The Financial Times reported in mid-July that Mastercard was examining a sale of a majority stake in Vocalink, the company contracted to run the infrastructure that sits behind core UK payments networks Bacs, Faster Payments and the LINK ATM network. The reported buyer is DeliveryCo, the industry-owned company being stood up to build the UK’s next-generation payments infrastructure, and the reported terms are a 51% stake at an implied value of roughly £400m. None of that is confirmed. Mastercard declined to comment to both outlets that asked, and the FT’s own sourcing, relayed by PYMNTS and citing two people briefed on the discussions, puts the talks at a very early stage and any deal as unlikely before next year, since DeliveryCo is still setting up its funding and governance.
While a deal may or may not materialise, I thought it would be interesting to examine what the implications of such a deal would be for the UK payments market, why Mastercard would be inclined to sell and why now?
American Banker’s John Adams got to the obvious framing first, and got it right as far as it goes. This is payment protectionism, a US network divesting a “nationally sensitive asset” to keep its goodwill with a government that’s increasingly uneasy about American control of critical infrastructure. This is fair, but for me that framing treats the sale as something being done to Mastercard.
When you start to look past the deal itself and to the numbers behind it, a second force comes into view that so far the coverage hasn’t followed. It appears to me that two different parties want the same company to change hands, for different reasons that happen to point the same way.
What’s actually for sale is Vocalink, not “the UK’s payment rails”
For those not well versed in UK payments infrastructure, it’s important to understand what is reportedly in discussion to be sold. If you get the ownership chain wrong here, you will likely jump to different conclusions.
Pay.UK is the Payment System Operator. It owns the Bacs and Faster Payments schemes, the rulebooks, the membership and the governance, and contracts the plumbing itself out to an operator. In December 2025 it extended three infrastructure contracts, for Faster Payments, Bacs and the Image Clearing System, to Vocalink, into the early 2030s. LINK, the UK’s ATM network, is a separate scheme again, owned by Link Scheme Holdings Ltd, the recognised operator under banking law, which has its own long-term contract with Vocalink for switching.
Put simply, you have four systems, two scheme owners and one contracted infrastructure operator underpinning all of them.
Neither scheme owner has shareholders in the equity sense, and both are membership-governed, not-for-profit bodies. So what’s Mastercard’s role in all of this? Well, Mastercard doesn’t own any of the schemes. It owns the operator, and since February 2025 it has owned it outright, after Santander, the last minority holder, sold its stake in the holding company.
The reported sale would move the ownership of one column, and only one.
Scheme ownership, operation and settlement for the four systems Vocalink runs, with the Bank of England’s own systems alongside.
| System | Scheme ownerUNCHANGED | Infrastructure operator
 TODAY · MASTERCARD 100%
AS REPORTED · DELIVERYCO 51%MASTERCARD 49% · MINORITY
|
SettlementUNCHANGED | Contract | |
|---|---|---|---|---|---|
RUN BY VOCALINK |  Bacsdirect debits and credits | Pay.UK | Vocalink | Bank of England | to early 2030s |
 Faster Paymentsreal-time payments | Pay.UK | Vocalink | Bank of England | to early 2030s | |
 Image Clearing Systemcheque imaging | Pay.UK | Vocalink | Bank of England | to early 2030s | |
 LINKATM network | Link Scheme Holdings Ltd | Vocalink | Bank of England | long-term contract | |
BANK OF ENGLAND OPERATED |  CHAPShigh-value payments | Bank of England | Bank of England owned and operated | Bank of England | owned outright |
| RTGSthe settlement engine | Bank of England | Bank of England owned and operated | Is the settlement layer | owned outright, renewed Apr 2025 |
The scheme owners do not change, and settlement stays across accounts at the Bank of England. Vocalink would keep operating all four systems under its existing contracts into the early 2030s, whilst the UK decides what the next generation looks like.
