Go Ahead, Lean Into AI and Smart Testing

VI-grade’s annual Smart Prototypes Summit (SPS), held in Udine, Italy, featured whole-room simulators. (VI-grade)

June 26, 2026

Sebastian Blanco

If you want to improve a vehicle’s dynamics, you’ll inevitably have to deal with wheels and tires.

The trouble is that understanding tires – how they behave in an endless variety of circumstances – is a complicated science. If tires were easy, VI-grade’s annual Smart Prototypes Summit (SPS), held in Udine, Italy, would look quite different.

Given VI-grade’s work on vehicle dynamics, it’s not a surprise that, each year, SPS features plenty of companies that work on tires and wheels, including simulation to on-road testing. One example is Ford Motor Company, which today measures, organizes and analyzes road load data using HBK products (VI-grade is part of HBK's Virtual Test Division) as an update to the nCode-based Corporate Road Loads system the automaker used for decades. Given that data now comes from any number of sources, including proving grounds, connected vehicles, an engineer or automated driver using a virtual vehicle on a virtual proving ground, all the information can get unwieldy.

“If I'm an engineer at Ford, and I'm responsible for one part of this particular vehicle,” Kurt Munson, manager of application engineering at HBK, told SAE Media at SPS. “Out of this sea of road load data that Ford has, how do I find the data that I need to understand? How do I make the wheel strong enough to not break?”

Ask Aqira

HBK’s answer is Aqira, a web-based system that manages test data and also allows for collaboration between physical tests and CAE simulations. Think of it as an efficient data filtering tool for the overwhelming amount of data that engineers are faced with in 2026. Aqira can ingest all the road load data along with metadata about the vehicle and other factors. Munson compared the Aqira dashboard to shopping for something on Amazon, but what you’re putting into your cart isn’t trinkets but information.

“Rather than, say, a TV, I'm buying, so to speak, the road load data that meets specific search criteria,” he said. “Instead of how big is the TV and what's the resolution, now it’s, what's the vehicle? Mustang. What model year? Oh, it's a 2032. Furthermore, I need wheel load data for the back left side. Those are my search criteria.

“All of those road load sets come together in Aqira and allow me, as the person that needs to make sure that wheel's strong enough to not crack, to go find that really easily. Out of terabytes of road load data for all kinds of different vehicles globally, I can find exactly the data I need to then go up and perform more detailed analysis to dig in and find things like, ‘under these load conditions, how long will the wheel last?’”

HBK’s Guido Bairati (right) and Tanneke Reinders speak at Smart Prototypes Summit 2026. (Sebastian Blanco)

Aqira’s data intelligence layer and other upgrades were part of a notable trend at SPS this year: smart testing and physical AI. The shift comes with a change in the name, leaving “Zero Prototypes Summit” in the past.

"For many years, especially at VI-grade, we were focusing on the simulation part of the development,” Guido Bairati, vice president of global sales at HBK, said during the summit. “Now, we’re seeing a so-called shift to left, which means front-loading activities in the first part of the development process as much as possible. For many years, we advertised the zero prototype vision. Now, we also realize the value of actually combining simulation and physical testing together, in order to get the most out of the two, supported by a data platform.”

Which leads us to AI. HBK application engineer Nicolas Baron told SAE Media that most of the AI that's built into HBK software is there to help users do something, as opposed to doing something for the user. For example, engineers can use AI to create and run Python scripts within Aqira.

“We could use AI to create those scripts for you easily, quickly, efficiently, and then inject that directly into the system, so that the AI version will be done through Python scripting, which AI provides to you very efficiently, even if you know nothing about Python scripting,” Baron said. A future update could allow the system to recommend data sets that are similar to the search terms but not technically included, he said.

Smart Testing at SPS with AI

Smart testing is an engineering framework built for the new automotive reality. By unifying virtual and physical testing with continuous data connections, smart testing eliminates disconnected processes and replaces them with an intelligence layer that correlates simulation data, physical test results, and real-world usage on a central platform. The output isn't a report, but a flow of actionable engineering insights that can accelerate decisions.

