Person
1 episodes · 20 quoted claims
Thomas von Tschammer says that before computer-aided design, car makers could explore only about five to ten physical prototypes per year.
“ultimately you could explore five and ten prototypes a year if you were lucky before you had to go to production”
Von Tschammer says finite element analysis software lets engineers simulate crash effects on a computer instead of crushing physical prototypes.
“software that we call FEA software, finite elements analysis software, that would predict, simulate the effect of the crush”
Von Tschammer says a single crash simulation can take days because it solves physics equations on large computer clusters.
“A single crash simulation can take days to be run because we are solving the equations of physics on the computer.”
Neural Concept uses hybrid training, which combines lower-fidelity numerical simulation data with real measurements such as wind tunnel tests.
“This is what we call hybrid training, essentially, where you combine numerical simulation that can be low fidelity because we are not capturing the physics very well and measurements.”
Jaguar Land Rover published its work with Neural Concept at Nvidia's GTC conference in March 2026.
“published their work with us at the latest Nvidia GTC conference that was back in March this year”
Jaguar Land Rover increased its external aerodynamic design evaluations from about 50 per day to 1,500 per day using AI.
“They went from fifty designs evaluated per day to 1,500 every single day in production.”
Von Tschammer says suppliers designing battery cool plates cut their development cycles by 80% using Neural Concept's AI.
“We have other suppliers as well that are designing battery cool plates to cool the battery that were able to reduce by 80%.”
Von Tschammer says Nvidia CEO Jensen Huang described AI as a five-layer cake with the application layer on top, which is where Neural Concept works.
“the AI as a five layer cake from the foundations to the top, and the lay the last layer at the very top is the application layer”
Neural Concept began in 2019 with a computer-vision-based AI architecture that took 3D CAD geometry as input and learned physics.
“We started in 2019 by building the first AI model architecture based originally on computer vision that could directly ingest three D geometries and learn from physics.”
Von Tschammer expects Neural Concept's pricing to move from per-seat licences toward pricing based on the value delivered.
“ultimately, we're gonna move to a world where pricing is gonna be based purely on value you're able to deliver for companies”
Von Tschammer says Western European and US carmakers take 40 to 60 months to develop a new car, from launch to production.
“it takes them between forty to sixty month for a new car development from the moment they want to launch it”
Von Tschammer says Chinese carmakers develop a new car in 18 to 24 months.
“In in China, it's eighteen to twenty four months.”
Von Tschammer says additive manufacturing (3D printing) promised free-form design about eighteen years ago but proved too expensive to scale in production.
“that was the promise of additive manufacturing eighteen years ago, right? That you could design freely, because then you could print anything”
Von Tschammer predicts the first foundation model for engineering physics will be for aerodynamics, because the physics is similar across components.
“I think it's gonna start with aerodynamics, right? This is the lowest fruit today”
Formula One teams have a cap on the CPU hours they can use for external aerodynamic simulations.
“Formula One teams are capped in to be more specific, the CPU hours that they can run, so CPU hours is the compute to run external aerodynamic simulations”
Von Tschammer says Formula One teams' simulation compute allowance depends on their previous-year ranking, with better-placed teams receiving less.
“depending on your rank from the previous year, you don't get the same number of hours for the next season”
Von Tschammer says Formula One cars are redesigned in fine detail between every race.
“Design of the car is changing between every single race, right?”
Von Tschammer says developing a car from scratch typically costs about a billion dollars.
“A car is typically a billion, right to develop, or from scratch.”
Von Tschammer says AI-led workflows in crash safety, aerodynamics and powertrain can give carmakers a 20% to 40% speedup in the first year.
“that can already lead you to 20%, 30%, 40% speed up on those flagship disciplines”
Von Tschammer says multidisciplinary AI-driven workflows are already producing 50% to 60% speedups in some specific disciplines.
“we are seeing the multidisciplinary automated AI driven workflows happening already for some specific disciplines, and we are seeing 50%, 60% speed up”