Swipe the table sideways, or expand it
That distinction isn’t me just being pedantic. A reported sale of Mastercard’s stake in Vocalink doesn’t touch Pay.UK’s ownership of Bacs, Faster Payments or the Image Clearing System, or Link Scheme Holdings Ltd’s ownership of LINK. So the sovereignty concern survives that correction perfectly well. Whoever owns Vocalink controls the infrastructure that clears over 90% of UK salaries, switches Faster Payments, and runs LINK’s ATM network.
So let’s try and understand the scale the operator (Vocalink) actually runs, using the numbers the scheme owners publish themselves. Pay.UK’s most recent annual statistics put Bacs at 6.864bn transactions in 2025, up 0.8% on the year, and Faster Payments at 5.548bn, up 9.0%, a step down from 13.1% growth the year before. The Image Clearing System processed 80.8m cheque images, down 15.9%. LINK’s own January 2026 release puts 2025 at 1.27bn transactions, down 8% on 2024’s 1.388bn, with £76.7bn withdrawn. Add LINK’s 1.27bn to Pay.UK’s own combined Bacs, Faster Payments and ICS total of 12.5bn, and the four systems Vocalink processes for come to roughly 13.8bn transactions in 2025. That is most of the plumbing British households touch every month.
So if we break down the sovereignty concern further, one layer is already sovereign. Bacs, Faster Payments and LINK all settle across accounts at the Bank of England; CHAPS, the UK’s high-value system, is owned and operated by the Bank outright, on an RTGS core it finished rebuilding in April 2025. What’s actually in play in the reported Vocalink sale is the retail clearing and switching layer specifically, whilst settlement itself already sits on the Bank’s own books. That’s also why Victoria Cleland, the Bank’s own Chief Cashier, chairs the RPIB.
The UK has tried this before, and the last attempt didn’t go cleanly
DeliveryCo isn’t the first plan to loosen this layer’s dependence on a single incumbent. The New Payments Architecture began life in 2015 as a blueprint from the Payment Systems Regulator’s (PSR’s) own Payments Strategy Forum, one new central infrastructure meant to replace Bacs and Faster Payments’ separate systems outright. The New Payment System Operator was created in 2017 and was renamed Pay.UK in October 2018, inheriting a procurement it hadn’t designed. In 2021, the Payment Systems Regulator found “unacceptably high risks” that the programme, as scoped, would not deliver value for money, and forced Pay.UK to narrow it drastically, in effect critiquing a design its own forum had produced years before Pay.UK existed to run it. It was quietly reset in 2025 into something called Interbank Infrastructure Renewal, folded into the National Payments Vision’s new delivery model, and Vocalink stayed the incumbent operator throughout, with its contracts simply extended into the next attempt. All in all, a pragmatic approach to risk management given the pivot, and one that bought everyone involved some time.
That history says little about Vocalink’s own competence, since the PSR’s finding concerned programme risk and value for money. It’s mainly a corrective to the assumption that owning the operator solves the harder problem. Buying Vocalink’s equity is a much smaller act than building a working replacement from scratch, a distinction easy to lose sight of.
The UK’s own recent record on rebuilding this infrastructure without changing who runs it is one attempted rebuild, one narrowed scope, and one initiative rebranding.
The governance chain now standing behind the fresh attempt is at least more discernible than the last one. The National Payments Vision sets the direction. The Payments Vision Delivery Committee sets strategy. The Bank of England’s Retail Payments Infrastructure Board, or RPIB, turns that strategy into design, and DeliveryCo, chaired by Barclays UK’s Vim Maru, is expected to build and procure it.
Pay.UK sits inside this as the operator of today’s systems throughout, holding an automatic seat on the RPIB. HM Treasury confirmed in April 2026 that the Payment Systems Regulator’s functions will be consolidated entirely into the Financial Conduct Authority, pending legislation, so the regulator that found the NPA’s risks in 2021 likely won’t exist as a separate body by the time DeliveryCo truly starts negotiating anything.
