Best of LinkedIn: Next-Gen Vehicle Intelligence CW 43/ 44

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

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

00:00:07: Frennus supports automotive enterprises and consultancies with market and competitive intelligence decoding disruptive technologies, customer needs, regulatory change, and competitive moves.

00:00:16: so product teams and strategy leaders don't just react but shape the future of mobility.

00:00:21: Welcome to the deep dive.

00:00:21: Our mission here is to synthesize the most relevant next gen vehicle intelligence trends we saw popping up across LinkedIn during calendar weeks, forty three and forty four.

00:00:31: And the big picture.

00:00:32: What's the story?

00:00:33: Well, it seems pretty clear.

00:00:34: The industry has definitely passed the talking phase.

00:00:36: The core theme really is this big acceleration of the software defined vehicle moving from, you know, the concept stage to actual real world execution, messy execution sometimes.

00:00:47: That's

00:00:47: it.

00:00:47: Exactly.

00:00:47: It's not about roadmaps anymore.

00:00:49: We're hearing about.

00:00:50: the real friction, the deployment challenge is happening right now.

00:00:53: So we've sort of distilled the insights into four main areas for you.

00:00:57: First, the strategic demands around hardening SDV architectures, then the huge shifts needed in validation, the move towards platform-based cybersecurity, and finally, how AI is rapidly changing user experience and autonomy.

00:01:10: OK, so let's kick off where any big transformation has to.

00:01:14: Strategy and the underlying architecture.

00:01:16: It feels like we're swapping PowerPoint slides for, well, practical decomposition and thinking in snacks now.

00:01:22: And when we talk SDV, it's not just tech upgrades anymore.

00:01:24: It's a whole business model pivot.

00:01:26: Oh,

00:01:26: absolutely.

00:01:27: It's become an existential thing, not just a nice to have.

00:01:31: Juan Han put it quite starkly, saying, for the European auto industry, it's not just a transition, it's a full transformation.

00:01:37: Moving from selling a car to selling a digital service, the SDV is really about reinventing the business around that lifecycle software revenue.

00:01:45: Trinivasa Prasad MN also touched on this, stressing how vital holistic integration is across, I think it was six core tech layers, especially the complex bits like the dynamic software stack and edge and connectivity.

00:01:57: And it's

00:01:57: fascinating how quickly this is moving beyond just the high-end models.

00:02:01: It's not solely a premium play now.

00:02:03: Rajesh Gopi Suresh highlighted the Hyundai Venue Launch in India.

00:02:06: They're actually calling it a debut SDV for that market.

00:02:09: It gets OTA updates for what, up to twenty controllers, and it runs an NVIDIA connected car navigation cockpit.

00:02:16: Yeah,

00:02:16: that shows it's democratizing.

00:02:18: Maybe faster than we thought.

00:02:19: It

00:02:19: really does.

00:02:20: But that speed immediately brings up tricky internal issues, right?

00:02:23: Control ownership.

00:02:25: This is where the business transformation really starts to bite.

00:02:27: Mayday Owls-Bezra made a sharp point about data ownership complexity.

00:02:31: A lot of the critical IP inside the vehicle, it actually belongs to suppliers right now.

00:02:36: So she argues that the vertically integrated OEMs, well, they have a clear edge deploying those valuable edge AI applications because they control that whole data pipeline since they're the cloud.

00:02:45: Hold on, that's a huge point.

00:02:47: If suppliers own the crucial IP, like Mayday suggests, Isn't that a massive ticking clock for traditional OEMs?

00:02:55: I mean, how do you become a software company

00:02:57: if

00:02:58: the very foundation of your future moneymaker, the data, is controlled by someone else?

00:03:02: Precisely.

00:03:03: It raises a fundamental question.

00:03:05: Can the OEMs who aren't vertically integrated even get to those later more profitable STV stages without I don't know, some kind of unprecedented collaboration.

00:03:14: Or new data sharing standards that just don't exist yet.

00:03:17: They're essentially fighting over the raw materials for the digital age.

00:03:20: And the clock is ticking because of the competition.

00:03:23: Felix Warthman really emphasized the internal upheaval needed.

00:03:26: He said incumbent R&D models, you know, built for these long multi-year cycles, they have to be totally redesigned.

00:03:32: Why?

00:03:32: Because the digital natives, especially out of China, they're already setting the pace.

00:03:37: Weekly OTA releases, you just can't match that.

