Best of LinkedIn: Next-Gen Vehicle Intelligence CW 21/ 22

00:00:00: Brought to you by Thomas Allgaier and Freinus, this edition highlights key LinkedIn posts on NextGen vehicle intelligence in weeks twenty-one and twenty two.

00:00:09: Furnace supports Tier One automotive suppliers with early stage market validation for their R&D efforts by combining secondary research direct OEM expert interviews and facilitated customer meetings.

00:00:20: You can find more info in the description.

00:00:22: Right,

00:00:22: and today we are extracting the top next-gen vehicle intelligence trends that we saw across LinkedIn over the last two weeks.

00:00:30: Yeah exactly if you manage a product roadmap or lead an engineering team In the mobility space this deep dive is entirely for you.

00:00:37: We're

00:00:38: cutting straight to strategic moves dictating where all those billions of R&D dollars Are actually flowing.

00:00:44: no fluff at all.

00:00:45: So to start us off, the most glaring signal right now is that the very definition of automotive value is shifting.

00:00:51: It's moving away from mechanical performance and going straight into a war over software platforms.

00:00:57: Oh absolutely!

00:00:58: The whole ecosystem control thing... Augustine Friedel actually shared a great example of this recently.

00:01:02: What did he say?

00:01:03: He argued that OEMs absolutely must own what he calls the operating layer.

00:01:09: It's that invisible layer between the user's intent, The AI execution and well...the monetization.

00:01:16: Right because if you don't own That Layer You Don't Own The Customer Relationship.

00:01:21: Yutong Wang had a really sharp observation about that too.

00:01:23: specifically looking at Apple CarPlay.

00:01:26: Oh right!

00:01:26: The shrinking role of carplay.

00:01:27: Exactly for years automakers just basically surrendered the dashboard to apple.

00:01:32: but now OEMs are fighting back because they realize if you give away the user interface, You completely lose the data loop.

00:01:38: You

00:01:38: lose the ability to learn from your driver?

00:01:40: Yeah

00:01:41: and you definitely lose The ability to sell them future services.

00:01:43: And the money They're spending To take that control Back is just staggering.

00:01:47: I mean look at Stellantis.

00:01:49: Oh man yeah Honey Ruta Dorley & Dr.

00:01:51: Wilhelm Gropner both posted about their pivot.

00:01:54: Yeah, they're basically trying to turn into a massive platform company.

00:01:57: it's a sixty billion euro bet on software-defined vehicles.

00:02:01: Sixty billion!

00:02:02: That's just wild.

00:02:03: and that specifically the STLA One Platform right?

00:02:06: The one launching in twenty twenty seven...That's

00:02:08: the one They want to consolidate over thirty different models onto that single eight hundred volt SDV architecture.

00:02:14: Wow okay.

00:02:15: so mass market brands like Stellantis can rely.

00:02:21: What about luxury brands?

00:02:23: Well, luxury brands have a totally different existential crisis right now.

00:02:26: Paul K actually highlighted this huge warning sign for Maserati.

00:02:30: Oh I saw that!

00:02:31: The fifty-seven percent sales drop.

00:02:33: Yeah it's a clear warning.

00:02:35: they need to redefine their identity beyond just engines.

00:02:38: How do you digitize the century of Italian mechanical heritage?

00:02:42: It is tough.

00:02:43: But then you look at Ferrari and Stefan Voldman noted there new facility loose

00:02:48: I think it's called.

00:02:49: Right, Loose!

00:02:49: It is a huge signal that they are entering the software era but doing so while desperately trying to preserve that brand DNA

00:02:56: Which brings up really good point.

00:02:58: This kind of feels like early smartphone wars doesn't?

00:03:00: Oh

00:03:01: for sure.

00:03:01: But

00:03:02: if every OEM wants to be a walled garden platform aren't North American OEMs risking alienating drivers?

00:03:09: Like ditching cloud features from these embedded AI platforms?

00:03:12: Well Robbie Illich pointed out preserving long-term control is exactly why they are doing it.

00:03:18: They want absolute control over the vehicle's brain.

00:03:21: But as a user, if I get in my car... ...I just want my digital life to sync perfectly.

00:03:27: If i'm forced use of proprietary embedded system.

00:03:30: that feels like huge step backward.

