Mastercard is buying the stablecoin rails that were meant to replace it
On 17 March 2026 Mastercard agreed to buy the stablecoin-infrastructure firm BVNK for up to $1.8 billion, the largest acquisition of its kind to date (CNBC, 17 March 2026). For two years the confident view in payments has been that cheap, programmable, near-instant stablecoin settlement, and checkout agents transacting machine-to-machine, would route value from payer to payee without an interchange leg and let merchants walk around Visa and Mastercard. The rail built to do that walking is now owned by the thing it was meant to bypass, and the people who funded it have been paid by the incumbent for the privilege.
That inversion is worth more than the irony, because it says something contestable about where this goes. The disintermediation trade is not merely losing; for an investor it is, at best, an exit to the network it was supposed to kill, and the reason is that the schemes have quietly made themselves close to regulation-proof. Interchange is the one number a court, the Federal Reserve or Congress has ever managed to cap, through Durbin on debit, the European caps, the swipe-fee settlement on credit. Everything Visa and Mastercard are now growing fastest sits outside that perimeter. When the regulated price is attacked, the schemes move revenue to the services and the rails no statute caps, and when a new rail threatens to carry volume around them, they buy it. A merchant or a regulator can win the interchange fight and still lose the cost war, because the cost has moved to a layer nobody is regulating.
Read the BVNK purchase, then, as a balance-sheet decision about where network revenue is allowed to sit rather than a bet on crypto. Interchange is the commodity line of the card business and the line two decades of regulation and litigation have spent compressing. The proposed US swipe-fee settlement and the European caps already in force have made the per-transaction acceptance fee the most contested figure in the industry, the one a counterparty can attack in a courtroom, at a regulator, or on Capitol Hill. A platform whose core priced good is being commoditised by forces it cannot control does not defend that good. It shifts the mix toward complements it can still price freely, and it acquires the substitute so that the substitution becomes its own product line rather than someone else’s.
The revenue mix shows the shift before the interchange squeeze has fully landed. In the quarter ended 31 March 2026, Mastercard’s value-added services and solutions net revenue grew 22%, or 18% on a currency-neutral basis, more than double the 12% growth (8% currency-neutral) of the payment-network line that carries the interchange-adjacent economics (Mastercard Q1 2026 earnings release). The company lists the drivers itself: “security solutions, digital and authentication solutions, business and market insights and consumer acquisition and engagement services, and pricing” (Mastercard Q1 2026 earnings release). Each is a service the scheme sets the price of, and none is a mandated acceptance fee fixed in a network rule and exposed to a settlement, so the part of the business no regulator caps is growing at twice the rate of the part that is.
Owning the settlement rail is what turns a favourable revenue mix into a strategy, because it answers the disintermediation thesis on its own premise. That thesis treats the new rail as a substitute for the scheme; Mastercard’s reply is to become the operator of the rail. Chief executive Michael Miebach framed the quarter on exactly that point: “we’re advancing agentic commerce with Mastercard Agent Pay and expanding our stablecoin solutions through the planned acquisition of BVNK” (Mastercard Q1 2026 earnings release). Agent Pay makes the network the authorisation and trust layer for machine-initiated commerce, the one part of agentic checkout an autonomous agent cannot provide for itself. BVNK supplies the orchestration between on-chain settlement and fiat rails, the plumbing a merchant would otherwise build or buy elsewhere. With both inside, value that moves over a stablecoin instead of through interchange still moves through a Mastercard-priced service, and the revenue the disruption was supposed to strand lands back on the incumbent.
The price is the tell that this is not a hedge. BVNK was valued at roughly $750 million at its December 2024 Series B, and Coinbase had come close to buying it for around $2 billion before that deal fell apart near November 2025 (CNBC, 17 March 2026). Mastercard’s up to $1.8 billion, including roughly $300 million tied to performance, beats Stripe’s $1.1 billion purchase of Bridge in 2024 and stands as the largest stablecoin-infrastructure deal struck (Mastercard press release, 17 March 2026). A network does not pay more than twice a recent private mark, in cash and earn-out, for something it considers optional. It pays that to own a rail before a rival or a large acquirer does.
What forced the timing was pressure on the acquiring side. Stripe and Shopify put USDC-on-Base settlement in front of merchants across 34 countries with local-currency or external-wallet payout, announced on 12 June 2025, and Stripe reported $94 billion of stablecoin volume settled over two years, monthly volume rising from under $2 billion to over $6.3 billion (Stripe newsroom, 12 June 2025). That landed roughly a year before the BVNK agreement, so it is the move the schemes are answering rather than fresh news, and the GENIUS Act framework that makes a compliant stablecoin bankable had by then removed the legal ambiguity that kept such paths experimental (S.1582, 119th Congress). Once a clean bypass exists, the incumbent either owns it or watches volume cross it, and Mastercard chose to own it.
The case against this reading is real and deserves stating plainly. The value-added-services figures are Mastercard’s own, audited at the segment level rather than line by line, and “pricing” sits last among the named drivers, which leaves open how much of the 22% is genuine new demand and how much is the scheme charging more for what it already sold. The stablecoin bet could fail outright, and a regulator that watched interchange relief evaporate into uncapped services could extend a cap to the services themselves, which is the one move that would unwind the whole strategy. Nothing here proves the migration offsets interchange compression rather than hedging a slice of it; a network can grow complements at 22% and still erode if the regulated base shrinks faster.
Grant all of it, including that the offset is unproven and may stay unproven for several quarters, and the position holds, because it does not depend on the offset being complete. It depends on the direction, and the direction is set by a $1.8 billion cheque and explicit guidance, not by a soft signal. Even if BVNK underperforms, the act of moving revenue off the capped line and onto rails the scheme owns is the durable manoeuvre, and the regulatory response that would matter, a cap on the services layer, does not exist and is not drafted. The contest is no longer whether the schemes are escaping the interchange perimeter, which they visibly are, but how long it takes a regulator to notice that the perimeter has moved.
For diligence through 2027, two conclusions follow and both cut against the consensus. Anyone modelling the swipe-fee settlement, the Credit Card Competition Act, or the European caps as durable savings on acceptance cost is modelling a transfer the schemes are already routing around; the relief is real on the interchange line and recaptured on the services line, so the merchant’s all-in cost of acceptance is set by the layer no one regulates and should be underwritten there. And any company whose pitch is that stablecoins or agents will let merchants bypass the networks is, on this evidence, building either an acquisition target for a scheme or a toll-payer to one, which is a different asset than the disruptor its deck describes. The number worth extracting from any payments business is the share of revenue that still sits on interchange or interchange-linked acceptance fees, because that share is what regulation can still reach, and the schemes have shown they intend to keep shrinking it.