04:20: Sixteen

00:04:20: billion

00:04:21: dollars!

00:04:22: Ouch!

00:04:22: That number isn't just like... a financial footnote, it feels like an existential crisis for that build-it-all-in-house mentality that was so strong a few years ago.

00:04:31: And

00:04:31: the fallout was real, wasn't

00:04:33: it?

00:04:33: Totally.

00:04:34: This failure led directly to a strategic pivot, slowing the EV rollout, pushing core projects like Project Trinity way out into the twenty thirties now.

00:04:44: So the most profound implication here is really the shift in strategy then.

00:04:47: I think so.

00:04:48: Andja Roerbeck analyzed the nuanced VW Rivian joint venture.

00:04:53: Right, the partnership model.

00:04:54: And she showed how these targeted partnerships are now becoming the industry blueprint.

00:04:59: After the carry-at lesson, it seems OEMs realize they have to collaborate on foundational platforms to scale their SDV ambitions.

00:05:06: While still keeping sovereignty over the software that really differentiates them.

00:05:09: Exactly.

00:05:10: You simply cannot afford to rebuild every single piece of the puzzle from scratch internally anymore.

00:05:16: The cost is just too high, the risk too great.

00:05:19: Okay, so that realization what you must master yourself versus what you can share.

00:05:24: that moves us neatly from the foundational architecture to the intelligence applications, right?

00:05:30: Yeah, perfectly automated driving AD and advanced driver assistance systems, ADS.

00:05:36: These are now huge strategic battlegrounds.

00:05:38: Especially around the compute platforms.

00:05:40: and who actually owns the IP.

00:05:41: Definitely.

00:05:42: We saw a great example of this push for strategic ownership in the post from Sharif Hussein about BMW choosing Qualcomm.

00:05:50: Right, for their automated driving super brain, as they called

00:05:53: it.

00:05:54: Yeah, and this isn't just, you know, buying a chip off the shelf.

00:05:56: It's framed as a strategic move to co-develop and actually own the Core AD

00:06:01: stack.

00:06:02: what the industry keeps calling compute sovereignty.

00:06:04: Exactly.

00:06:05: The ability to control the core tech, have flexibility, manage power efficiency across their full vehicle lineup.

00:06:12: Rather than just seeding control to a turnkey supplier, it's about owning your destiny, technologically speaking.

00:06:18: And this choice, it reflects a key insight from Henry Guino, doesn't

00:06:21: it?

00:06:22: It really does.

00:06:22: He argued that the STV crown, it goes to contracts, not just TOPS, you know, Terra operations per second.

00:06:30: Raw power isn't the only game anymore.

00:06:32: So what are the winning models then?

00:06:34: Well, the market seems to be splitting.

00:06:36: You have the comprehensive platform providers like Qualcomm with their digital chassis concept.

00:06:41: Offering

00:06:41: a whole

00:06:42: stack.

00:06:42: Right.

00:06:43: And then you have the high-performance specialists like NVIDIA really focused on that heavy L-three and L-four AI compute.

00:06:50: Interesting.

00:06:51: And he noted mobilized revenue seems to be stalling.

00:06:54: Yeah, he connects that to the market now.

00:06:56: demanding more open, high compute, sensor-fused platforms, which maybe isn't modally's traditional strength compared to these newer, more open approaches.

00:07:05: Okay, speaking of sensors, Dr.

00:07:07: Matthias Traub had a critical observation from the autonomous event.

00:07:11: What's that?

00:07:11: He talked about the fragility of trust in autonomous systems, basically saying redundancy has to be baked in absolutely everywhere, compute, power supply, and crucially, sensors.

00:07:22: Can't have single points of failure when safety is paramount.

00:07:25: exactly and he quoted missy Cummins who stated it really plainly There will be no autonomous car relying on cameras.

00:07:33: only never never never.

00:07:34: Wow,

00:07:35: that's pretty definitive.

00:07:37: a full stop on the vision-only approach.

00:07:39: Yeah

00:07:39: reinforcing the need for robust sensor fusion for safety and importantly for public trust

00:07:44: and the immediate market impact of just ADAS, never mind full autonomy, is already tangible, isn't it?

00:07:50: Frank Turlett pointed this out.

00:07:51: How

00:07:51: so?

00:07:52: Well, ADAS is reshaping auto insurance and collision repair right now.

00:07:57: On one hand, these systems are projected to cut annual loss costs, maybe two to three percent by twenty thirty, just because of fewer low-speed accidents.

00:08:04: Fewer fender benders, good news.

00:08:06: But on the other hand, when accidents do happen, the repairs themselves are becoming far more technically demanding.

00:08:11: Ah, because of all the sensors and calibration.