00:03:32: It might feel like that.

00:03:33: yeah but counter argument all about latency and functional safety.

00:03:37: Embedded platforms process locally on cars silicon.

00:03:40: So zero latency

00:03:41: Exactly zero latency and total privacy, which is critical when you tie AI to actual driving functions.

00:03:48: But uh...you can't be a software platform company without fundamentally changing the hardware underneath

00:03:54: Right!

00:03:54: You can't run embedded AI on a legacy spaghetti bowl of wiring.

00:03:58: Satya Pradhan detailed this perfectly When he talked about shift toward zonal architecture.

00:04:02: Yes

00:04:02: tearing out those AD scattered electronic control units

00:04:05: Exactly ripping out eighty ECUs and replacing them with one central brain.

00:04:10: It's a massive physical change,

00:04:11: And managing that central brain is pushing the industry toward a massive open source movement.

00:04:16: Matt Crowley just shared huge update on that front.

00:04:19: Oh!

00:04:19: The LG & Google announcement

00:04:21: Yeah They unveiled single system-on ship Android automotive solution for STVs... ...and they are open sourcing it in June.

00:04:30: twenty twenty six Opersourcing

00:04:32: it in twenty twenty sixth.

00:04:34: Wow So, Andy Key raised a really strategic question about this which is if Android Automotive OS is open-sourced who actually controls the remaining value layer?

00:04:43: It's great because the physical layer foundations are changing too.

00:04:47: Clyde Bullock and Aaron Leiba shared details on GM's Connectivity Hub module.

00:04:51: right integrating all of the radios in antennas natively Yeah!

00:04:54: And Ralph Dynline highlighted ZF's chassis two point O system.

00:04:57: it unifies braking steering damping under one central software control.

00:05:02: But wait, if the base layers you know things like Eclipse SDV or Zephyr RTOS are becoming open source commodities aren't OEMs basically just assembling open-source Lego bricks?

00:05:11: It

00:05:12: kind of looks that way yeah.

00:05:13: So

00:05:13: where's the defensible moat If everyone has the same foundational code?

00:05:17: well The Moat is in the execution.

00:05:18: it's your APIs having certification confidence and modular execution.

00:05:23: Anyone can use the Open Source Bricks but not Everyone Can Build a system That flawlessly passes automotive safety certifications at eighty miles an hour.

00:05:31: Ah, okay.

00:05:32: That makes sense but managing this centralized open source architecture requires just massive engineering power

00:05:40: which is exactly the bottleneck

00:05:42: right.

00:05:42: and Lucas Tim shared a stat that blew my mind.

00:05:45: he said engineers waste up to eighty percent of their time on compliance paperwork instead of actual engineering.

00:05:51: eighty percent.

00:05:51: yeah.

00:05:52: human beings literally cannot write and document code fast enough for these central compute platforms.

00:05:57: so the industry's trying But not AI in the dashboard.

00:06:01: Peter Bosch posted about Kariad running a three hundred and fifty person AI hackathon.

00:06:06: Oh, wow!

00:06:06: Three-hundred and fifty people.

00:06:08: Yeah they were building agents specifically to automate Volkswagen's developer workflows.

00:06:12: That's huge.

00:06:13: And Bruno Finco mentioned something similar.

00:06:15: He said MoviDot is running hands on workshops for engineers To actively deploy AI agents onto automotive processes.

00:06:21: It was basically massive rescue mission For bogged down engineers.

00:06:25: but Dr.

00:06:26: Dirk Alexander Molitor issued a pretty stark warning about this.

00:06:30: What does he say?

00:06:31: He warned that AI simply cannot fix broken architectures, you can't just throw AI at twenty year old legacy code.

00:06:38: it needs AI native software and really clean APIs to actually work right.

00:06:43: if the system isn't designed To be read by machine The agent is going to generate garbage but when It works we are seeing A I mood into real world driving too.

00:06:54: like how

00:06:55: Andres C. Yordike highlighted here location reasoning, they are bringing deep Location Intelligence directly into agentic AI.

00:07:04: So the car isn't just processing a camera feed, it understands its geographical context.

00:07:09: Yes fascinating but you know Magnus Osberg made a really grounded point about all this tech.

00:07:13: he reminded everyone that transformations like MBOS are actually about agile methods.

