Business models
Strategic moats
Part IThe Story
Every day, approximately 329 billion times a year, a signal travels from a plastic rectangle — or a phone, or a watch, or increasingly a line of code — through a network that touches more than 200 countries, and the answer comes back in roughly 1.8 seconds. Approved or declined. The message is trivially small, a few hundred bytes, and yet that message is the global economy in miniature: a buyer's intent, a seller's trust, a bank's guarantee, and an invisible toll collector sitting between all three. Visa Inc. does not lend money. It does not hold deposits. It does not issue a single card. It builds nothing that the end consumer ever touches. What Visa operates, at the most irreducible level, is a protocol — a set of rules, a routing system, a trust layer — that has embedded itself so deeply into the infrastructure of modern commerce that most people cannot distinguish between Visa and money itself.
The distinction matters, because it explains everything that makes this business historically unusual: net revenue margins that routinely exceed 50%, a market capitalization that dwarfs the banks who collectively birthed it, and a competitive position so durable that the Department of Justice has spent the better part of two decades trying to figure out whether it constitutes a monopoly. The answer keeps arriving in the form of more transactions, processed at higher volumes, on more devices, in more countries, by a company that employs roughly 30,000 people to supervise a machine built for a trillion dollars of annual payment volume.
By the Numbers
The Visa Machine
~$35.9BNet revenue, FY2025 (est.)
~329BTotal transactions processed annually
4.9BPayment credentials worldwide
~175M+Merchant acceptance locations
~14,500Financial institution clients
~$600B+Market capitalization
~67%Operating margin (FY2024)
200+Countries and territories served
The Fresno Drop
The origin story is, by any reasonable standard, insane. In September 1958, Bank of America — then the largest bank in the United States, the institution A.P. Giannini had built by lending to immigrants and farmers in California — mailed 60,000 unsolicited credit cards to the residents of Fresno, California. No applications. No credit checks of the sort we'd recognize today. Just cards, arriving in mailboxes, pre-approved and ready for use, in what the bank internally called "the Fresno Drop."
The idea belonged to a middle manager named Joseph Williams, who had studied the failures of earlier charge card programs — Diners Club, launched in 1950 as a restaurant charge card for Manhattan's expense-account class; Carte Blanche; various retailer cards — and concluded that the fundamental problem was chicken-and-egg. Merchants wouldn't accept a card unless consumers carried it. Consumers wouldn't carry a card unless merchants accepted it. Williams's solution was brute force: saturate an entire city with cards simultaneously, then recruit merchants by pointing to the installed base. Fresno was chosen because it was a midsized, self-contained California city where Bank of America had dominant market share.
The result was both a triumph of market creation and an operational catastrophe. Fraud was immediate and rampant. Delinquencies soared. The BankAmericard program lost an estimated $20 million in its first year — a staggering sum for a bank in 1959. But the concept worked: merchants signed up because customers had cards, and customers used the cards because merchants accepted them. Bank of America had, through a controlled detonation, ignited the first self-sustaining two-sided credit card network in history.
— Dee Hock, Visa founder, reflecting on the early card eraMoney is nothing but alphanumeric data… it is guaranteed alphanumeric data. It is a mathematical guarantee by one party to another.
What Bank of America had stumbled into — and would take nearly two decades to properly understand — was not a lending product but a network topology. The credit function was almost a red herring; the real innovation was the authorization and settlement system that allowed a transaction to occur between two strangers, intermediated by two separate banks, in real time. Every subsequent development in Visa's history, from consortium formation to IPO to its current war with the DOJ, flows from this structural insight: the network is the product, and the network is the moat.
The Consortium and the Chaordic Vision
By the mid-1960s, Bank of America faced a problem it could not solve alone. The BankAmericard had proven the concept, but American banking law — specifically the McFadden Act of 1927 and the Douglas Amendment of 1956 — prohibited interstate branch banking. A bank in California could not issue cards to customers in New York. So in 1966, Bank of America began licensing the BankAmericard brand and system to other banks across the country. The licensees could issue their own BankAmericards, process transactions through a shared system, and settle with each other through Bank of America as the central node.
This was the birth of the four-party model that still governs global card payments: the cardholder, the merchant, the issuing bank (the cardholder's bank), and the acquiring bank (the merchant's bank). The network operator — BankAmericard, later Visa — sits in the middle, routing authorization messages and managing settlement between issuer and acquirer. Critically, the network operator never touches the money. It moves information about money, and charges a fee for the privilege.
But the licensing model quickly became unruly. By 1968, the program had hundreds of participating banks, spiraling fraud, chaotic settlement, and a governance structure that essentially required competitors to cooperate. Enter Dee Hock.