Vocalink’s own accounts explain the price, before Mastercard’s strategy does
One aspect which raised a few eyebrows from the FT’s coverage was the reported price for the 51% stake in Vocalink. Comparisons have already been made in the broader coverage. PYMNTS set the 2016 £700m against the reported £400m within days of the story breaking, and Payment Expert reported the same two figures alongside each other. What that coverage hasn’t done is open the one document that actually explains the gap between them. Payments:Unpacked, the newsletter Northey Point runs, opened Vocalink Limited’s own 2024 accounts, filed at Companies House, on the day the story broke.
Vocalink Limited turned over £221.5m in 2024, up 1.5% on 2023’s £218.2m, the same year Faster Payments volume grew 13.1% on a scheme Vocalink doesn’t own. The composition of that turnover matters too. Revenue from external customer contracts grew 6.3% to £177.1m, whilst internal revenue charged to other Mastercard entities fell 15.3% to £44.4m as intra-group licence fees ended, so even the external line grew at half the pace of the rail. Pay.UK’s and LINK’s contracts with Vocalink aren’t public, so the accounts show the result, without providing the pricing mechanics. Part of that result is a mix effect, with cheque volumes through the Image Clearing System falling 16.9% in 2024 and LINK’s own transactions falling 7%, both sitting in the same combined book as Faster Payments’ growth. The company also swung to an operating loss of £20.0m and a loss before tax of £12.4m (2023: a £1.9m profit). Management’s own explanation is specific. Costs rose 8.2%, predominantly from investment (capex up from £13.3m to £21.2m) and a financial penalty, and the company expects to return to operating profitability in 2025. Some of that is simply what infrastructure demands. Running critical national systems at growing volumes means refreshing and scaling the technology underneath them, so a portion of the forgone profit is investment the volumes themselves require, rather than money going missing.
Statutory accounts for a wholly owned subsidiary are an imperfect window in any case, and these disclose intra-group IP licence arrangements changing during the year too, with the intellectual property for Bacs and other products returning to Vocalink from another Mastercard entity. Read that way, a single loss-making year with a named cause and real caveats is weaker evidence of a structural ceiling than the revenue line sitting next to it. 1.5% growth against a rail growing 13.1% is the number actually carrying the argument.
The penalty itself was £11.9m, settled with the Bank on 26 June 2025 and announced that July, for failing to fix governance and risk-management shortcomings by a February 2022 deadline the Bank itself had set, the first such fine against a financial market infrastructure firm. The accounts provide for it in full as a 2024 charge, having disclosed it only as a contingent liability the year before, and the penalty has since been paid. The Bank’s own final notice sets out the mechanics. A £20m penalty was reduced 15% for cooperation and a further 30% for settlement, imposed because the firm had confirmed compliance with a 2021 remediation direction that an independent reviewer later found had not been delivered. That provision is also most of the cost story. Of the £18.3m by which costs rose in 2024, £11.9m is the penalty, so strip it out and costs rose nearer 3%, which makes management’s one-off reading of the loss more credible, not less.
Sovereignty carries some value independent of margin, a separate, fair argument for later. But the more impactful objection would be that there is no gap here at all. The price is flat because Vocalink’s own revenue grew 1.5% against a rail that grew 13.1%, which explains the whole story without any theory about Mastercard’s portfolio. Vocalink’s own revenue line is the mechanism, and Mastercard’s own choice is the separate question it leaves open. Pay.UK and LINK have just extended Vocalink’s contracts into the early 2030s, so a company happy to hold a loss-making asset through a decade-long cycle would normally fix it or fund it. Mastercard’s own numbers explain why it might sell into a flat price now.
Mastercard’s own numbers point to where the capital is going instead
Mastercard’s own results say where that capital allocation preference actually sits. In its second-quarter 2026 results, published at the end of July, value-added services and solutions net revenue grew 20%, against 10% for the payment network, with almost nothing of that from acquisitions and dispositions this time. The full 2025 year told the same story, with value-added services net revenue up 21%, or 18% excluding acquisitions on a currency-neutral basis, growth Michael Miebach called “a clear demonstration” of the company’s growth algorithm. Nearly 40% of Mastercard’s transactions are now tokenised, Mastercard Move grew transactions more than 35% in 2025 across over 17bn endpoints, and in June the company launched Agent Pay for Machines, its bet on agentic and machine-to-machine commerce, with more than thirty launch partners settling across cards, bank accounts and stablecoins.
