F1 2026: The Data Economy and the Price of Unverified Numbers

On 25 November 2026, Formula 1 and the FIA confirmed an agreement to bring General Motors' team — later branded Cadillac F1 — onto the grid from 2026, attached to a $450m anti-dilution fee. That weekend I reopened the Liberty Media revenue tracker I have kept since 2026: $1.145bn in 2026, $2.573bn in 2026, $3.409bn in 2026. Three data points, three measurement cycles, three different documents. I realised I had consumed those figures for years without ever asking how they were produced.

Around the same time, a report ran through my internal analysis pipeline and came back with a single label: f1. Nine analytical sections, complete headings, zero facts. No team, no driver, no circuit, no session, no date. Technically the system was not wrong. It was honest to an uncomfortable degree: with no data, the only way to preserve integrity is to declare that there is no data.

I have followed F1 since 2026 and have not missed a Grand Prix. It took a blank spreadsheet to make me ask the question my job demands: where was this number measured, by whom, and who paid for it to exist?

The Production Line Behind an F1 Number

Formula 1 runs on five data streams. The first is Formula One Management's revenue, published quarterly and annually by Liberty Media. The second is the prize-money pool redistributed to teams. The third is the FIA cost cap, policed by the Cost Cap Administration. The fourth is media rights, where every contract is sold on audience volume and viewer loyalty. The fifth is the driver market, where on-track performance is converted into contract value.

These five streams feed each other on a single currency: the belief that the number both sides are looking at is the same number.

The 2026 season disrupts all five at once. The new power unit regulations split output 50/50 between combustion and electric power, mandate 100% sustainable fuel, and raise energy recovery substantially. Cadillac F1 joins as the eleventh team. Audi takes over Sauber, Aston Martin switches to Honda power, Alpine moves to Mercedes, and Red Bull runs a Ford-supported power unit. In the United States, a new broadcaster replaces ESPN after years of incumbent coverage.

My time inside a V.League club gave me one simple threshold: when a wage bill exceeds 68% of revenue, the club is in danger, and the safe band sits near 50%. Formula 1 does not operate on that ratio, but the principle holds. Any system that lets costs run ahead of revenue is buying time with unaudited data.

Four Calculations Nobody Publishes

Start with Cadillac F1's $450m anti-dilution fee. The new team pays it for the right to share the pot, and by the common reading it flows to the incumbent teams. Split ten ways, that is $45m each. It sounds large. Placed beside the annual redistribution, the picture changes. With F1 revenue near $3.4bn and the team share at roughly half, each of the ten teams receives an average of about $150m to $160m per season. An eleventh entrant dilutes that share, subject to the phasing negotiated in the Concorde Agreement.

If the net loss per team lands between $10m and $15m a season, then $45m covers only three to four seasons. This is the arithmetic teams do not publish, because publishing it means admitting the true lifespan of a deal.

The second calculation sits in the US media rights. Widely reported figures put Apple TV's payment at roughly $140m a year over multiple seasons, after an ESPN era priced materially lower. Across 24 races, $140m a year works out to nearly $5.8m per Grand Prix.

That figure only means something alongside cost per viewer. Recent US audiences for F1 have hovered between one million and 1.3 million average viewers per race. Divide it out and the broadcaster is paying roughly four to five dollars per viewing, per race. In sports media, that is the price of confidence in growth rather than the price of the current audience. Flat viewership makes the deal a loss. Growth at the pace of the last three seasons makes it cheap.

The third calculation is the cost cap — the only place in F1 where money is forced to declare itself. In 2026 the cap stood at $145m per team. In October 2026 the FIA found Red Bull in procedural breach and in minor overspend, around $2.16m, below the 5% threshold that defines a material breach. The penalty issued in October 2026 was a $7m fine plus a 10% reduction in aerodynamic testing time over twelve months. Aston Martin was also sanctioned procedurally in the same cycle.

The notable part is not the size of the penalty. It is that for the first time in the sport's history, a performance dispute was settled through a balance sheet rather than through scrutineering. A team lost 10% of its aerodynamic testing allocation — a slice of next season's development capacity — because of a line item. Power inside F1 had shifted from the technical office to the finance office.

