Best of LinkedIn: Next-Gen Vehicle Intelligence CW 05/ 06

00:00:00: Brought to you by Thomas Algeier and Frennus, this edition highlights key LinkedIn posts on next-gen vehicle intelligence in weeks five and six.

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

00:00:18: so product teams and strategy leaders don't just react but shape the future of mobility.

00:00:24: You know that phrase shaping the future always sounds so aspirational.

00:00:28: a little bit Yeah,

00:00:29: but looking at the chatter from the last couple of weeks.

00:00:32: It feels like the industry is finding out that shaping the futures just really Really hard work?

00:00:38: it's definitely less dreaming and more engineering right now.

00:00:41: if I had to put A label on what we saw in weeks five and six of twenty-twenty-six i'd say We've moved From The PowerPoint phase Into.

00:00:48: Well, maybe the panic phase.

00:00:49: Or

00:00:50: the execution phase if we're being polite.

00:00:51: Wait...the execution phase?

00:00:52: Execution

00:00:53: Panic might be most accurate term.

00:00:55: actually We are seeing this massive reality check.

00:00:58: All of the glossy presentations about software defined vehicles The SDVs.

00:01:02: They're gone.

00:01:03: And now Now it's engineers and accountants staring at incredible mess involved in building them.

00:01:10: It is a mess that touches everything organizational fractures, a massive trust gap with consumers around AI and well-a sales landscape that is just punishing anyone who gets it wrong.

00:01:22: It's the hangover after the hype.

00:01:24: so we have three big themes to get into today.

00:01:27: first That internal struggle The operating models And why?

00:01:31: The org chart Is basically killing innovation okay.

00:01:34: second the big pivot To ai in this growing Trust Gap.

00:01:37: and finally A really brutal look at the market dynamics from Tesla slump to the strange new geopolitics of manufacturing.

00:01:45: Let's start inside the building then.

00:01:47: for the last few years, The narrative was so simple we're becoming tech companies right?

00:01:50: We just need to hire more software engineers

00:01:52: exactly Just higher enough coders and you'll magically turn into a Silicon Valley giant.

00:01:56: Yeah.

00:01:57: But looking at a post from Augustine Friedel that whole theory seems dead.

00:02:02: on arrival.

00:02:03: It is he made A very short point...the bottleneck isn't talent anymore not really!

00:02:07: The OEMs know how To recruit smart people.

00:02:10: So, what's the problem?

00:02:10: The

00:02:11: problem is the operating model.

00:02:13: You can't just bolt a fast-moving software team onto a rigid hundred year old manufacturing hierarchy and you know expect magic to happen.

00:02:23: so it's the difference between who was doing the work And how the work actually gets done.

00:02:28: Friedel mentioned struggles with funding governance Accountability

00:02:32: structural stuff.

00:02:33: It sounds like they have the players but they're trying to play completely different game.

00:02:39: Well, Friedel

00:02:40: notes a shift.

00:02:41: It used to be.

00:02:42: we will build everything ourselves now because those models are failing.

00:02:46: it's shifting too.

00:02:47: We need to own the control points

00:02:49: meaning you don't have to write every single line of code

00:02:51: Exactly But you absolutely must own the data loops and The big compute choices.

00:02:58: that's the

00:02:58: key in this.

00:02:59: connects directly To what Frank Petsnik was arguing he brought up.

00:03:02: Conway is law.

00:03:03: I feel like this comes up Every single time we talk about legacy auto trying to do software

00:03:07: because it's the fundamental law of the universe for them.

00:03:11: Conway's Law basically states that your software architecture will always, ALWAYS end up mirroring your organizational structure.

00:03:20: So if you're organization is full of silos engine team here Chassis Team there Infotainment over there You get siloed software.

00:03:30: Pet Snick argues these companies are still structured like factories rigid handovers, clear hierarchies.

00:03:37: But software needs to be fluid!

00:03:39: And if you try and build an agile software stack inside that Rigid Factory org chart...

00:03:43: You fail—you just get spaghetti code that doesn't talk to itself….

00:03:46: …but even if you could fix the org-chart —the sheer scope of work seems to be drowning

00:03:51: people.".

00:03:52: I saw a visual from Venetius Tadduzain at The Iceberg.