00:08:14: Precisely.

00:08:15: It requires higher skill, more precision, extensive documentation.

00:08:21: It's a whole new ball game for the repair industry.

00:08:23: Meanwhile, looking globally, we saw ambitions accelerating.

00:08:26: Mady Al's Bizarra shared some insights from the Dubai World Conference.

00:08:30: Dubai is always pushing the envelope.

00:08:31: What's their goal?

00:08:32: Automating twenty-five percent of all transportation by twenty-thirty.

00:08:36: Pretty bold.

00:08:37: Very ambitious.

00:08:38: But what stood out globally wasn't just the driving automation goals.

00:08:42: It was the powerful, immediate integration of advanced AI, specifically generative and agentic

00:08:49: AI.

00:08:49: How's that being used?

00:08:50: To transform business processes.

00:08:53: She mentioned GE Eli connecting virtual design and manufacturing using AI, achieving remarkable speed and cost savings.

00:09:00: The focus has really moved beyond just the driving tabs.

00:09:02: Optimizing the entire development lifecycle with AI.

00:09:05: Interesting.

00:09:06: Yeah, and that focus on radical development speed, that leads us perfectly into our third theme, doesn't it?

00:09:11: It does, because given the complexity and those high-performance demands we just talked about, the traditional way of building automotive software simply won't cut it anymore.

00:09:19: So the conversation around how we build the SDV is almost as revolutionary as the SDV concept itself.

00:09:26: You could say that.

00:09:27: Marcus Rettstat highlighted a major mindset shift that really defines this theme.

00:09:32: Code first.

00:09:33: Code first.

00:09:34: Okay, what does that mean in practice?

00:09:35: It means the old way, where engineers manually maintain detailed specifications in interface documents.

00:09:41: That's kind of out.

00:09:42: Okay.

00:09:43: Instead, the specs and interfaces are derived from the code itself.

00:09:47: Automatically, in many cases.

00:09:48: Ah, so the code becomes the single source of truth.

00:09:51: Exactly, which fundamentally speeds up iteration and, crucially, reduces human error in keeping docs and code aligned.

00:09:59: And this code-first agility, it's being heavily supported by open source initiatives, right?

00:10:04: Yeah, definitely.

00:10:04: We saw the Eclipse Foundation's S-Core project mentioned.

00:10:07: That's that about.

00:10:08: It's modular, designed for safety critical systems, and importantly, it's backed by major OEMs like Mercedes, BMW, and VW.

00:10:16: So

00:10:16: the big players are collaborating on the foundation.

00:10:18: Yes,

00:10:19: and it shows how industry quality standards like AceBice are having to evolve.

00:10:23: Validation is still absolutely key, but it has to be embedded into the continuous development flow, not something tacked on at the end.

00:10:31: And we saw other examples of collaboration targeting this development speed problem too.

00:10:35: Like

00:10:35: what?

00:10:36: Sirat Maitin introduced Horizon, that's the open source platform from Google and Accenture.

00:10:41: Google and Accenture, interesting pairing.

00:10:43: What's Horizon designed for?

00:10:45: Basically, to streamline the heavy lifting stuff like code integration, build processes, test pipelines, the generic stuff.

00:10:53: Okay,

00:10:53: handling the non-differentiating infrastructure.

00:10:56: Exactly.

00:10:56: The goal is simple.

00:10:58: Offload that generic work so OEMs can focus their precious engineering resources on actual product differentiation.

00:11:05: That's where their competitive IP really lives, after

00:11:07: all.

00:11:07: Makes sense.

00:11:08: And this focus on platform development, it connects directly back to the need for both speed and robust security, doesn't it?

00:11:14: Absolutely.

00:11:14: Nithin Bhaskar Alamana discussed turning the vehicle into essentially a cloud application platform.

00:11:20: Using things like service-oriented architecture, OTA.

00:11:23: Right, and principles like GAIX, which focus heavily on controlling data access and ensuring privacy-first data handling.

00:11:31: So secure modular updates.

00:11:33: Yeah, he showcased a remote-controlled speed limiter app as an example, something that can be securely deployed and updated remotely as a modular function without touching the whole vehicle software.

00:11:43: With all that connectivity and continuous updating comes pretty profound risk, right?

00:11:48: Huge risk.

00:11:49: Christian Zimmerman used Cybersecurity Awareness Month to really emphasize the operational reality of those UNESCO regulations, R-one-fifty-five and R-one-fifty-six.

00:11:59: Which mandates cybersecurity management across the life cycle.

00:12:02: Exactly.

00:12:03: And he stressed the need for integrated IT plus OT defense.

00:12:07: It's not just IT security anymore.