00:07:17: Yeah It's cultural shift not just a text shift

00:07:19: Exactly!

00:07:20: Its people first leadership getting mechanical engineers and software developers to speak the same language.

00:07:25: So let me ask you this, with all of these tech AI right now is less a robot chauffeur for the consumer and more super interned by the bog down automotive engineer?

00:07:35: Yeah exactly.

00:07:36: It's ghost writing test scripts checking safety compliance so that humans can actually design their car.

00:07:41: but it brings up a good pivot.

00:07:43: How does all those new architecture feel to the driver?

00:07:46: Right...the

00:07:47: consumer perspective?

00:07:48: Well, that leads us straight into the booming cockpit domain controller market.

00:07:52: The CDC.

00:07:53: The localized supercomputer behind the dashboard.

00:07:56: Angela Slocrentis and Martin Cedarfist shared some crazy data on this...

00:08:00: Oh!

00:08:00: The shipment projections?

00:08:01: Yeah.

00:08:02: CDCs are hitting twenty three million shipments in twenty twenty five And they project it will reach forty.

00:08:07: nine point five million by twenty thirty.

00:08:09: That's a sixteen point five percent compound annual growth rate.

00:08:12: It

00:08:12: is exploding But the design philosophy around those CDCs is finally shifting.

00:08:17: Multi-car stand observed that Mercedes Benz was actually returning to physical control

00:08:21: Really?

00:08:22: Less screens?

00:08:23: Yeah,

00:08:23: moving away from screen maximalism and trying to find a true interface balance.

00:08:27: Thank goodness Tim Barlett noted something similar.

00:08:30: He said True Progress Is Actually Invisible Tech.

00:08:33: he used the Audi Q III's integrated cockpit as benchmark for that.

00:08:37: Right because when tech demands your attention it becomes safety issue And there is a serious safety reality check happening right now.

00:08:44: Oh for sure!

00:08:45: Steve Greenfield pointed out that Tesla's full self-driving claims just do not hold up to rigorous scrutiny, the gap to true autonomy is still massive.

00:08:54: But then you have companies like BYD taking a completely different approach to building trust.

00:08:59: Alvin Foo shared that BYD is fully covering ADS damages to accelerate adoption.

00:09:04: They are paying for crashes if the system isn't engaged.

00:09:07: Exactly,

00:09:07: it's a huge market signal to build consumer confidence.

00:09:10: That's a massive bet.

00:09:12: But Venusia's Ted Design and Ricky Kwok argued that safety Isn't just about the car.

00:09:16: they said safety as a system property

00:09:19: meaning It requires physics first roadside infrastructure not Just smarter onboard software.

00:09:24: The environment has to be smart too right

00:09:26: which makes me wonder.

00:09:27: I mean we're adding Wi-Fi six Google Gemini cameras, predictive co-pilots.

00:09:31: Are we building safer cars or just highly entertaining crash proof living

00:09:36: rooms?

00:09:36: That is the ultimate question right?

00:09:39: Distraction is deadly but the goal of agentic AI isn't entertainment it's contextual awareness.

00:09:46: A living room doesn't know its about to hit black ice But a physics aware AI does.

00:09:52: That's a really good distinction and it brings us to a final, really provocative thought-to-leave you with drawing on an insight from Sean Sia.

00:09:59: Oh I love this point!

00:10:00: Right

00:10:01: the tech world is totally obsessing over cloud AI right now.

00:10:04: but OEMs might actually be Yeah,

00:10:08: because they have the fleets.

00:10:09: They have factories and complex hardware software integration that tech giants just do not possess

00:10:15: Exactly!

00:10:15: They have massive streams of real world operating data.

00:10:19: Silicon Valley can't easily replicate millions of metal machines moving at eighty miles an hour.

00:10:23: It's

00:10:24: a huge structural advantage

00:10:26: it really is.

00:10:27: If you enjoyed this episode new episodes drop every two weeks.

00:10:31: Also, check out our other editions on electrification and battery technology.

00:10:35: Future mobility in market evolution...and commercial fleet insights!

00:10:39: Thanks so much for diving with us today.

00:10:40: Yeah thank you And don't forget to subscribe.

00:10:42: So never miss an insight.

00:10:44: We will catch ya next time.