Twenty Years of Swipe-Fee Lawsuits, and Visa's Fees Only Went Up

A merchant lawsuit against Visa and Mastercard was filed in 2005.¹ Twenty years later it is still going. In that time it has produced a $7.25 billion settlement that an appeals court threw out, a $5.54 billion settlement that finally stuck, a $30 billion deal a federal judge refused to even preliminarily approve, and — as of late 2025 — a fresh $38 billion proposal that has not been approved either.⁸²³⁶ Through all of it, total U.S. swipe fees did not fall. In 2024 they hit $111.2 billion, up from $100.8 billion the year before.⁷ Two decades of litigation, four settlement numbers, and the meter kept running the whole time.

The official story is that this is an existential antitrust fight — the merchants of America versus a payments duopoly, with billions on the line and the outcome in doubt. The truer story is quieter and more unsettling: the litigation is not a threat Visa is fighting off. It is a recurring expense Visa has learned to absorb, the way a toll road budgets for road repairs. The settlements are the cost of keeping the road open, not the price of being shut down.

Four settlement numbers, and only one of them is cash

Start with the figures, because the press routinely scrambles them into a single villain. There are really two separate tracks. One is damages — backward-looking cash for fees already paid. The other is injunctive relief — forward-looking changes to the rules and rates. The cash track is the smaller one: after the Second Circuit vacated an earlier $7.25 billion deal in 2016, a $5.54 billion damages settlement was approved in 2019 and affirmed on appeal in 2023.⁸² The eye-watering $30 billion and $38 billion numbers belong to the other track entirely — they are projected fee savings over five years and beyond, not money anyone writes a check for.³⁶ A headline that says 'Visa to pay $38 billion' has quietly confused a forecast with a fund.

Once you separate the tracks, the scale snaps into perspective. The only cash that has actually changed hands across two decades is the $5.54 billion damages settlement.² Set that against a business that booked $35.9 billion in net revenue in fiscal 2024 alone — Visa's share, not counting Mastercard's — and the entire historical cash cost of the most famous antitrust fight in payments is roughly two months of Visa revenue, spread over twenty years.⁵ This is not a company bleeding out. It is a company expensing a line item.

Visa is being sued over a fee it does not collect

Here is the detail that reframes the whole dispute, and almost no coverage carries it. The swipe fee merchants rage about — interchange — is not Visa's revenue. Visa's own 10-K says so plainly: interchange reimbursement fees 'are paid by acquirers to issuers' and are 'set independently from the revenue we receive.'⁴ The money flows from the merchant's bank to the cardholder's bank. Visa is not the recipient. What Visa controls — and what the lawsuit actually targets — is the rule: the network requirement that merchants accept the cards at the rates issuers set, the honor-all-cards condition, the no-surcharge constraints. Visa is the rulemaker, not the toll-taker on this particular toll. The antitrust claim is about market power over the rules, not about a pile of interchange cash Visa is hoarding.

That distinction is why the settlements keep landing on rule tweaks rather than refunds, and why the headline savings numbers are so soft. The November 2025 deal promises a 0.1 percentage-point reduction in swipe fees phased over five years and a cap on standard consumer rates at 1.25% for eight years.⁶ A tenth of a percentage point. On a fee base growing with card volume every year, a rate concession that small can be swamped by the next year's transaction growth before the ink dries — which is exactly how fees can keep climbing in dollar terms while a settlement 'cuts' the rate. The merchants get a lower number on a bigger pile.

Why a judge tore up the $30 billion deal

If the settlements were toothless, you would expect the court to wave them through. The opposite happened, and that is the most interesting part. On June 25, 2024, Judge Brodie denied preliminary approval of the $30 billion deal, saying she was 'not likely to grant final approval.' Her 88-page opinion, released in early July, found the deal failed to treat merchants equitably and offered relief inadequate next to what merchants might actually win at trial.³ A judge looked at the headline number, did the same arithmetic, and concluded the concession was too thin to count. The bench, in other words, saw through the projection — the very thing the press kept reporting as a $30 billion win for merchants.

And yet the rejection changes less than it seems. A vacated deal in 2016, an approved one in 2019, a rejected one in 2024, a fresh one in 2025 — the pattern is the point.⁸²³⁶ Each cycle, the networks come back with a new number and a slightly rejiggered set of rule concessions, the merchants object, a judge weighs it, and the underlying machine keeps running. Visa booked a $2.5 billion litigation provision in fiscal 2025 for this case and related matters — a reserve, set aside calmly, the way you'd budget for a known recurring bill.⁵ The settlement process behaves less like a verdict and more like a pressure valve: it lets accumulated merchant fury escape in periodic, survivable bursts, and the system resets.

The honest objection: maybe this time it bites

The fair counter to all this is that 'managed cost' is a story you can only tell while you're winning, and the recent record cuts both ways. A judge did reject the $30 billion deal precisely because it was too small — that is the system working, not the system captured.³ The $38 billion proposal still needs court approval and is already drawing opposition from the National Retail Federation, so it is not a quiet rubber stamp.⁶ And the litigation is only one front: the same rule-making power that draws these suits also draws regulators, and government-built payment rails keep eroding the card networks' monopoly on moving money. A pressure valve works until the pressure stops being releasable — until a court orders structural rule changes the networks can't dilute with the next year's volume growth. The honest read is that the duopoly has absorbed every blow so far, and that nothing about absorbing the last twenty guarantees absorbing the next.

Twenty years of litigation has produced four settlement figures, one written verdict striking a deal down, billions in reserves, and a fee total that climbed the entire time. That is not the profile of a company under siege. It is the profile of a company that owns a road everyone has to use, faces a recurring repair bill, and pays it. The lawsuit was never really about whether merchants would win. It was about how expensive it would be for the road to stay open — and so far, the toll-keeper has decided the answer is: affordable.