Hock was a loan officer at a small bank in Seattle — the National Bank of Commerce — that had joined the BankAmericard program. Born in 1929 in North Ogden, Utah, he was an autodidact and an organizational theorist before such a thing had a name, consumed by questions about how complex systems self-organize. He saw the BankAmericard chaos not as a management failure but as an architectural one: the system was trying to be both centrally controlled (by Bank of America) and inherently distributed (across thousands of independent banks). Hock believed a new kind of organization was needed — one he would later call "chaordic," a portmanteau of chaos and order. As he wrote in One from Many: VISA and the Rise of Chaordic Organization, the goal was to create an entity owned by its members but governed by principles rather than hierarchy, capable of evolving without central command.
In 1970, Bank of America agreed to relinquish control. The licensed banks formed National BankAmericard Inc. (NBI), a membership corporation, with Hock as its first president and CEO. He was 41 years old, running an organization that existed more as a concept than an institution. Over the next six years, Hock did three things that would prove foundational:
🔗
Hock's Three Moves
The structural decisions that built the Visa network
1970–73
Built the electronic authorization system, BASE I and BASE II, which allowed real-time transaction routing among thousands of banks — the first version of what would become VisaNet.
1974
Expanded internationally, creating IBANCO to coordinate non-U.S. bank members, a structure that would evolve into Visa's federated global organization.
1976
Renamed the system "Visa" — a word Hock chose because it was pronounceable in virtually every language, evoked permission and passage, and carried no national identity. The blue-and-gold stripe was designed to suggest a sunset (or sunrise, depending on your mood).
The name change was more than branding. It was a declaration that the system had transcended its origin as a Bank of America product. Visa was now a protocol — a standard for how money moved — owned collectively by the banks that used it. Hock's insight was that in a network, the central operator's power comes not from control over participants but from the rules of engagement between them. He wrote the rules. The banks competed with each other on everything else.
The Invisible Architecture
To understand why Visa is one of the most profitable businesses ever constructed, you have to understand what happens in the 1.8 seconds between a card swipe and an approval beep.
A consumer taps their Visa card at a coffee shop. The merchant's point-of-sale terminal sends the transaction data — card number (or token), amount, merchant category code — to the merchant's bank, the acquirer. The acquirer routes the authorization request through VisaNet to the cardholder's bank, the issuer. The issuer checks the account: Is there credit available? Is the card reported stolen? Does the transaction pattern match the cardholder's history? If yes, the issuer sends an approval code back through VisaNet to the acquirer, which sends it to the terminal. The coffee shop hands over the latte.
This happens, on average, 639 million times per day.
The transaction is not yet settled — actual money won't move between banks until the end of the day or later — but the authorization is a binding commitment. The issuer has guaranteed payment. The merchant has certainty. And Visa has earned approximately 10 to 15 basis points of the transaction value for routing the message.
The economic architecture is astonishingly elegant. Visa's revenue comes from three primary streams. First, service fees — charged to issuers based on the payment volume of Visa-branded cards, essentially a royalty on the network's brand and infrastructure. Second, data processing fees — charged for each transaction that travels through VisaNet, a pure throughput fee. Third, international transaction fees — a premium charged when the issuer and acquirer are in different countries, reflecting the additional complexity of cross-border settlement and currency conversion.
Crucially, Visa does not collect interchange — the 1.5% to 3% fee that flows from the acquirer to the issuer on every transaction. Interchange is the lifeblood of the card ecosystem, funding rewards programs, fraud protection, and issuer economics, but it flows through the Visa network, not to Visa. Visa sets the interchange rates — and this is where much of the regulatory and legal risk lives — but the money goes to the banks. Visa's own take is a fraction of a percent, multiplied by enormous volume.
The result is a business with almost no credit risk, no balance sheet exposure, negligible marginal cost per transaction, and operating margins that have consistently exceeded 65%. Visa is, in the taxonomy Alex Rampell of Andreessen Horowitz has used, "the original protocol business" — more analogous to TCP/IP than to a bank, except that it charges a toll on every packet.
— Alex Rampell, General Partner, Andreessen HorowitzVisa is really a software company. The entire product is software and a brand.
The Debit Offensive
For the first thirty years of its existence, Visa was primarily a credit card network. Debit — the ability to spend directly from a bank account — was a sideshow. That changed in the 1990s, and the way it changed reveals something important about Visa's strategic instincts.
Debit cards existed before Visa got involved, running over regional ATM networks like STAR, NYCE, and Pulse, which used PIN authentication. These networks charged low fees — a flat rate per transaction, typically under a dime — and had no ambitions beyond ATM interoperability. Visa saw an opportunity: create a signature-based debit card that ran over the Visa network, looked and felt like a credit card to the consumer, and carried interchange rates far higher than PIN debit.
The Visa Check Card, launched in 1994, was a masterstroke of product positioning. From the consumer's perspective, it was a debit card — money came directly from their checking account. From the network's perspective, it was routed like a credit transaction, through VisaNet, with Visa-level interchange fees. Banks loved it because they earned more per transaction than PIN debit. Merchants hated it because they paid more. Consumers were largely indifferent — they tapped or swiped and moved on.
