Checkered flags, blurred cars, thrilled fans and headset-wearing pit crews feeding split-second information to drivers hurtling past at more than 200 mph. A grand prix lasts about two hours. But the work that decides it starts weeks earlier and never really stops: hours of collaboration, data and pressure that spectators don’t get to see.
“Formula 1 is an extremely data-dense sport,” Aston Martin Aramco F1’s chief information officer, Fabrizio Pilotti, tells TechInformed.
Teams in the sport started collecting data as early as 1975 through telemetry, and since then the data has not stopped flowing.
“As storage and computation also improved over the last years and decades, teams started to enter the arena of petabytes of data,” he says. “[Now] it’s absolutely impossible to crunch on and intelligently extract engineering information out of this enormous amount of data gathered from all the testing, simulation and racetrack.”
Still, that doesn’t keep teams from trying. Pilotti and TechInformed discussed the nature of data collection at high speed in real time, and how the team is using emerging technologies to stay competitive.
Ahead of the curve
The Aston Martin Aramco team’s factory in Silverstone, UK, processes data volumes that have long since surpassed any meaningful capacity to analyze unaided — a gap AI has stepped in to fill.
According to Pilotti, the relationship between F1 and AI predates the current wave of commercial enthusiasm for the tech by several years. He says teams were deploying machine learning algorithms to process sensor data from hundreds of onboard systems — tires, aerodynamics, power units, cooling — well before the term entered mainstream corporate vocabulary.
“What you see currently on the market was a thing for us maybe five years ago,” Pilotti says. “Currently we’re looking at the next level.”
Right now, this includes partnerships with Cohere, the enterprise AI company, and Cognition, a software development AI firm.
Data processing at speed
Every car that loops around the track generates a continuous stream of readings from sensors, Pilotti explains.
The raw feed from there is used to maintain a digital twin: a continuously updated simulation of the physical car that models how each strategic variable, such as fuel loads, tire compound and pit window, plays out across the remaining laps of a race.
At Mission Control in Silverstone, 45 engineers monitor live data during every race weekend, feeding analysis upward through a chain that reaches the pit wall, the garage and ultimately the driver.
Inside the partner ecosystem
Indeed, while one driver sits at the wheel, many more people work to make the car as fast as possible.
That was on display at Aston Martin Aramco’s inaugural AMR Network Technology Forum, held at the team’s Silverstone campus in July 2026 ahead of the British Grand Prix to mark the launch of its new partner platform, the AMR Network.
The event brought together senior figures from its technology partners, including Cohere, ServiceNow, Cognizant, Cognition, Arm and more to discuss AI, machine learning and high-performance computing in elite sport.
David Ingham, digital partner in media and entertainment at IT consulting firm Cognizant, related his company’s own infrastructure to the team’s.
“There’s a real parallel between Cognizant and Aston Martin, because we’re in a position to demonstrate how we bring together huge amounts of immersive information and data,” he said.
Ryan Lewis, head of UK and northern Europe at enterprise AI firm Cohere, said companies like his want to “give people the ability to offload the mundane, routine tasks that slow them down, so they can focus on the decisions that really matter in F1 teams.”
He added that deploying models securely within a team’s own infrastructure is critical when data is too sensitive to risk exposing to external systems.
Simon Cox, chief transformation officer at ServiceNow, described a partnership with Aston Martin that dates back to 2023, starting with employee joiner-mover-leaver processes — onboarding, internal moves and departures — and now extending into race freight logistics and travel.
“A lot of our work is making sure that, behind the scenes, all the things the public doesn’t see on race day are running properly so the whole operation works,” he said.
Gabie Boko, chief marketing officer at NetApp, said her company’s role is connective tissue across the ecosystem.
“If you took all of us away from Aston Martin, the one thing they’d still have to manage is their car, their people and their races,” she said, describing NetApp’s focus as “data infrastructure” — moving data as much as storing it. “There’s no better test environment for that than something running at the speed of an F1 weekend.”
Built for the pit lane
Powering AI inside the car is semiconductor firm Arm. Aston Martin F1 has in the last year overhauled its car’s architecture, including the ECU, engine, gearbox, aerodynamics, telemetry and a newly built in-house pit stop system.
The pit stop system, developed from scratch in roughly nine months after previously being supplied by Mercedes, is built around Arm’s Cortex-M series and is “nearly fully autonomous,” according to the team, with the exception of the gun trigger, which by regulation must remain a manual action.
John Kourentis, director of go-to-market for automotive solutions at Arm, said it’s been more than 18 months since Arm and Aston Martin began exploring use of the Ethos-U55, a compact, power-efficient IP block for running small machine learning models on embedded hardware.
The team is now working to move more analysis onto the car itself.
One project aims to replace a heavy optical speed-over-ground sensor with a machine-learned model that estimates the car’s sideslip angle from wheel speed, ride-height laser, damper load-cell and accelerometer data, reducing weight while preserving the insight needed to understand tire wear.
A second effort trains models to read tire and brake condition directly from onboard thermal imaging, working around limited telemetry bandwidth and removing the need to reconcile separate photo, broadcast and telemetry sources after the fact.
“Now you’re processing it in real time, which gives you insight and lets you make decisions immediately, rather than through a manual collect-log-analyze cycle,” Kourentis said. “Everything you’re doing is about optimizing the speed of decision-making on the vehicle.”
Within design, the team is using models trained on a smaller set of high-fidelity computational fluid dynamics simulations to estimate the aerodynamic effect of design changes without running every full simulation, and to accelerate analysis of wind tunnel data, which is strictly rationed by the FIA.
Kourentis noted that balancing real-time processing, power efficiency and safety requirements within a race car’s tight power and cooling limits is a central design challenge for Arm’s silicon partners.
Cheaper compute, faster cars
Formula 1 introduced a budget cap in 2021, limiting what teams can spend on car performance. Against that backdrop, Pilotti says, the steady decline in the price of compute and storage, following the long curve of the electronics industry, has effectively subsidized the team’s AI ambitions.
“Every 12 to 18 months, the crunch gets cheaper,” he says. This freed-up cash can be redirected into engineering decisions rather than infrastructure.
As compute becomes more efficient, this also helps the team with its sustainability efforts.
Aston Martin has installed solar panels across its Silverstone campus and monitors its energy consumption in real time on displays throughout the facility.
Efficiency gains from AI reduce processing overhead; reduced overhead cuts power consumption; lower consumption shrinks the carbon footprint. Pilotti treats this not as a separate initiative but as a natural consequence of the team’s drive toward operational efficiency.
Looking ahead
Pilotti identifies agentic AI — systems capable of executing multistep tasks autonomously — as the most immediate frontier. Quantum computing, he believes, is also coming, though the use cases in racing remain speculative. The team is in early conversations with potential partners but has not yet committed to a specific application.
Whatever the future brings, the crew serving the drivers is standing ready — some wearing fireproof suits, others building firewalls. Their mission is all the same: shave another tenth of a second off the lap time.