None of that points to Mastercard exiting infrastructure altogether, and it would be naive to read it that way. In October 2025 the company committed €250m to three new data centres in France, and Mastercard’s European president, Kelly Devine, framed it explicitly as building resilience against “natural or geopolitical events,” the same kind of sovereignty-adjacent infrastructure it’s reportedly negotiating to sell in Britain. The difference is differentiation. France is a moat Mastercard prices and controls end-to-end. A regulated interbank switch it runs under contract, on pricing Pay.UK sets, carries a clear ceiling.
The historical valuation comparison provides colour on that mechanism, and it warrants real caveats. Mastercard paid roughly £700m, after adjusting for cash and certain other estimated liabilities, for 92.4% of VocaLink in 2016, plus an earn-out of up to £169m never confirmed as paid. If we treat the £700m base as implying a whole-company value then of about £757m, the reported £400m for 51% today implies about £784m, on the same rough division. Both are control stakes, though they sit on different terms. A 92.4% buy-out of a whole company prices on a different basis than a 51% stake sold to a buyer whose seller stays on the register with 49%, and grossing a control stake to 100% assumes the retained minority is worth the same per share as the block sold, an assumption that typically overstates it.
Net it out and the price has moved very little in nominal terms over a decade in which UK consumer prices rose by roughly 37%. If you uprate the 2016 deal’s implied £757m by that rise, it comes to roughly £1,037m in 2026 money, against the £784m the reported terms imply, a real-terms fall of about a quarter.
Britain may be buying the pipe at the moment the economics moved upstairs
Here’s another genuine tension the sovereignty and control framing skips past. Sovereignty over critical infrastructure has value that doesn’t reduce to today’s margin mix. Resilience, data residency, and the ability to set the rules for a systemically important utility without asking permission from a foreign parent are amongst the objectives the Bank of England’s own delivery model names explicitly, independent of what the asset currently earns. I’d put that as a fair, standing rebuttal that isn’t fully answered by anything below.
Owning the rail is not the same as owning the value, and that gap is exactly what I’ve been arguing is reshaping payments more broadly.
In The Long Trace, I’ve made the case that money is a ledger entry reverting to that form after a roughly seventy-six-year detour through cards, the one stretch in five millennia of monetary history where settlement got slower and acceptance costs got higher at the same time.
For card networks, the consequence is that the bundled toll of the card era is unbundling. Disputes, guarantees, credit and identity are being rebuilt as separate, priced services on top of instant account-to-account rails. If Britain reclaims Vocalink, it reclaims the processing layer just as the industry’s own profit pool has settled somewhere else entirely, in layers a bare infrastructure contractor was never built to capture.
None of that is an argument against the deal. DeliveryCo’s mandate, per the Bank of England’s own framing, is to build next-generation infrastructure, a bigger job than inheriting Vocalink’s existing book of contracts, and a UK-owned operator with a clean governance chain behind it is a defensible outcome on sovereignty grounds alone. If the reported deal happens, Britain gets a strategically important asset it can govern and ensure continues to operate on its own terms. Whether it also gets a bigger share of where payments profit actually sits is a separate question, because that share was never inside Vocalink to begin with.
The £400m is unconfirmed and may never be the number that gets struck. The piece worth watching is whether DeliveryCo’s brief, once it has an owned operator to build on, extends upward into the dispute-and-guarantee layer the essay describes, or stops at replacing the wires it already has.
The RPIB’s design consultation, open since 25 June with responses due 11 September, already reaches beyond like-for-like replacement, naming account-to-account payments at the point of sale as an additional option to cards. Whether disputes and guarantees follow into that blueprint is the part worth reading it for. Sovereignty over the wires is a real win, but it is the smaller half of the prize. The disputes, guarantees and identity built on top of a rail are where the money actually went, and no change of shareholder puts them back on the rail.