How F1 Teams Are Turning to AI to Improve Performance on the Track

Behind the roar of engines and the flash of speed lies a silent revolution driven by algorithms, simulations and real‑time intelligence.

McLaren and other F1 teams are using AI to simulate scenarios, optimise race strategies, and streamline operations

Beyond the track, AI is being deployed for fan personalisation, commercial outreach, and business efficiency,AI helps engineers model car performance using digital twins and real-time data from races and tests

When you picture a Formula 1 garage, images of gleaming carbon‑fibre monocoques, mechanics with torque wrenches and tyres being swapped in a blur come to mind. What you’re less likely to visualise is rows of servers, streams of telemetry data, generative AI models running probabilistic simulations, and digital twins whispering guidance to the engineers. Yet that is very much the racetrack of today — and tomorrow.

At McLaren’s Technology Centre in Woking, the team stewards one of the sport’s more deliberate, behind‑the‑scenes bets on artificial intelligence. As Dan Keyworth, McLaren’s director of business technology, describes, “We’re an organisation that’s used traditional machine‑learning tech products for a long time.” But what’s newer — and potentially game‑changing — is how generative AI is being folded into strategy, operations and commercial outreach.

“What AI allows us to do from a generative perspective is to actually game out more of those actual scenarios and go, ‘What will happen?’” Keyworth says — whether that means choosing the optimal moment to pit or which tyre compound to switch to mid‑race. In some cases, the predictions are already “pretty accurate” — to an “almost scary” degree, he admits.

Three Fronts of AI Deployment

Keyworth identifies three domains where McLaren sees the greatest return from AI:

Car performance and race strategy,Day‑to‑day operations and engineering workflows,Commercial, fan engagement and marketing

#### Car Performance & Race Strategy

This is where AI’s edge is perhaps most visible to fans (if they care to look). McLaren uses machine learning to absorb terabytes of sensor data, often hundreds of readouts per second per car; to model the real conditions a vehicle may face. Those data streams feed digital twin models: 3D virtual replicas of cars, underpinned by physics, materials, environment and system models. Engineers can test scenarios in silico, sometimes doing in minutes what would take days or weeks in a wind tunnel or real track test. Digital twinning in F1 isn’t new, but AI and simulation sophistication is elevating its role.

In the McLaren/Dell AI collaboration, their so called “AI Factory” helps process incoming telemetry, detect anomalies, and frame actionable insights.

Other teams are following similar paths. Aston Martin Aramco utilises large “data lakes”; vast repositories of historical and live data, paired with machine learning to forecast tire wear, temperature curves and aerodynamic performance. Clare Lansley, CIO at Aston Martin, notes that freeing up engineers from repetitive modelling tasks allows them to focus more intensely on the car itself.

At the Red Bull–RB outfit (via its technical wing in Faenza), AI is also being used to reduce the need to run hundreds of brute‑force simulations. As RB’s Head of Vehicle Performance Guillaume Dezoteux says, fewer but more accurate models shorten design loops.

#### Operations & Engineering Workflow

Behind the spectacle of grand prix weekends lies a logistical and engineering challenge of extraordinary scale: coordinating parts, scheduling upgrades, managing inventory, staffing across time zones, and maintaining a racing factory that never truly sleeps.

Here, AI helps streamline repetitive, data‑intensive tasks. Through automation, predictive maintenance, anomaly detection and workflow optimisation, teams can reduce the “busy work” engineers historically wrangled with freeing them to do higher value modelling or creative design. This aligns with broader discussions on whether AI Agents Will Steal Your Job Or Help You Do It Better?.

McLaren, for instance, sees AI not as a “labour replacement” but a “labourious replacement”; a means to pull away the overhead that bogs people down.

“You want to unlock your team to do the things you hired them for,” Keyworth says.

“You want to unlock your team to do the things you hired them for,” Keyworth says.

#### Commercial & Fan Engagement

Perhaps surprisingly, the commercial side of F1 is also seeing AI uplift. McLaren uses AI to tailor fan experiences, especially in emerging markets such as the U.S., where the sport is seeing renewed interest. By analysing engagement patterns, local time zones, and content preferences, they can personalise when and how to deliver content to fans, sponsors and partners.

In effect, the same data pipelines that feed race strategy can feed marketing funnels. Over time, the hope is that more immersive, responsive fan interactions, whether in AR/VR, in‑app experiences or even bespoke content will mirror the precision of what’s happening on track. This also ties into how AI & Museums: Shaping Our Shared Heritage are leveraging similar technologies for visitor engagement.

The Connectivity Imperative

All the modelling, AI, and simulation in the world means little without the infrastructure to carry data safely and fast. In Formula 1, connectivity is life. Without low-latency, reliable networks, both on site and back at the factory none of these innovations can function.

McLaren’s answer: mobile data centres. These are compact server units that travel with the team to every circuit, deployed locally to support real-time storage, processing and relay back to Woking. This approach highlights the growing importance of distributed computing and edge AI, a trend also seen in Elon Musk’s Big Bet: Data Centres in Orbit.

Partnerships with technology providers like Cisco and Google are essential to build redundant, high-throughput networks that can carry vast telemetry, video, strategy models and more. Without that infrastructure, Keyworth notes, “no car can be on the track safely.”

Challenges and Constraints

Even in a sport where the appetite for innovation is insatiable, deploying AI at this level carries constraints.

Model fidelity & validation: AI models and twins must be calibrated carefully and continuously validated against real data.,Computational cost & latency: Running complex simulations in real time requires serious compute horsepower.,Regulatory limits: F1 places restrictions on testing and simulation budgets.,Cultural adoption: Engineers must learn to trust and adopt AI insights.,Data privacy & security: Sensitive racing data must be protected from leaks or breaches.

Looking Ahead: What AI Might Enable Next

Smarter scenario planning: Digital twins will be used to run “what-if” simulations far in advance of race day.,Generative co-design: AI might suggest radical new car components, unbound by legacy thinking.,Live adaptive modelling: AI models could update mid‑race, recalibrating on-the-fly.,Collaborative sustainability tech: Shared innovation in sustainable design beyond the race day.,Data-rich fan immersion: From predictive dashboards to personalised race feeds, AI will deepen the audience connection.

On race day, fans see the speed, the strategy calls, the pit stops and winners lifting trophies. Behind that facade, though, is a quietly intensifying arms race: one of models, data pipelines, silicon and simulations. Formula 1 has always rewarded technical ingenuity, and AI is now one of its sharpest tools. A deeper dive into the technical aspects of data processing in high-performance environments can be found in publications like those from the Association for Computing Machinery (ACM).

Teams that learn not just to use AI but to embed it Into every twist, every sensor, every strategy stand to convert those digital decisions into milliseconds on the tarmac; and in this sport, milliseconds are everything.