AI in Formula 1 2026: What’s Actually Being Delivered, Team by Team

Formula 1’s 2026 regulation reset has helped trigger a wave of AI-branded activity across the grid, though not all of it is new. Williams, Mercedes and Cadillac signed fresh AI-first deals this year. McLaren’s Gemini and Red Bull’s Oracle strategy agent are the AI layer on top of sponsorships that are years old already. Some have shown the working. Red Bull has named its strategy agent. Aston Martin has put engineers, partners and numbers on stage. Others have sold the campaign harder than the technology. AI isn’t arriving in F1 during a wave of public trust. It’s arriving while large parts of the audience are already asking whether the technology is moving too fast.

Key Points: AI in Formula 1 2026

• Not all of 2026’s AI activity is new. Williams/Claude, Mercedes/Microsoft and Cadillac/TWG AI are fresh deals, while McLaren’s Gemini layer sits on a Google partnership dating to 2022 and Red Bull’s Oracle strategy agent sits on a title deal running since 2021
• Disclosure levels vary sharply. Aston Martin and Red Bull have named their tools and cited specific figures. Williams and McLaren lean heavily on creative campaigns with far less technical detail
• Williams’ Claude is confirmed for trackside data interrogation and tyre degradation comparisons, bounded so it does not decide when to pit. McLaren’s Gemini is more specifically described: a named regs search tool, a cross-system data interface and named creative tools
• US public sentiment on AI has moved sharply negative through 2026: an Economist/YouGov poll found over 70% of Americans think AI is advancing too fast, and only 18% of 14 to 29 year-olds say they feel hopeful about it
• Given that backdrop, staying vague about how much a team leans on AI looks like a defensible commercial choice as much as a gap in disclosure
• AI is the latest in a line of industries using F1 as a fast route to mass-market legitimacy, following tobacco through the 2000s and crypto’s 2021-22 boom and bust

Why Now: The 2026 Reset

Not everything in this piece landed in the same window. Williams, Mercedes and Cadillac signed new AI deals in early 2026. McLaren’s Gemini and Red Bull’s Oracle strategy agent are new layers on top of much older sponsorships, Google has backed McLaren since 2022, Oracle has been Red Bull’s title partner since 2021. What they share is timing on the AI-specific escalation, not the underlying deal. The 2026 technical regulations are the biggest reset in over a decade: a near-even power split between combustion and electrical systems (53/47 in practice), active aerodynamics, fully sustainable fuel. Teams that could previously lean on years of accumulated simulation data were suddenly working with none of it. Oracle’s own material puts it plainly, the new rules demand deeper, more granular modelling than anything the sport has needed before. That’s a real problem every team on the grid was facing at once, whether the AI answer to it came from a brand-new partner or an existing one adding a new tool.

The other half of the timing is money, but not in the same way for everyone on this list. Google, Oracle and Amazon are established tech giants funding this from mature businesses that already existed before the current AI boom: cloud, search, enterprise software and retail. They already understood F1’s value from their existing relationships before Gemini or the Oracle strategy agent ever entered the picture. Anthropic is the opposite case. It’s an AI-native company with no prior F1 relationship, and it only reached the scale and investment needed to be a credible bidder for a team-level partnership recently. Five years ago it couldn’t have signed this deal. In 2026 it can, at the same moment Williams needs exactly what it’s selling. The rules created the problem for the teams. Two very different kinds of money showed up to answer it.

Perplexity is the clean contrast. It bought Lewis Hamilton’s reach, not a team’s pit wall. Reporting at the time confirmed Perplexity held conversations with Red Bull, Aston Martin and Williams before going individual instead, branding on the helmet, collaborative content, no claim about strategy, car development or race execution. An AI brand doesn’t need to pretend it’s changing lap time to justify spending in F1.

That’s where the split starts.

Case Study: Williams and Claude

Williams named Claude its Official Thinking Partner in February, with language covering race strategy, car development and operations. That’s a huge phrase. If you use it, fans are entitled to ask what the thinking is. Five months of coverage since has produced four public campaign moments: the livery reveal, the Monaco headset campaign and race suit and livery Pattern of Thought, an F1 Academy livery at Silverstone, and a public lap simulator at the Piccadilly Circus Fan Zone. On the technical side, independent reporting confirms Claude sits alongside engineers during practice, qualifying and races, helping them interrogate data, comparing live tyre degradation against a Friday simulation run, checking how a setup change affected performance earlier in the season. It’s explicitly bounded: Claude does not decide when to pit, and engineers validate its outputs before it’s allowed near a live session. That boundary isn’t just a Williams choice either. The FIA is currently developing its own rules on how AI can be used across the sport, with keeping humans at the centre of race-day decisions the stated aim. Where exactly that line ends up sitting is still being worked out by the governing body, not just by the teams.

None of that came from Williams itself. It took independent reporting to surface it. Williams’ own channels have stayed at campaign level throughout, including on routine content: a standard post-race debrief graphic now carries the same “thinking with Claude” tag as the flagship moments, and public comments underneath have started to notice the gap between the branding and the substance.

