Visa Inc

Visa's record clears most of Buffett's quality tests, but the price cannot be checked against value yet.

Visa runs VisaNet, a card network that routes and settles payments between consumers, card issuers, banks that accept cards and sellers, and it earns fees rather than lending money. In fiscal 2025 net revenue was $40.0B and net income was $20.1B, with no loss years across 15 years of filings. Return on invested capital averaged 31% and free cash flow averaged 107% of reported profit, so the business has needed little capital to grow, though the operating margin of 60% in the latest year sat below its average of 63%. Total debt of $25.2B is covered 41× by operating income, but the filings give no diluted share count, so the share price of $407.28 cannot be set against an estimate of value. Our research status for Visa is Not enough evidence, because the valuation and retained earnings tests cannot be completed from the filings.

Visa runs a card payments network, VisaNet, that routes and settles transactions between consumers, the banks that issue cards, the banks that accept them, and sellers. Visa does not lend money or issue cards, so it does not carry credit risk. It earns fees for each transaction it processes and for the payment volume it supports, plus fees for added services such as fraud tools and advisory work. In fiscal 2025, net revenue was $40.0B and net income was $20.1B.

• growth in nominal payments volume
• growth in processed transactions
• cross-border volume and currency conversion
• value-added services such as fraud tools and advisory work
• offset by client incentives paid to banks and sellers

Inside the circle? Volume-linked fees have passed every long-run test in the filings, with 0 loss years and an average return on equity of 36%, although share buybacks shrink the equity base that return is measured on. Visa's fees are set or capped by regulators and courts in many countries, and new payment rails and stablecoins could bypass the network, so a ten-year view needs wider error bars than the numbers alone suggest. The filings give no diluted share count, so we cannot judge the share price against an estimate of value.

Visa's profits rest on a payments network that issuers, acquirers and sellers already use. Switching away is costly, and each new connection adds to the network's reach, so the protection sits in the network rather than in any single product. The record backs this up with 0 loss years across 15 years of filings, an average operating margin of 63% and an average return on invested capital of 31%. Capital needs are light, with capital spending at 1.1× depreciation, but buybacks shrink book equity, so a return on equity of 36% flatters the picture and return on invested capital is the cleaner test. The filings do not separate brand and switching costs from regulation, give no gross margin history and give no diluted share count, so we cannot yet test the price against value.

Visa keeps most of its earnings and returns cash mainly through share buybacks, with dividends as a smaller and steadier payment. The average dividend payout is 22% of profit, and cumulative retained earnings stand at $97.3B. Reinvestment needs little cash, since capital spending runs at 1.1× of depreciation and estimated maintenance capex was $959.4M. Buffett's rule is that buybacks make sense only below intrinsic value, and the filings do not give the share count or share price history needed to test that. Management is plain about its litigation exposure and says the accruals could be higher or lower than the escrow deposits.

• Buybacks are the larger cash return to owners in fiscal 2025, and the MD&A reports a repurchase programme with funds still remaining. The filing gives no share count trend, so we cannot say whether buybacks have shrunk the share base over time.
• Litigation takes cash that could otherwise reach owners. Owner earnings are $20.3B and normalised owner earnings are $19.7B, and the excerpts do not split the gap between litigation and other items.
• Debt is modest against cash flow. Total debt is $25.2B, and net cash of -$8.0B means borrowing exceeds cash once the two are set against each other. The debt is used to fund general corporate needs and a euro bond issue, not to buy back shares directly.

• Can't tellThe $1 retained testCan't tell

  Needs both a price history and positive retained earnings across the period.

  Why this test
• Can't tell

Discount rate: 10-year Treasury (5.2%) plus a 3% hurdle. End-of-day prices from free public sources, may be delayed or missing.

Buffett's circle of competence principle asks whether you can follow the economics well enough to estimate the future. Visa's fee model is easy to follow, but the filings show regulators in many countries setting or capping its fees, which makes the future harder to estimate. The principle of a durable competitive advantage points to the network itself: the filings describe a network that issuers, acquirers and sellers already use, and that is costly to leave. Pricing power is the harder question. Visa sets default interchange rates, yet the filings warn that when those rates cannot be set at optimal levels, issuers and acquirers may find the system less attractive. On owner earnings, normalised owner earnings of $19.7B sit close to reported profit, which is the cash test Buffett asked for. The little or no debt principle is met, with total debt of $25.2B and a payback of 1.3 yrs of normalised owner earnings. The one dollar retained test cannot be run, because it needs a share price history that the filings do not provide. Rational capital allocation depends on repurchases, which Buffett said make sense only below intrinsic value, and we cannot test that without a share count. Interest rates act like gravity on long-dated cash flows, and with the ten-year Treasury yield at 5.2%, distant fees are worth less today. The margin of safety principle cannot be applied, so no conclusion on value can be drawn.

• Diluted share count and diluted earnings per share in the next annual report

  Without a diluted share count we cannot estimate value, test the margin of safety or judge whether buybacks are made below value.

• Interchange caps and the outcome of the interchange multidistrict litigation

  The filings say the interchange fees at issue may be reduced, and lower caps would cut revenue directly.

• The litigation provision and the balance of the U.S. litigation escrow account

  The provision rose sharply in fiscal 2025, and the filings say the accruals could be higher or lower than the escrow deposits.

• Client incentives as a share of net revenue

  The filings say incentives rise with volume, so a growing share would mean less of each dollar reaches Visa.

• Stablecoin settlement volume and how the GENIUS Act rules are applied

  The filings say stablecoins could disrupt existing networks in cross-border and business to business payments.

Questions the filings couldn't answer

• What is the diluted share count, and what is the share price history over the period? Without these, neither the margin of safety test nor the one dollar retained test can be run.
• How much of the return on equity comes from buybacks that shrink the equity base, and how much from better operating returns?
• How much of the fiscal 2025 litigation provision will recur, and how would an outcome in the interchange case change owner earnings?