00:03:41: a, a, a, a,

00:03:44: a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a,

00:04:04: And that move, you know, away from domain controllers towards zonal central compute, it's paying off in physical ways too, not just software flexibility.

00:04:12: Melondelica shared a great stat about BMW Group.

00:04:15: Moving to this new structure for the new class, they cut harness weight by about thirty percent.

00:04:19: Thirty percent.

00:04:20: Yeah, that's like six hundred meters of cable gone per car.

00:04:23: Huge savings in cost, complexity, weight, directly from the architecture change.

00:04:28: It turns the vehicle into more of a standardized platform.

00:04:31: Which is why Prasanthigarh Bahalla argued, the real fight now isn't just OEM versus OEM, it's between the semiconductor ecosystems.

00:04:39: NVIDIA, Qualcomm, ARM, even RISCV is gaining ground.

00:04:44: Controlling that foundational compute stack?

00:04:47: Well, that dictates the future.

00:04:49: Silicon is the new strategic high ground.

00:04:51: It's about setting the rules for the whole digital

00:04:53: car.

00:04:54: And this fundamentally changes how we have to think about the network itself, connecting silicon to safety.

00:04:59: Chauncey argues the real magic isn't just the AI.

00:05:02: It's the automotive ethernet fabric underneath.

00:05:04: It's quality.

00:05:05: It's reliability.

00:05:06: He pointed out, you know, in modern break by wire systems, just ten milliseconds of network jitter.

00:05:12: That can mean failure.

00:05:13: The communication path is the safety system now.

00:05:16: And he observed traditional OEMs often still organize networks by domain, kind of siloed.

00:05:21: While the leaders, they're architecting the whole vehicle like one unified compute platform, a deterministic safety critical data center on wheels.

00:05:28: And that need for determinism is pushing standardization efforts.

00:05:33: Makes sense right to manage the complexity and risk.

00:05:35: Robert Day highlighted blueprint efforts like the so offy and auto wear foundation collaboration.

00:05:40: They're key because they help everyone build on shared sort of pre salinated blocks Minimizes integration pain.

00:05:46: good point, and we shouldn't forget the established standards either.

00:05:49: Ahmed Muneer detailed how the autosar Ethernet stack is still vital.

00:05:53: for that high-speed data exchange.

00:05:55: you need for modern a dose and Efficient service-oriented communication.

00:06:00: It's still very relevant.

00:06:01: Okay, so if the dream is weekly OTAs, like we heard, that puts insane pressure on QA, on validation.

00:06:08: How do you compress years of testing in today's?

00:06:10: How are companies actually trying to do this shift left thing for real?

00:06:15: Robert Faye defined it well.

00:06:17: He said testing isn't slow because testers are slow.

00:06:19: Right.

00:06:20: It's slow because feedback comes too late, integration logic breaks, the system's often busted before testing even begins.

00:06:26: And the only way around that lateness seems to be massive virtualization.

00:06:31: Andrew Rohrbeck talked about how virtual ECUs, VECUs, they're totally changing the validation game.

00:06:37: They let you do system-level testing using multi-VECU co-simulation.

00:06:41: So engineers can actually build confidence in the whole integrated system way before they have the physical hardware.

00:06:47: Okay.

00:06:48: Jeremy Dahan even showed synopsis demonstrating a whole virtual car running simulating complex compute environments.

00:06:55: Yeah.

00:06:55: So the tech is there.

00:06:57: But then there's the process problem, the culture clash.

00:07:00: Grinthian Rajan highlighted this friction between dynamic software and, well, static processes.

00:07:05: Yeah, that's a big one.

00:07:06: He talked about the clash between something like the SOP models, which is the nineteen nineties, right?

00:07:10: It assumes development stops at SOP, start a production, the SDV reality where the system really starts living after SOP with OTA updates every Tuesday.

00:07:19: How do you make those two worlds meet?

00:07:20: You really can't, not easily anyway.

00:07:23: That continuous evolution means security has to fundamentally change too.

00:07:27: It can't be a pre-launch checklist anymore.

00:07:29: It has to be a dynamic, real-time thing.

00:07:32: Continuous vulnerability management, that's a run-time job now, not just something you do before SOP and tick a box.

00:07:38: Makes sense.

00:07:39: And this need is driving the rise of platformized security solutions.

00:07:42: Stuff designed to meet R- one fifty-five and ISO two one four three four rules at scale, continuously.