00:03:54: Oh!

00:03:55: I loved it!

00:03:56: It's so accurate for budget meetings happening right now.

00:03:59: Explain it for us.

00:04:00: The car?

00:04:00: The thing you touch?

00:04:01: The screens?

00:04:02: The apps…?

00:04:02: That's just the tip, maybe ten percent of all

00:04:04: things.

00:04:04: And the other ninety percent is?

00:04:06: Below-the water line.

00:04:07: it's the validation and simulation compliance over-the air infrastructure.

00:04:13: So Zainz point that industry has been obsessed with polishing a tiny tip at iceberg while ignoring massive chunks of ice underwater.

00:04:22: And thats where cost explosion comes from.

00:04:25: If you don't build that underwater infrastructure correctly, none of the shiny features on top actually work.

00:04:31: You can just patch a car like you patch an app.

00:04:34: The validation tail is enormous!

00:04:37: We saw two very different ways to tackle this ice burn problem This week.

00:04:42: On one hand... you've got BMW.

00:04:45: Simon Terraz gave us look at new class.

00:04:47: That's the clean sheet approach.

00:04:49: BMW basically ripping out all old wiring.

00:04:52: They're moving to centralized compute, getting rid of dozens and dozens of little ECUs.

00:04:57: And separating the hardware life cycles from the software ones.

00:05:00: It's a huge bet!

00:05:01: it

00:05:01: is an existential bet on completely new platform.

00:05:03: Now compare that with the update on Volkswagen from Raoul Arredondo.

00:05:07: they are launching something called the China Electronic Architecture or CEA.

00:05:12: Well its really fascinating contrast isn't?

00:05:14: UWs effectively admitting that can solve the iceberg problem fast enough in Germany

00:05:19: So their developing this architecture IN CHINA for China.

00:05:22: But it's not just for China, that is the key!

00:05:24: The goal is what?

00:05:25: Speed.

00:05:26: Speed... Our redondo says they're using this to cut development time by thirty percent.

00:05:31: They are essentially importing innovation velocity from east back into their global stack.

00:05:37: It a huge sign of the center gravity on how car software has shifted.

00:05:42: A humbling moment in German engineers I'm sure.

00:05:45: Just as these companies are figuring out software-defined vehicles, the goalposts outdated.

00:05:57: Yeah, they're pushing this concept of the AIDV—the AI-defined vehicle.

00:06:00: Is it just rebranding?

00:06:02: I mean everything is AI now...I

00:06:03: don't think so.

00:06:04: not in this case.

00:06:05: In a standard SDV —a human rights code that says if X happens then do Y. It's rule based Right!

00:06:10: And an AID V. The AI isn't just a feature like voice command ...it becomes the runtime decision layer.

00:06:15: Its making judgment calls in real time about how the car behaves

00:06:19: Which sounds amazing for performance.

00:06:21: But brings us straight to our second big theme Trust.

00:06:25: Because if the car is making judgment calls, I need to know why it made that call.

00:06:29: That's the multi-billion dollar question!

00:06:33: And we saw a really insightful report from BCG and Bosch which was shared by Mark Brunson, Christopher Dennell & Dennis Raib.

00:06:41: They surveyed thousands of drivers.

00:06:43: The results were pretty stark.

00:06:46: This Is The Trust Gap.

00:06:47: It's huge.

00:06:47: The one stat that really jumped out at me was it over sixty percent of users said they need to know what the driver assistance system is doing

00:06:55: right now, They don't.

00:06:56: most drivers are using trial and error To figure out if their car sees the lane lines or?

00:07:01: The car ahead

00:07:02: which is just incredibly dangerous.

00:07:03: I hope the car sees that truck as not a safety strategy.

00:07:06: Not at all.

00:07:06: And the report highlighted That phantom braking where the car Just slams on the brakes for no apparent reason Absolutely destroys trust.

00:07:12: did the moment that happens people just turn the feature off for good?

00:07:15: So adoption isn't about how many features you can list on a spec sheet.

00:07:20: It's about whether the driver can understand the machine, but

00:07:29: And this is where a post from Daniel L about NVIDIA becomes so relevant.

00:07:33: He talked about their new toolkit Alpamayo.