Case Study: McLaren and Gemini

McLaren’s Google partnership dates to 2022, and it was Android, Chrome and Pixel branding for years before Gemini entered the picture. Gemini itself only debuted on the car in November 2025 at the US Grand Prix, replacing Chrome as the lead branding, so what’s new for this cycle is the AI layer, not the relationship. On what that layer is, McLaren has shown a little more than Williams, though it still leans heavily on the creative side. The team names a custom regulations tool that lets trackside engineers search and compare dense sporting rules during a race, plus a natural language interface built with Google Cloud that pulls from multiple internal systems at once, described specifically as helping engineers cross-reference lap-time metrics against historical data. On the commercial and creative side, McLaren names its tools too: Nano Banana and Omni for concept development and campaign ideation, market research synthesis for sponsorship pitches, and Gemini’s image generation used for helmet and livery brainstorming.

The gap is there. Just not as big as the branding makes it feel. Both teams describe a bounded, named engineering tool alongside a much larger and more visible creative campaign, Vegas Sphere projections and a Gemini Paddock activation on one side, Pattern of Thought and Piccadilly Circus on the other. Neither team has published anything close to a case study on competitive impact.

The Rest of the Grid

Two teams have gone considerably further than either Williams or McLaren.

Oracle has been Red Bull’s title partner since 2021, so this isn’t a new sponsorship either. What’s new for 2026 is the AI Strategy Agent, and Red Bull’s Oracle deal names that tool outright: live trackside from the Australian Grand Prix, built to automate data collection and interpret historical and real-time race inputs so engineers can respond faster to safety cars, weather shifts and rival strategy calls. It’s described as working alongside human engineers rather than replacing the pit wall call, but the function itself is named and specific in a way Williams and McLaren’s flagship tools aren’t.

Aston Martin has gone furthest of all. Its AI stack runs across nine-plus named partners, CoreWeave for compute, Cohere for enterprise generative AI, Arm, NetApp, Cognizant and others, and the team put engineers and executives on stage at its own technology forum to talk about it on the record. Team principal Adrian Newey put it plainly: feeding in the team’s own wind tunnel, CFD and track data and using AI to spot patterns a human might not catch quickly enough, rather than anything internet-facing. The team has also put a number on it, between 10,000 and 100,000 race scenarios simulated per weekend.

Mercedes (Microsoft Azure) and Ferrari (AWS) sit in between. Mercedes cites hard figures, over 400 sensors per car generating more than 1.1 million data points a second, and describes Azure and Azure Kubernetes Service supporting simulation, race strategy modelling and performance analysis. Ferrari’s most concrete claim is a virtual ground speed sensor built on Amazon SageMaker and a pit stop synchronisation system described as up to 80% faster than manual methods. Both are more specific than Williams or McLaren’s flagship claims, though neither has gone as far as Aston Martin or Red Bull in naming exact functions and putting people on record about them.

Cadillac’s version of this deal is different from every other team on the grid, because it is not a normal sponsorship. In February the new team named TWG AI its primary and exclusive AI partner, covering vehicle simulation, aerodynamic correlation between digital and physical testing, race strategy modelling, manufacturing and logistics, and driver performance analysis. TWG AI and the Cadillac F1 team share the same parent company, TWG Global, so this is closer to a holding group using its own team to showcase its own AI division than a traditional sponsorship. That makes Cadillac less of a sticker deal and more of a built-in advert for the parent group’s wider technology ambitions.

This isn’t the whole grid, either. Audi has picked up ElevenLabs for voice AI, Perk for logistics and travel automation, and Extreme Networks for AI-driven cloud infrastructure, none of it pitched as a competitive edge, all of it smaller in scale than anything covered above. The AI presence across the 2026 grid runs well beyond the seven projects profiled in this piece. It’s just that most of the rest sit closer to Audi’s end of the scale than Aston Martin’s.

The Public Isn’t Sold On This Yet

In the US at least, the public mood around AI is ugly. An Economist/YouGov poll in May found more than 70% of Americans think AI is advancing too fast, a view that holds across every age group and both major parties. Gallup found only 18% of 14 to 29 year-olds say they feel hopeful about AI. A separate survey put anxiety about AI’s rise at 56%, against just 42% who said they felt excited by it. The Axios Harris Poll, an annual measure of corporate reputation, has shown several major AI companies ranking below institutions like ICE in public trust.

That’s the room these deals are walking into. F1 sells human nerve. Driver feel. Pit-wall judgement. That’s the product, and it’s now carrying AI branding on the cars, the kit and the paddock activations at exactly the moment a large chunk of the US public is telling pollsters it doesn’t trust the companies behind that technology. Staying vague about exactly how much a strategy call owes to a model, rather than an engineer, isn’t necessarily a failure to disclose. It may simply be a brand and a team both reading the room.

