It’s long Away from the pedal bike to Formula 1. But that is exactly the quantum leap that the AI-based startup Neural Concept and its co-founder and CEO Pierre Baqué have achieved in just six years.
In 2018, the company’s fledgling software helped create the world’s most aerodynamic bike. Today, four out of ten Formula 1 teams use a further development of the same technology.
Over time, Baqué’s company secured deals with aerospace suppliers such as Airbus and Safran, earning a $9.1 million Series A raise in 2022. Now employing 50 people, Switzerland-based Neural Concept is working towards a Series B round, while its software is helping historic Formula 1 teams like Williams Racing find their way back to the top of the world’s premier form of motorsport.
However, while Formula 1 cars rely on 1,000 horsepower hybrid V6 engines, Baqué’s first practical application of the technology was through human power.
Pedal power
In 2018, Baqué studied at the Computer Vision Laboratory of the École Polytechnique Fédérale de Lausanne, working on applying machine learning techniques to three-dimensional problems.
“I was put in touch with this man who was leading this team and designing the sixth or seventh generation of bicycles, and their goal was to break a world bicycle speed record,” Baqué said. That guy was Guillaume DeFrance and the team was IUT Annecy from Université Savoie Mont Blanc. The cycling team had already gone through half a dozen iterations of bike designs.
“I came back to him two days later with a form that looked almost like that of the current world record holder,” Baqué said. Impressed, the team asked for further iterations. The result, according to Baqué, was “currently the most aerodynamic bicycle in the world.”
That’s a strong statement, but one that is backed up by several world records from 2019. We’re not talking about wing-shaped downtubes or knobby rims to reduce air resistance. This bike is fully enclosed and the cyclist sweats in a composite cocoon, completely protected from the wind.
The core technology is a product called Neural Concept Shape or NCS. It is a machine learning-based system that makes aerodynamic suggestions and recommendations. It fits into the broad field of Computational Fluid Dynamics (CFD), where highly skilled engineers use advanced software packages to perform three-dimensional aerodynamic simulations.
CFD is much faster than creating physical models and throwing them into wind tunnels. Yet it is also hugely systems intensive and relies largely on people making good decisions.
At its core, NCS helps engineers avoid potential aerodynamic pitfalls while pushing them in directions they might not have thought of. In “Co-Pilot Mode,” an engineer can upload an existing 3D shape, providing a starting point, for example.
NCS will then enter its neural network to suggest improvements or modifications and possible paths in a 3D game to choose your own adventure. The human engineer then selects the most promising proposals and puts them through further testing and refinement to achieve aerodynamic glory.
Not just “cheating the wind”
NCS is not only useful in racing, but also in the automotive and aerospace industries. “The path to widespread acceptance in such companies is slow,” Baqué said of working in the more conservative aerospace industry. “So we started working more closely with the automotive industry, where the needs are a little more urgent and can change quickly.”
Neural Concept secured contracts with several global suppliers, including Bosch and Mahle. Aerodynamics is becoming increasingly important in the automotive world as manufacturers look for increasingly aerodynamic vehicles that offer the greatest possible range from a given size battery pack.
But it’s not just about cheating the wind. NCS is also used in the development of battery cooling plates, which, when made more efficient, can keep the battery at its optimal temperature without consuming too much energy. “Enormous gains can be made,” said Baqué, meaning even greater reach.
While the ultimate testing ground for these technologies is always the road, the ultimate laboratory is Formula 1. A global motorsport phenomenon since 1950, Formula 1 is currently experiencing an unprecedented wave of popularity.
The power of Netflix
The Netflix series “Formula 1: Drive to Survive” has brought the excitement of Formula 1 to a whole new audience. While this series focuses on the politics and drama between teams, success on the track has much more to do with aerodynamics. This is where Neural Concepts comes into play.
Baqué started watching Formula 1 before Netflix was even a twinkle in Reed Hastings’ eyes. “I’ve always watched, since the days of David Coulthard and Michael Schumacher.”
Today, parts developed using his company’s software run in this premier class of global motorsport. “It’s a great sense of achievement,” said Baqué. “When I founded the company, I felt it was a milestone. Not just in Formula 1, but just to have parts designed with the software on the road. And yeah, every time that happens, it’s a great, great feeling.”
Formula 1 is also an extremely mysterious sport. Of the four teams Neural Concept works with, only one was willing to be named as a customer, and even that team remained pretty tight-lipped about the entire process.
Williams Racing is one of the most traditional teams in Formula 1. Founded in 1977 by racing legend Frank Williams, the team was so dominant in the 1990s that it won five constructors’ world championships, including three in a row from 1992 to 1994.
But like most sports, success for Formula One teams is cyclical and Williams is currently in a period of rebuilding. The team finished last in the 2022 season and only rose to seventh place last year.
NCS is one of the tools helping Williams regain its competitive edge. “We are using this technology in a variety of ways, some of which will improve our simulation, and other methods we are working on will help produce better CFD results the first time,” said Hari Roberts, head of aerodynamics technology at Williams.
Again, CFD simulations are time-consuming and costly, which is exacerbated by Formula 1 regulations that limit a team’s ability to test. Physical time in the wind tunnel is very limited, and each team has a limited budget for computing time that they can use to develop their cars.
Any tool that can help a team get their aerodynamic designs into shape quickly is a potential advantage, and NCS is very fast indeed. Baqué estimated that a full CFD simulation, which would normally take an hour, would take just 20 seconds through NCS.
And because NCS does not perform actual physics-based calculations, but instead makes AI-driven guesses based on its network of aerodynamic insights, it is largely exempt from Formula 1’s draconian restrictions. “Anything we can do to gain more knowledge and therefore more performance from every CFD and wind tunnel run gives us a competitive advantage,” said Roberts.
But the teams still have to pay for it. Baqué said NCS costs vary depending on team size and type of access, but are typically between 100,000 and 1 million euros per year. Considering F1 teams also operate under an annual cost cap of $135 million, that’s a significant commitment.
Williams’ Roberts was unwilling to point out specific parts or lap time improvements thanks to the NCS software, but said it had impacted his car’s performance: “This technology is used as part of our toolset for the aerodynamic development of the car.” Therefore “We can’t directly attribute the lap time to this, but we know it improves our correlation and the speed at which we can investigate new aerodynamic conditions.”
Beyond aerodynamics
The ceaseless march of AI will not end here. There is talk of artificial agents on the pit wall determining the racing strategy and even the vehicle set-up.
“It’s a fascinating time as the growth in the AI/ML industry is exponential,” said Roberts. “However, it is also a real challenge facing anyone engaging with technology today. What new tools are we dedicating to research, development and adoption?”
That’s not the kind of intrigue that will captivate the average Drive to Survive viewer, but for many F1 fans, the race behind the race is the ultimate source of drama.
As for Neural Concept, the company continues to push deeper into the non-motorsports side of the automotive industry, working on developing more efficient electric motors, optimizing cabin heating and cooling, and even participating in crash testing.
Baqué said the company’s software can help engineers optimize a car’s crashworthiness while saving unnecessary weight. However, the company can currently only perform crash simulations for individual components, not entire cars. “This is one of the few applications where we have reached the limits of performance,” he said.
Maybe another application for that The EU’s emerging AI supercomputing platforms?