00:07:35: Right,

00:07:36: This Is The Thinking Out Loud feature Exactly!

00:07:38: NVIDia released the AI tool kit for autonomous vehicles and key differentiator that you can literally read its log file to see it's reasoning.

00:08:09: But

00:08:14: proving that logic is safe must be a nightmare.

00:08:18: Yeltsin-Parlak raised this whole issue around validation, everyone's talking about end to end AI where you just feed data in one and get driving out the other.

00:08:26: but how do feeling, or context.

00:08:30: You can't!

00:08:35: If then statement with a neural network, it's black box.

00:08:38: You can't easily prove that won't do something totally crazy in his specific situation

00:08:43: which is why Muhammad al-Khazandar work on explainable AI Is so critical.

00:08:48: we have to be able to open the Black Box.

00:08:50: But you know while The engineers are fighting With That the regulators have surprisingly stepped up Which is

00:08:55: rare.

00:08:55: usually regulation legs technology by A decade.

00:08:58: but this week We saw massive move.

00:09:00: Voss van den Hoovel and Alex Kendall shared the news about new UN regulation.

00:09:04: Oh yeah, the catchy title Global Technical Regulation for Automated Driving Systems.

00:09:09: It doesn't exactly roll off-the-tum but it is so important!

00:09:13: It applies to level three automation and up And sets a clear harmonized safety benchmark.

00:09:19: The system must be at least as safe As a competent human driver

00:09:23: Company.

00:09:23: human driver seems like a pretty low bar depending on who's driving, but I guess legally that is huge deal.

00:09:28: It's licensed to scale.

00:09:30: it harmonizes rules across Europe Japan the UK and its compatible with US self-certification model.

00:09:38: this is boring but essential paperwork.

00:09:41: lets you sell product globally.

00:09:43: Before we move on from tech, I have to mention this wildcard idea that Jeshwant posted about the user-defined vehicle.

00:09:49: Ah yes!

00:09:50: The idea of telling your AI what tune my suspension for rush hour traffic?

00:09:56: Exactly.

00:09:57: or hey car make the brakes more aggressive because i'm late.

00:10:00: it sounds cool but the liability questions are just

00:10:03: Well, completely.

00:10:04: If I tell the AI to change the suspension and then crash is it my fault?

00:10:08: The AIs' fault...the manufacturers...?

00:10:10: The corporate lawyers would have a heart attack before that ever shipped!

00:10:12: ...I think so but it shows where people's imagination is going.

00:10:15: People want their car to adapt not the other way around

00:10:18: True.

00:10:19: But let's snap back to reality.

00:10:22: We've talked software orgs we've talked AI trust.

00:10:24: Let us look at scoreboard now Because while tech is advancing sales charts for weeks five or six looks turbulent.

00:10:32: Turbulent is a polite way to put it.

00:10:35: For some players, It's more like crashing.

00:10:37: We got some hard numbers from Peter Knapp and Matthias Schmidt about the start of twenty-twenty six.

00:10:42: Let's Start with The Big One Tesla.

00:10:45: Its Rough.

00:10:46: In Europe, Tesla sales were down forty four percent in January And in Norway which Is usually the EV Capital Of The World They Were Down Eighty Eight Percent.

00:10:55: Eighty eight percents.

00:10:56: That Isn't A Slump Thats A Collapse Is Just The Model Y Is Getting Old.

00:11:00: That's part of it.

00:11:01: The lineup is aging, but Schmidt explicitly cited CEO political contagion.

00:11:06: He's arguing that Musk's political activities are actively alienating a key demographic of buyers in Europe.

00:11:12: So the brand is taking a hit because of the persona?

00:11:14: It seems so!

00:11:15: It's a heavy price to pay for social media engagement.

00:11:18: But wasn't just Tesla, BYD took a hit too.

00:11:20: Right they were down about thirty percent Yes...but

00:11:23: that seemed to be largely due to subsidy cuts.

00:11:26: It shows how sensitive these markets still are.

00:11:28: government cash You take away the incentive.

00:11:31: The demand curve just drops off a cliff,

00:11:33: and yet there was one company popping champagne in the middle of all.

00:11:36: this way moe.

00:11:37: their valuation tripled to a hundred twenty six billion dollars

00:11:41: And they're planning expansions in Tokyo and London.

