Thinking About Thinking
F1 teams can't outspend each other anymore, so they're buying brains
Does anyone remember that SAP advertisement with Clive Owen, where he thinks about thinking?
I don’t know why I think about that ad so often, especially in today’s world, where we have AI models doing a lot of that for us.
Maybe because SAP has been a longtime client of mine, or maybe because they were on to something early.
I’ll move on to the point. Anthropic gave itself a job title at Williams: Official Thinking Partner.
Every tech sponsor in Formula 1 sells compute, cloud, or connectivity. Anthropic‘s Claude is selling the cognitive work itself, which happens to be the one thing nobody has figured out how to put on a billboard.
That’s why “Pattern of Thought” exists.
The value of an AI knowledge tool is invisible. You can’t photograph an engineer reading hundreds of pages of technical regulations faster than he could last season. So, ad agency Mother strapped consumer headsets on Alex Albon, Carlos Sainz, and James Vowles, captured their reactions under pressure in the sim and at the desk, and turned that into a visual that lands on helmets and race suits in Monaco.
They sold the romance of thinking.
Williams Racing Team Principal, James Vowles came up through Mercedes’ strategy. He thinks in data, and a team clawing its way up the order has everything to gain from operational speed without the dynasty-sized ego. A hungry operation makes a better case study than a dominant one.
Zoom out, and the timing makes sense. Every serious team now races with a tech giant in the garage. Oracle has been inside Red Bull since 2021, and this February it extended the title deal and pushed an AI strategy agent trackside; Hannah Schmitz’s strategists already run roughly 25% more race simulations thanks to the cloud behind them. McLaren went the other way with Google, training Gemini models to predict race outcomes. Ferrari runs its strategy engine on AWS, which also ingests 1.1 million data points from each car every second, generated by 300-odd sensors. The aero math is just as stark. A CFD simulation that used to take 40 days now clears in under 12 hours.
So the compute war is both settled and permanent. What’s left to win is judgment.
Here’s the mechanic nobody puts on the broadcast. Since 2021, F1 has operated a cost cap set at $215 million for 2026. You can’t buy your way to the front anymore. Wind tunnel time is rationed, CFD is metered, and salaries are capped, so every team is solving one problem: more performance per regulated dollar. A tool that lets an engineer digest a new rule, debug a model, or pressure-test a strategy in an afternoon instead of a week isn’t a gadget. It’s cost-cap arbitrage. You’re buying thinking speed in a formula that has capped almost everything except how well your people think.
That’s the real reason Williams was the right lab. Mid-grid was the point. James Vowles came up through Mercedes strategy, he thinks in data, and a team clawing up the order has everything to gain from operational speed and no dynasty-sized ego to protect.
If Anthropic can show that a full racing team moves faster by thinking with Claude, it has written a sales deck for every mid-sized enterprise watching the broadcast. And the Formula 1 calendar is the live demo.
