Video · 2026-07-01 · 1h 29m · 6 moments

1000 Designs a Day: Neural Concept's Thomas von Tschammer on AI-Native Engineering

✦ AI generated

timeline · colored by role

Mechanism

Thanks to AI, engineers get simulation results in minutes instead of days, which means exploring thousands of design options a year instead of just tens or hundreds.

Von Tschammer traces automotive design from physical prototype-crashing, to CAD/FEA numerical simulation, to AI, noting each shift cut turnaround time and multiplied the number of designs engineers could explore.

transcript

Thomas von Tschammer: thanks to AI, you don't get results in days, but you get them in minutes. And if you do get results in minutes, it means that you don't explore 60 designs a year, maybe 100 designs a year, but now thousands of different options. And that drastically accelerates your development cycles.

Data◆

Jaguar Land Rover went from evaluating 50 aerodynamic designs per day with traditional computational solvers to 1,500 designs per day in production using Neural Concept's AI.

Citing JLR's public NVIDIA GTC presentation, von Tschammer says AI adoption took the automaker's daily aerodynamic design throughput from 50 to 1,500 designs, alongside other clients seeing 80% faster development cycles and better-performing parts.

transcript

Thomas von Tschammer: With AI, and that was what the company a couple of months ago. They went from 50 designs to validity per day to 1,500. Every single day in production. So, you can imagine the level of speedup that it brought to them.

Mechanism

AI-driven engineering workflows are not a black box that spits out a finished design; they act as an assistant that surfaces options while engineers stay in the loop to validate steps and make final trade-off decisions.

Von Tschammer explains that Neural Concept's AI ingests specs, proposes a space of design options, and interacts directly with CAD and simulation tools, but engineers remain the ones guiding and validating decisions because engineering always involves multi-disciplinary trade-offs.

transcript

Thomas von Tschammer: Today what we've seen is that this AI-driven engine workflows are not in black box, right? They are today an assistant to the engineer so that the engineer can take the right data and form design decisions... We are rather in workflows where the AI ingests the spec sheet, understands the requirements, we set up the base model, but this will be done in interaction with the engineer.

gives example · 3

Data

Western and US automakers take 48 to 60 months to develop a new car from concept to launch, while Chinese automakers do it in 18 to 24 months.

Von Tschammer cites a stark gap in vehicle development timelines, noting Western OEMs need 48-60 months versus 18-24 months in China, calling closing that gap the top competitive priority for US and European carmakers.

transcript

Thomas von Tschammer: some numbers to develop a Western Europe or US OEM, it takes time between 48 to 60 months for a new car developed. From the moment they want to launch it to the actual moment it hits the plant and it's being manufactured, right? In China, it's 18 to 24 months.

explains mechanism · 1

Fact

Formula 1 teams have their aerodynamic simulation compute capped in CPU hours, with better-ranked teams from the previous season receiving fewer hours for the next season, deliberately equalizing competition.

Von Tschammer reveals that F1 imposes CPU-hour caps on aerodynamic simulation that scale inversely with a team's prior-season ranking, preventing the wealthiest teams from simply out-computing rivals.

transcript

Thomas von Tschammer: today Formula 1 teams are capped in... the CPU hours that they can run... to run external aerodynamic simulations, right? And what's even more interesting is that depending on your ranking from the previous year, you don't get the same number of hours for the next season. The idea is that you want to try and make it more equal across teams.

Anecdote◆

Engineers have seen AI generate designs they initially would have dismissed as broken, only to discover the designs actually outperform anything humans came up with, forcing them to rethink their own intuitions.

Drawing a parallel to AlphaGo's 'Move 37,' von Tschammer describes engineers reacting with disbelief to AI-proposed designs that look wrong at first glance but turn out to beat human-designed baselines, prompting engineers to revisit their intuitions.

transcript

Thomas von Tschammer: Hey, very impressive. The AI model came up with a design that I would have never thought would be good. If you had shown me this design like this, I would have said, 'Hey, scrap this. This is not going to work.' But actually, those designs are better than anything we could come up with.

extends · 1gives example · 1rebuts · 1

Highlight slides

JLR's AI Leap in Aerodynamic Design✦ from: Jaguar Land Rover went from evaluating 50 aerodynamic designs per day with traditional computational solvers to 1,500 designs per day in production using Neural Concept's AI.Broader Industry Impact✦ from: Jaguar Land Rover went from evaluating 50 aerodynamic designs per day with traditional computational solvers to 1,500 designs per day in production using Neural Concept's AI.Broader AI Impact on Design Cycles✦ from: Jaguar Land Rover went from evaluating 50 aerodynamic designs per day with traditional computational solvers to 1,500 designs per day in production using Neural Concept's AI.AI Designs That Look Broken—But Win✦ from: Engineers have seen AI generate designs they initially would have dismissed as broken, only to discover the designs actually outperform anything humans came up with, forcing them to rethink their own intuitions.An AlphaGo 'Move 37' Moment✦ from: Engineers have seen AI generate designs they initially would have dismissed as broken, only to discover the designs actually outperform anything humans came up with, forcing them to rethink their own intuitions.

Related episodes

Video · a16z Podcast

Is Defense the Next Trillion-Dollar Category? | a16z American Dynamism Summit

→

Article · Interconnects

Kimi K3: The open-weights escalation

→

Video · BG2 Pod

China, China, China. Breaking Down China’s Tech Surge | BG2 w/ Bill Gurley and Brad Gerstner

→