00:12:09: It has to cover the entire vehicle lifecycle.

00:12:11: and that regulatory pressure, it requires specific, deep technical solutions down in the weeds.

00:12:16: For

00:12:16: sure.

00:12:16: Chrivan Suryanarayana discussed some foundational requirements.

00:12:20: Like

00:12:20: what?

00:12:20: Things like the necessity for robust approaches for software authenticity and integrity checks, effective key management, and hardware trust anchors.

00:12:28: Especially regarding secure bootloaders, right?

00:12:30: If you don't control the foundational hardware boot process, then

00:12:34: everything running on top of it is potentially vulnerable.

00:12:37: You need that root of trust.

00:12:38: And that security defense?

00:12:40: It needs to match the incredibly rapid release cycle of the SDV we were just talking about.

00:12:47: You can't have security be the bottleneck.

00:12:49: Exactly.

00:12:50: Mostafa Alchemy of VX Labs frames security not as a bottleneck, but actually as an accelerator.

00:12:55: Oh,

00:12:56: so.

00:12:56: Well... by pushing solutions like intrusion detection and prevention systems, IDPS for in vehicle defense and continuous monitoring through things like extended detection and response XDR.

00:13:08: Okay, right.

00:13:09: You can actually keep pace with the massive speed of the SDV release cycle.

00:13:13: The automation required by code first development actually enables the kind of continuous security monitoring needed for compliance and safety.

00:13:20: So it all ties together.

00:13:21: It's clear that everything from the foundation, the architecture to the execution, the development practices, it's all moving towards these integrated systems, rapid iteration and shared open platforms.

00:13:30: Driven by complexity and let's be honest, cost.

00:13:33: Okay, so wrapping this up, what does this all mean for you, the listener?

00:13:38: Well, I think we've seen the SDV narrative really solidify.

00:13:41: It's moved way beyond just a conceptual buzzword.

00:13:45: Now it's about tangible architectural and organizational change.

00:13:48: Change that often involves massive financial bets and these crucial strategic partnerships we talked about.

00:13:54: Right.

00:13:54: And it feels like the industry has really absorbed the lesson of that, uh, that sixteen billion dollar carrier challenge.

00:14:00: Yeah,

00:14:00: that number hangs heavy.

00:14:01: So the focus now is less on if software is defining the car.

00:14:06: It almost entirely on how to execute that definition safely, securely and crucially profitably at scale.

00:14:13: And the trend seems clear.

00:14:14: absolute collaboration to solve those common problems, middleware, validation tools, open source foundations.

00:14:20: While

00:14:20: still retaining sovereignty over the highly differentiating features like ADAS performance and the unique cockpit experience, you collaborate on the plumbing, compete on the features.

00:14:30: And the

00:14:30: constant theme running through all these sources is that massive shift needed in an internal competency, isn't it?

00:14:35: Definitely.

00:14:36: As Stefan Faye noted, particularly for the European industry.

00:14:40: There's a need to move definitively from merely commissioning critical software layers.

00:14:45: from suppliers

00:14:46: to actually mastering them internally themselves, taking ownership.

00:14:50: Exactly.

00:14:51: But if we connect this to the bigger picture, we have to talk about the revenue side.

00:14:54: Ah,

00:14:55: yes, the money.

00:14:56: We saw Augustine Friedel state pretty bluntly that revenues from digital services and features on demand are still being significantly overestimated by many industry stakeholders.

00:15:07: Really?

00:15:08: Even with all the hype.

00:15:09: Yeah, he said willingness to pay and actual conversion rates are way lower than expected even for some of those cutting-edge AI use cases.

00:15:17: That raises the final maybe provocative question for you, the listener, doesn't it?

00:15:21: Go on.

00:15:22: The industry is spending billions, literally billions, to transform the vehicle into a digital platform.

00:15:28: We've seen the enormous cost of that organizational transformation.

00:15:32: But when will the digital revenue finally validate that massive investment?

00:15:36: Right.

00:15:36: When does the SDV turn from the expensive cost center we saw on Theme One into the profitable digital service engine it has always promised to be?

00:15:45: Something

00:15:45: to mull over as you shape your next strategic roadmap.

00:15:48: Definitely food for thought.

00:15:50: So if you enjoyed this deep dive, new additions drop every two weeks.

00:15:53: Also check out our other additions on electrification and battery technology, future mobility and market evolution, and commercial fleet insights.

00:16:02: Thank you for listening and don't forget to subscribe.

New comment

Your name or nickname, will be shown publicly
At least 10 characters long
By submitting your comment you agree that the content of the field "Name or nickname" will be stored and shown publicly next to your comment. Using your real name is optional.