Three connected factors that can make for smarter prototypes were on full display at SPS: validation moving earlier through simulation and model-based design; engineering disciplines running in parallel rather than sequence; and continuous development that extends well past the production line and into software-defined vehicle functions, ADAS, powertrain calibration, and OTA updates.

“Everybody [at SPS] is talking about how to reduce [development time], and the reduction is really in the front end,” Tanneke Reinders, executive vice president for simulation and validation at HBK, said, during a press panel at SPS. “But we should also not forget how key the physical testing is. The next part is what we are building at the moment is what we call engineering data intelligence. AI will really accelerate our understanding and create faster intelligence out of the data.” Reinders said she spoke with an engineer at SPS who told her that the development time their team is facing for a new model is just two years. She admitted that it will take a lot of work to get to that level, but between new simulators, hardware in the loop rigs, simulation models, and real prototypes, that target becomes more feasible.

Given all of the data generated during new vehicle development today, Bairati said there’s little question that new AI tools are required.

“When it comes to a large amount of data, I don't think you can do a lot without AI, and I think it's not a coincidence that nowadays all our customers are talking about data,” he said at SPS, before giving an example of speaking with a durability engineer at an unnamed Detroit OEM who wanted HBK’s help to extract some sort of intelligence from all of the test data the team had collected in the last 15 years.

Of course, when there’s talk about sharing data, questions of ownership and security often follow. HBK president Ben Bryson told SAE Media that companies can put barriers on what access their suppliers and collaborators get to the data, but the biggest benefits will come to those who work well with others.

“What our customers need are people who can collaborate and work in an open ecosystem, not be protective, closed, and preventative from doing business,” he said. “The companies that will win in this cycle are the companies that are going to help be solution-oriented, and some customers will want us to own the data, and others won't, and the ones that don't will still want us to provide them with a tool that can get them to the insights, and that's what they're aiming to achieve.”

One way that data openness, done right, can help shorten development times is by smoothing out steps all along the way, from initial design to start of production.

Alessandro Pino, vehicle dynamics control manager for Bugatti Rimac (left) and Enrico Maria Talarico, vehicle dynamics manager, speak at Smart Prototypes Summit 2026. (VI-grade)

“If you think about something like an nCode that's helping on durability and reliability, right at the front end of the design cycle, the more open [the OEMs] are with the design shape, sizes, dynamics, structural loads, the things that they're trying to solve – whether it be the next Ferrari or the next family vehicle – the more open they are in that test, the more you can then create a real seamless development chain for them,” Bryson said. “You start to do ride and handling testing, structural load and dynamic testing in the physical world and then you can take some of that data and go into the production environment and learn at that stage and create a big ring around that data. The more open that interface is, the better we are to help solve their problems. That's where the opportunity lies. The more open we are – but in a controlled sense, since cybersecurity is really key – but the more open we are, both from us and from the customer, the more you can accelerate innovation and deliver those insights faster.”

Over two-and-a-half days at SPS, dozens of OEMs and suppliers shared stories about how they used simulation tools and AI to improve and shorten development times. Alessandro Pino, vehicle dynamics control manager for Bugatti Rimac, said the sports car start-up is a virtual-first company. That way, they can take a small group of physical prototypes to the track and prove out what the simulations showed.

“We don't have a fleet of 100 vehicles,” he said.” We have probably 10 prototypes, so we have to make sure that we do our homework before the virtual world, and that is true for a suspension system, this is true for an ECU or the code that goes in the ECU because then we're going to merge all of this in our dynamic simulation tool. We know that the virtual really equals the physical, and we are not [learning] something new about the car once we go on the track. We know already about the car, and we have a high-fidelity digital twin, and the track testing is really a certification, not really much development. Of course, we're going to do some calibration to adapt the models, but that's really not development. The true core of the development has been done virtually. Why? Because we have a very important identity to preserve. It's very important that the vehicle dynamics identity is being developed from the ground up.”