It’s long from pedal bikes to Formula 1. But that’s exactly the quantum leap AI-based startup Neural Concept and its co-founder and CEO Pierre Baqué have made in just six years.
In 2018, the company’s new software helped develop the world’s most aerodynamic bike. Today, four out of 10 Formula 1 teams use a development of the same technology.
Along the way, Baqué’s company signed contracts with aerospace suppliers such as Airbus and Safran, earning a $9.1 million Series A raise in 2022. Now with 50 employees, Switzerland-based Neural Concept is working on a Series B round, while its software helps historic F1 teams like Williams Racing find their way back to the top of the world’s premier motorsport.
However, where Formula 1 cars are based on 1,000bhp hybrid V6 engines, Baqué’s first practical application of the technology was human.
Pedal power
In 2018, Baqué was studying at the Computer Vision Lab at École Polytechnique Fédérale de Lausanne, working on applying machine learning techniques to 3D problems.
“I got in touch with this guy who was leading this team, designing the sixth or seventh generation of bikes, and their goal was to break a world bike 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 bike design iterations.
“Two days later, I came back to him with a shape that almost resembled the current world record holder,” Baqué said. Impressed, the team asked for more repetitions. The result was, according to Baqué, “the most aerodynamic bike in the world right now.”
That’s a strong statement, but it’s backed up by several world records set in 2019. We’re not talking about spoiler-shaped tubes or dimpled rims to reduce drag. This bike is fully covered, with the rider sweating 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 trained engineers use advanced software suites to perform 3D aerodynamic simulations.
CFD is much faster than carving physical models and throwing them in a wind tunnel. However, it is also extremely system intensive and highly dependent on human beings making good decisions.
At its core, NCS helps engineers avoid potential aerodynamic pitfalls while pushing them in directions they might not have considered. In “copilot mode,” an engineer can upload an existing 3D shape, providing a starting point, for example.
NCS will then dig into its neural network to suggest improvements or tweaks to possible paths in a 3D choose-your-own game. The human engineer then selects the most promising proposals and runs them through further testing and refinement, repeating the path to aerodynamic glory.
Not just “tricking the wind”
NCS is useful not only for racing but also for the automotive and aerospace industries. “The path to widespread adoption in these kinds of companies is slow,” Baqué said of working in the somewhat conservative aerospace industry. “So we started working more with the automotive industry, where the needs are a little more intense and will change quickly.”
Neural Concept secured contracts with several global suppliers, including Bosch and Mahle. Aerodynamics is increasingly important in the automotive world, with manufacturers increasingly looking for more aerodynamic cars that offer the most range possible from a given size battery pack.
But it’s not just about cheating the wind. NCS is also being used to develop things like battery cooling plates that, if made more efficient, can keep the battery at its optimal temperature without using up too much power in the process. “There are huge gains to be made,” Baqué said, meaning even greater reach.
While the ultimate ground for these technologies is always the road, the ultimate laboratory is Formula 1. A global motorsport phenomenon since 1950, F1 is currently experiencing an unprecedented surge in popularity.
The power of Netflix
The Netflix series “Formula 1: Drive to Survive” brought the excitement of F1 to a whole new audience. While this series focuses on the politics and drama between the teams, success on the track is much more about aerodynamics. That’s where Neural Concepts comes in.
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 with the help of his company’s software run in this pinnacle of world motorsport. “It’s a wonderful, wonderful sense of accomplishment,” Baqué said. “When I started the company, I saw it as a milestone. Not just Formula 1, but just to have components designed with the software on the road. And, yeah, every time that happens, it’s a great, great feeling.”
Formula 1 is also a highly secretive sport. Of the four groups that Neural Concept works with, only one was willing to identify itself as a customer, and even that was pretty thorough about the whole process.
Williams Racing is one of the best-known teams in Formula 1. Founded in 1977 by racing legend Frank Williams, his team was so dominant in the 1990s that it won five constructors’ world championships, including three in a row since 1992 until 1994.
But, as in most sports, success is cyclical for Formula 1 teams and right now, Williams is very much in a rebuilding phase. The team finished dead last in the 2022 season, rising to just seventh last year.
NCS is one of the tools helping Williams regain its competitive edge. “We are using this technology in a number of ways, some of which improve our simulation and other methods we are working on will help achieve better results for the first time in CFD,” said Williams Head of Aerodynamic Technology, Hari Roberts.
Again, CFD simulations are time-consuming and expensive, a situation 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 also has a limited budget for the computing time they can use to develop their cars.
Any tool that can help a team quickly formulate their aerodynamic designs is a potential asset, and NCS is very fast indeed. Baqué estimated that a full CFD simulation that would normally take an hour would take just 20 seconds via NCS.
And, since NCS doesn’t perform actual physics-based calculations, but makes AI-based guesses based on its network of aerodynamic courses, it’s largely exempt from F1’s draconian restrictions. “Anything we can do that allows us to extract more knowledge and therefore more performance from each CFD and wind tunnel run gives us a competitive advantage,” said Roberts.
But the teams still have to pay for it. Baqué said the cost of the NCS varies depending on the size of the group and the type of access, but typically ranges from €100,000 to €1 million per year. Considering F1 teams also operate under an annual cost cap of $135 million, this is a significant commitment.
Williams’ Roberts wasn’t prepared to point to any specific components or lap time improvements thanks to the NCS software, but said it had an impact on their car’s performance: “This technology is used as part of our toolset to develop the car aerodynamically . So we can’t directly attribute the lap time to it, but we know it helps our correlation and the speed with which we can explore new aerodynamic conditions.”
Beyond aerodynamics
The relentless march of AI won’t stop there. There is talk of artificial factors on the pit wall calling the shots for race strategy and even car setups.
“It’s an exciting time as growth in the AI/ML industry is exponential,” said Roberts. “However, it is also a real challenge facing anyone involved in technology today. What new tools are we spending time exploring, developing and adopting?”
This isn’t 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 the Neural Concept, the company continues to push deeper into the non-motorized side of the automotive industry, working to develop more efficient electric motors, optimizing cabin heating and cooling, and even getting into crash testing.
Baqué said the company’s software can help engineers optimize a car’s crashworthiness while removing excess weight. But, for now, the company can only do crash simulations on individual components, not entire cars. “This is one of the few applications where we’ve hit performance limits,” he said.
Maybe another app for it The EU’s emerging AI supercomputing platforms?
