Ferrari, IBM, and the Other Side of the AI-in-F1 Bet

Ronni Holmvig Strøm · 2026-05-25

When IBM and Ferrari rolled out new AI-powered features for the Scuderia Ferrari HP app at the Miami Grand Prix on May 1, the framing was familiar: a watsonx-powered AI Companion, a unified Game

When IBM and Ferrari rolled out new AI-powered features for the Scuderia Ferrari HP app at the Miami Grand Prix on May 1, the framing was familiar: a watsonx-powered AI Companion, a unified Game Center, an enhanced Race Center, and a press release full of numbers about engagement. A year after the relaunch, IBM reports a 62% lift in views, a 56% rise in average race-day active users, and a 35% bump in cumulative downloads. The interesting part is not the numbers. It is what IBM and Ferrari decided to do with watsonx in a sport already crowded with AI partnerships.

Most AI Deals in F1 Are About Winning. This One Isn't.

The recent run of AI sponsorships in Formula 1 has aimed at the pit wall. Oracle has been Red Bull's title sponsor since 2022 and a tech partner since 2021. AWS has been the sport's analytics backbone for longer. In February, Anthropic signed its first major sports deal with Atlassian Williams, with Claude branded as the team's "Official Thinking Partner" for race strategy, car development, and operations. The implicit promise of these deals is performance, a tenth of a second somewhere in the lap against a low regulatory ceiling.

IBM and Ferrari have pointed watsonx the other way. The Scuderia Ferrari HP partnership is titled Official Fan Engagement and Data Analytics Partner, and the data analytics in question runs toward the fan, not the front wing. The enhanced Race Center surfaces real-time driver data through practice, qualifying, and race day, but the audience is the person on the couch, not the strategist on the wall. The product is the relationship, not the race.

That is a different bet about where enterprise AI value actually sits. The teams have already been buying AI for years to chase laptime, and the marginal return is shrinking. The nearly 400 million Tifosi that Ferrari claims as its global fanbase are a more or less untouched substrate. A standalone fan app is, as TechCrunch reported, still rare among the eleven teams. Ferrari, McLaren, and Williams are among the few really running their own. The competitive frontier in sport, IBM is wagering, is no longer just the car. It is the fan graph.

The AI Companion Arrived Before the Italian Translation

There is a smaller moment in the TechCrunch piece that says more about how AI gets adopted than the AI Companion does. The old Ferrari fan app was not available in Italian. Ferrari is an Italian company. Many of its most devoted fans are Italian. And the most basic form of personalization, picking the right language for the user, was missing until IBM came in to overhaul the thing.

It is easy to read this as a cute detail. It is also a common pattern. Product teams reach for the large language model before they fix the part of the product that a competent intern could have shipped three years ago. The AI Companion is genuinely useful. But the discipline of "making each of them feel like we know them," in the words of Ferrari's new head of fan development Stefano Pallard, has been within reach of any team willing to do internationalization work for a decade. That it took watsonx to surface the gap is the part worth noticing.

We suspect this will be the dominant shape of enterprise AI rollouts for a while. The LLM gets the budget. The boring infrastructure improvements ride in alongside it, framed as features of the AI launch, because nobody was going to fund them on their own. Powered by AI turns out to be, in part, a procurement mechanism for fixing things that should have been fixed already.

The Question the Presser Doesn't Answer

IBM's release is careful about what it claims. The numbers are engagement metrics: views, sessions, active users. None of them are revenue. None of them are loyalty in the sense Pallard invokes when he talks about "loyalty that lasts." The case that a 62% lift in views inside a fan app translates to merchandise sold, races attended, or hearts won is not made, because in May 2026 it cannot yet be made. Year-one engagement data is what you ship when retention data isn't ready.

The interesting test is whether, in two seasons, IBM can show the fans who used the AI Companion behave differently from the ones who didn't. If yes, watsonx becomes the template for every team's customer relationship layer. If no, the AI Companion was a very expensive way to ship the Italian translation.