00:11:44: It's really a tale of two markets.

00:11:46: consumer EV sales are wobbling because price and politics.

00:11:50: But the robo taxi model selling miles not cars is proving its economics to investors.

00:11:55: There's also a cultural divide that might explain some of the pain for European manufacturers.

00:12:01: Philip Rash shared some stats on tech acceptance.

00:12:04: Ah, yes!

00:12:05: The Germany versus Asia split?

00:12:07: Exactly He noted only thirty-three percent of Germans like this idea with lots of software and cars.

00:12:12: compare it to eighty four percent people in India.

00:12:14: That is massive strategic trap for German OEMs isn't it?

00:12:18: Their home market—the people they talk about in Munich actively rejects very technology.

00:12:25: So if you only listen to your neighbors, You end up building a car that fails in Shanghai.

00:12:29: Pretty much!

00:12:30: You're designing for skeptics while all the growth is coming from enthusiasts.

00:12:35: Speaking of Shanghai The geopolitics are being flipped on their head.

00:12:39: Osger Nurettin Puskal discussed the new globalization.

00:12:43: This Ford and Geely news.

00:12:46: Well, it's historic.

00:12:47: It really

00:12:48: is.

00:12:48: we're hearing talks about Ford using its European plants to build vehicles that are based on Geely technology.

00:12:55: Just think about that for a second.

00:12:57: For thirty years the model was Western Technology Manufactured in Eastern factories to save on cost

00:13:02: and now it's western manufacturing assets The factories the unions the infrastructure building cars using Chinese intellectual property Chinese technology.

00:13:12: total

00:13:12: reversal?

00:13:13: It's a survival strategy.

00:13:14: As Puskal puts it, the tech leadership is coming from China but physical assets are still in the West.

00:13:19: so partnering is only way to bypass tariffs and close that tech gap quickly.

00:13:24: It's an admission that the tech part of car no longer a Western-dominated field.

00:13:36: And Manglino's point is that this a great proxy for what actually happening on the ground.

00:13:48: Because SDVs need, as he put it... A BRUTAL amount of physical infrastructure!

00:13:53: We're talking about power management chips, Physical layer chips.

00:13:57: You can't software define a car On Vibes!

00:13:59: You

00:14:00: NEED CILICON!

00:14:01: And if TI is selling more automotive ships It means fundamental shift is real.

00:14:06: Regardless whether Tesla had bad month The physical plumbing of the entire industry is being upgraded.

00:14:11: And speaking of plumbing, we have to mention connectivity.

00:14:14: Jenny Lynn Holman and Magnus Leonhardt posted about tests with Scania, Tillia & Erickson.

00:14:20: The

00:14:20: five G corridors?

00:14:22: We often think a Five-G is just fast internet for my phone but for autonomous freight it's critical safety infrastructure...

00:14:28: Is this for big trucks right?

00:14:29: Exactly!

00:14:30: If you have self driving truck It needs be able talk into cloud.

00:14:33: practically zero latency.

00:14:35: They're testing this in Sweden now.

00:14:36: It's just another piece of that iceberg we talked about, you can't have the autonomous truck without the invisible network supporting

00:14:42: it.

00:14:43: So if we pull all these together from the organizational struggles to the AI trust gap To the market reshuffling What is a big headline for your this week?

00:14:53: For me its said The industry has finally realized That software defined isn't feature.

00:14:58: You buy off shelf Its way you exist And existing that way is painful.

00:15:03: You can't just hire a few coders and keep being the traditional car company, you have to change everything your org chart Your validation methods even who you partner with globally.

00:15:13: The dreaming phase is definitely over.

00:15:14: now They actually have to make it work Make it safe and find a way to sell it to a public That is still very skeptical.

00:15:22: Do all of that while competing With the Chinese industry as moving thirty percent faster.

00:15:26: It's gonna be a very long year.

00:15:27: Well we'll be here to track.

00:15:29: If you enjoyed this episode, new episodes drop every two weeks.

00:15:33: Also check out our other editions on electrification and battery technology future mobility in market evolution and commercial fleet insights.

00:15:40: Thanks for listening!

00:15:41: Make sure to subscribe so that don't miss the next one.

00:15:44: See You In The Next Deep Dive.