The fourth calculation belongs to the driver market. Cadillac F1 announced Sergio Pérez and Valtteri Bottas for its debut season. Purely on pace, that is not the fastest pairing a new entrant could buy. But a seat is not priced on lap time alone. It has three layers: car development capability, consistency that protects crash costs, and the commercial value of the audience attached to the driver. Pérez brings Mexico and Latin America. Bottas brings eight seasons of top-tier operational data. For a team building a factory, hiring engineers and standing up systems from zero, verifiable operational data is cheaper than unverified peak speed.

Lewis Hamilton's move to Ferrari, announced in early 2026 for the 2026 season, shows how that third layer operates. The value of the deal did not sit in lap time, which had been publicly measured for eighteen years. It sat in the global audience attached to the driver, in the media equity the team acquired, and in the re-rating of the Ferrari brand after the announcement. The market does not pay for speed. It pays for the predictability of the cash flow speed produces.

The Blind Spot: A Spreadsheet That Looks Complete

All four calculations share one property. Each rests on a number published by one party and consumed by another, with no independent reconciliation in between.

Revenue traces back to Liberty Media's audited financial statements. The cost cap traces back to the FIA's audit function, which has a process. Sponsorship has neither. Every team sells sponsors a composite report of broadcast reach, streaming reach, social engagement and exposure-value equivalents. Three measurement systems, merged into one slide, reconciled by nobody. The sponsor buys a percentage of a total it cannot verify.

That is why I keep this line intact: in the F1 economy, the most expensive number is the one nobody verifies, because it behaves like a verified number — and looks more complete than the real thing.

The blank report my pipeline returned is the extreme version of the same problem. It had nine sections, analytical headings, correct formatting. It lacked only data. Skimmed, it looked like a genuine report. Checked, it collapsed in three seconds.

The Contrarian Read

Many people in the industry will argue that F1's core problem is on-track integrity: technical trickery, regulatory grey areas, post-race scrutineering disputes. That argument has merit, and I am not dismissing it. It only holds when the game is decided on the track.

For more than a decade, almost every major F1 dispute has been settled off it — in ledgers, in contracts, in Concorde rooms. The real fight happens where no camera points, and that is also where the data is weakest.

A second popular belief strikes me as wrong: that more data produces better decisions. In F1, more data usually raises confidence in a wrong number, because volume hides source quality. A team with a sophisticated aerodynamic model can still be consistently wrong if wind tunnel and track correlations drift. Adding data to a biased feedback loop only makes a team wrong faster, and more sure of itself.

The boundary condition is clear. If a team operates a closed, verified loop — simulation matching track data, track data matching race results — then large data sets genuinely create advantage. The catch is that the loop is not free, and not every team can pay to build it. Most are operating on the belief that they already have.

Three Actions, With Deadlines

Back at my own spreadsheet, three commitments.

First, immediately: attach a source, a publication date and a measurement method to every F1 figure used in analysis. Without all three, the figure is flagged provisional and cannot anchor a conclusion.

Second, within one season: separate audited figures such as revenue and cost cap from self-reported figures such as sponsorship performance. The two must live in separate columns and never be summed into one total.

Third, before the 2026 season begins: rebuild the forecasting model, because the new power unit regulations strip historical data of most of its reference value. The old model rests on old power assumptions. Keeping it for comfort is the most expensive way to be wrong.

One bad data point does not sink a racing team. A chain of decisions built on bad data, repeated across seasons, does.

Closing

Every record on track begins with a lap and ends with a line in a spreadsheet. A driver's value does not live in the current contract, but in how the market re-prices him after each season. And the track is where emotion gets traded, but professionals must read the balance sheet before they read the timing sheet.

In 2026 I started keeping F1 notes out of curiosity about speed. In 2026 I still sit in front of the same spreadsheet, asking a different question: who pays for this number to be measured, and if the payer disappears, does it still exist?

A team leaving the grid is not a full stop; it is the most honest financial statement that team ever published. By the same logic, a number with no source is not data. It is an assumption waiting for someone to pay.

F1 fans in 2026 will have more numbers to argue about than ever. What deserves more argument is where they come from. When the next race ends and a new record is announced, try one question before sharing it: what instrument measured this, and who published it?