Not every brand is just staying quiet, though. Some are working around the distrust directly, through the fans rather than the spec sheet. McLaren’s heritage-inspired Gemini livery at Silverstone was built as a fan engagement piece first, tying Gemini to nostalgia and the M2B rather than to anything technical. Claude’s Piccadilly Circus simulator does something similar: put someone behind a wheel for a lap of Silverstone with the AI framed as a companion for the ride, and the whole conversation about race strategy or engineering disappears. It becomes a fun few minutes rather than a trust question. Whether that’s a deliberate answer to the polling or just good marketing instinct, the effect is the same: winning fans over directly is a lot easier than winning an argument about what the model is doing back at the factory.

A Familiar Pattern: Tobacco, Crypto, Now AI

This isn’t the first time an industry has used Formula 1 as the place to buy legitimacy in one go. Tobacco got there first, starting with Gunston backing privateer Brabhams in 1968, building through decades of title sponsorships and, in British American Tobacco’s case, buying an entire team outright. It peaked in the mid-2000s as a final push ahead of the EU’s advertising ban in 2006. Ferrari held out longest, still carrying Marlboro’s parent company support years after the branding itself disappeared from the car in 2010.

Crypto followed the same shape on a shorter timeline. The sector arrived around 2019 to 2021 and exploded fast, nine of ten teams carrying a crypto sponsor at the peak. Then it broke just as fast. FTX collapsed in November 2022, Mercedes stripped the branding mid-season, Tezos dropped Red Bull and AlphaTauri dropped Fantom within weeks of each other, and some regulators forced several teams to cover sponsor logos entirely over the same period. What survived consolidated into a handful of steadier, longer-term deals rather than the free-for-all of 2021.

Both waves picked Formula 1 for the same underlying reason AI has now landed there too. There isn’t another sport quite like it for this purpose. Twenty-four races, multiple continents, the same brand and cars appearing on a global broadcast roughly every fortnight for ten months of the year. Football draws bigger single-match audiences, but nothing else tours the way F1 does, one product, one media cycle, a new country every couple of weeks. For an industry trying to look established fast, that’s a locked-in global stage no single-market sponsorship can match.

AI may not end the same way. That’s not the point here. The pattern is familiar by now: an industry needs credibility fast, and F1 gives it a travelling billboard with a global broadcast attached.

Coffee Corner Take

At least seven AI-linked projects across the grid now, and only two have really shown their working. Aston Martin put nine partner logos on a stage and let Adrian Newey explain, in his own words, what the models are doing with the team’s own data. Red Bull named its tool on day one and described exactly what it’s meant to catch. Everyone else is somewhere between a case study and a car sticker.

That gap isn’t accidental, and I don’t think it’s sinister either. Look at the polling. Most of the US public thinks this stuff is moving too fast. Barely a fifth of young adults feel hopeful about it. Some AI brands are already testing badly on trust. If I ran the commercial side of an F1 team right now, I’d think twice about standing up and saying “the model made that pit call,” even if it were true. Motor racing still sells human judgement under pressure. Announce too loudly that the judgement is partly software, and you risk undercutting the exact thing your fans turned up for.

There’s another motive here, and it has nothing to do with what any of these tools do. Seven AI-linked projects didn’t appear across F1 purely to help an engineer read a rulebook faster. They queued up because a car on the grid is the fastest way to tell the market “we’re the AI of choice, and our rivals aren’t.” Once Oracle is on Red Bull, Gemini is on McLaren and Claude is on Williams, every AI company still outside the paddock starts to look like it missed the boat. That’s not collaboration, it’s positioning.

That doesn’t let anyone off the hook completely. Williams calling Claude an “Official Thinking Partner” across race strategy, car development and operations, then pointing mainly to a headset campaign and a fan simulator, is still a mismatch between the claim and the evidence. Same goes for McLaren dressing a regs search tool in the same language as a livery reveal. If the substance is there, and Red Bull and Aston Martin suggest it can be, showing it doesn’t have to mean overclaiming. It can be as simple as naming the tool and describing what it does. At Aston Martin and Red Bull, there’s something you can point at. Elsewhere, there’s still a lot of fog, and given what the US public currently thinks about AI, that fog might be the point rather than an accident.

I went through the Claude simulator in the Williams activation at Piccadilly myself, and it worked. A lap of Silverstone, Claude framed as the thing riding along with you, and nobody in that queue was thinking about race strategy or data centres. Then came the clever bit. After the lap, it asked what I thought of my own driving, gave its own read, called me a “red liner,” heavy on the throttle, which lined up with the fastest lap in my session, and matched me over sixty percent with Alex Albon. That wasn’t just a fan simulator. That was Claude sitting in the room, guiding you straight into using it without making it feel like you’d gone looking for an AI tool. Fair play to whoever built it. Clever marketing. But it’s not the same thing as showing what the technology does inside the team.

Tobacco and crypto both ended somewhere the sponsors didn’t choose. F1 keeps being the stage an industry reaches for right when it needs the reach most and for the rest of this grid, it’s still more story than evidence. Worth remembering that the next time a car goes past with a new logo on the sidepod.