00:07:49: Bob Monkman mentioned the red hat and bicone partnership as an example.

00:07:52: The security has to keep pace with the software updates.

00:07:54: All right, let's shift to the front end.

00:07:56: The stuff customers see and touch.

00:07:58: the AI driven innovations.

00:08:00: This is where it gets real for the user experience.

00:08:03: Augustine Friedel had a pretty strong take here.

00:08:05: He urged OEMs to go beyond just sprinkling AI features around.

00:08:09: The goal should be to become actual AI companies.

00:08:12: Yeah, he was

00:08:13: quite sliding Tesla and X-Ping as examples companies where AI isn't an add-on.

00:08:17: It's like the core DNA of the product.

00:08:19: And you see that playing out in the cockpit already.

00:08:22: Magnus Osberg.

00:08:23: presented the new MBUX virtual assistant avatars from Mercedes-Benz.

00:08:28: They've got the star logo, a human-like one, and even something called little bands.

00:08:31: Little bands.

00:08:32: Yeah.

00:08:33: And it's all powered by generative AI running on their MBOS.

00:08:36: So it's not just voice commands.

00:08:38: It shows a real strategic push towards building a personalized, almost human-like interaction model.

00:08:44: Interesting.

00:08:46: Now moving towards autonomy, Justin Sweetlove noted Addy is obviously the foundation.

00:08:50: L-II adoption is set to grow quite a bit, moving into more mainstream cars.

00:08:54: Right, the building blocks.

00:08:55: But he stressed the biggest hurdle isn't really the tech itself anymore, it's trust.

00:09:00: and user education.

00:09:02: He pointed out that something like forty percent of UK drivers, they don't fully grasp their AD systems when they buy the car.

00:09:09: If people don't get it, they won't use it right or trust it, which kind of defeats the point.

00:09:13: And that trust gap gets way bigger when you talk level three autonomy.

00:09:17: Yeah.

00:09:18: Parthagaswamy really highlighted the inherent contradiction in L three, didn't he?

00:09:22: Yeah,

00:09:22: the paradox.

00:09:22: The driver is legally allowed to be eyes off, but they also have to be ready to take back control instantly if the system asks.

00:09:28: Tricky.

00:09:29: He raised serious questions about whether human reflexes are always good enough for that safe handoff.

00:09:35: Plus, there's still significant legal fog around L-III, especially in the US.

00:09:40: It feels like L-III is stuck in this awkward middle ground.

00:09:43: But while passenger car L-three is complicated, the commercial side is showing some hard numbers.

00:09:48: Tensekip shared compelling data on autonomous trucking.

00:09:51: Oh yeah, the trucking case is strong.

00:09:53: Aurora's platform showing fourteen percent fuel savings just from consistent optimized driving.

00:09:59: And Goldman Sachs projects that market hitting what, eighteen billion dollars by twenty thirty?

00:10:03: Eighteen billion, wow.

00:10:05: Driven

00:10:05: by operating costs potentially plummeting from over six dollars a mile down to maybe under two dollars per mile for driverless trucks.

00:10:12: That kind of cost difference makes the business case pretty hard to ignore.

00:10:16: Absolutely undeniable.

00:10:17: Hashtag tag outro.

00:10:18: So wrapping up these two weeks.

00:10:20: The main takeaway feels like the SDV conversation is fully grown up now.

00:10:24: It's real.

00:10:25: The merging of SDV architecture and AI is speeding up like crazy, and it's forcing established players to adopt totally new R&D methods, chase platform stability, and really rethink their products from the ground

00:10:36: up.

00:10:37: The question isn't if software runs the car anymore, is how fast can you update and validate that software safely and reliably?

00:10:45: And this big shift, away from traditional manufacturing towards platform control, It raises some deep existential questions for everybody in the auto world.

00:10:54: Think about the points we discussed.

00:10:56: Suppliers potentially owning critical in-vehicle IP, like maybe Al's Bizarra noted, and then major chip companies like NVIDIA, not just supplying parts, but building their own robot taxes, as Philip Rash pointed out.

00:11:07: So here's something that you want.

00:11:09: What's the bigger long-term threat for traditional OEMs?

00:11:13: Is it losing control over the vehicle's data?

00:11:15: Or is it losing control over the core compute and semiconductor strategy that defines what the vehicle can even do in the future?

00:11:21: That's the tightrope they're walking right

00:11:37: now.