By the early 2000s, Visa's signature debit volume was growing faster than any other product segment. This triggered the first major wave of merchant litigation: the Wal-Mart case, formally In re Visa Check/MasterMoney Antitrust Litigation, filed in 1996 and settled in 2003 for $3 billion — at the time, the largest antitrust settlement in American history. The merchants' core claim was that Visa and Mastercard used "honor all cards" rules to force merchants who accepted Visa credit cards to also accept the higher-fee signature debit cards, even when cheaper PIN alternatives existed.
Visa settled. It adjusted its rules. And then it kept growing debit volume, because the underlying consumer behavior was irreversible — people wanted to pay with plastic, and debit was eating checks faster than credit was eating cash. The Durbin Amendment, passed as part of the 2010 Dodd-Frank Act, capped debit interchange for large issuers at roughly 21 cents plus 0.05% per transaction, cutting revenue to issuers and indirectly squeezing Visa's pricing power on debit. Visa absorbed it. The machine adapted and kept running.
The IPO That Ate Its Creators
For nearly four decades, Visa existed as one of the strangest entities in corporate history: a nonprofit membership association, owned collectively by thousands of competing banks, governed by committees, generating no profit for itself because every dollar of value was passed through to its members. It was, by Dee Hock's design, more akin to a cooperative utility than a company. And it was, by any financial measure, wildly undervalued.
The restructuring that transformed Visa from a cooperative into a corporation — and then into a public company — was driven by two forces: litigation risk and competitive envy. Mastercard had gone public in 2006, and the market's reception was rapturous. Visa's member banks saw what a public listing could do: monetize their ownership stakes, fund litigation reserves, and separate themselves from the legal liabilities accumulating around interchange fees.
The restructuring was Byzantine. Visa had evolved into a federated structure: Visa U.S.A., Visa International, Visa Canada, and Visa Europe, each with its own governance and membership. In October 2007, these entities (except Visa Europe, which remained member-owned) were merged into a single Delaware corporation, Visa Inc. The member banks received shares — Class B common stock, carrying litigation liability obligations, and Class C common stock — while newly issued Class A shares would be sold to the public.
📈
The Largest IPO in History (at the time)
Visa's March 2008 public offering
Nov 2007
Visa Inc. files S-1 registration statement with the SEC, proposing a $10 billion offering.
Mar 18, 2008
IPO prices at $44 per share — 406 million Class A shares — raising approximately $17.9 billion in gross proceeds.
Mar 25, 2008
Shares begin trading on the NYSE under the ticker "V." First-day close: $56.50, a 28% pop.
Jun 2016
Visa acquires Visa Europe for approximately €21.2 billion, reuniting the global network under a single corporate umbrella.
The timing was cosmically terrible — and strategically perfect. Visa went public on March 25, 2008, six months before Lehman Brothers collapsed and the global financial system seized. The banks that had owned Visa saw their own stocks crater by 50%, 70%, 90%. Bear Stearns had already been fire-sold two weeks before Visa's IPO. But Visa — because it bore no credit risk, held no mortgages, had no balance sheet exposure to the instruments that were devouring Wall Street — sailed through the crisis. Its stock dropped, yes, but far less than the banks. And then it recovered faster, and kept going.
Within a few years, a remarkable inversion occurred: Visa's market capitalization surpassed that of JPMorgan Chase, Bank of America, Citigroup — the very banks that had created it, governed it, and surrendered ownership. The consortium's child had outgrown every one of its parents. As of mid-2025, Visa's market cap exceeds $600 billion, larger than all but a handful of companies on Earth. The entity that Dee Hock built as a cooperative experiment in organizational theory had become one of the most valuable corporations in the history of capitalism.
— Charlotte Hogg, CEO of Visa Europe, Harvard Business Review, 2021We should be seen as a platform-based technology company like Google, despite being an established financial institution.
The VisaNet Fortress
The physical infrastructure that processes those 639 million daily transactions is as close to indestructible as commercial technology gets. VisaNet operates across four primary data centers — in Ashburn, Virginia; Highlands Ranch, Colorado; London; and Singapore — each capable of running independently, each hardened against natural disasters, terrorism, and cyberattack. The system can handle up to 65,000 transactions per second and performs up to 100 billion computations per second in fraud screening alone.
Rajat Taneja, Visa's president of technology, joined the company in 2013 after stints as CTO at Electronic Arts and fifteen years in R&D leadership at Microsoft. Under his direction, Visa spent $3.5 billion rebuilding its data platform from scratch — not merely upgrading legacy systems but rearchitecting them for modern workloads: tokenization, mobile commerce, real-time fraud detection, and eventually generative AI. Over the same five-year period, Visa invested approximately $12 billion in technology overall.
The tokenization investment alone tells you something about the strategic imagination at work. When e-commerce exploded, the 16-digit card number on a physical card became a liability — static credentials traveling over the internet, exposed to every data breach. Visa's response was to replace the card number with a dynamic digital token, a surrogate value that could be scoped to a specific device, merchant, or transaction. By 2025, Visa had tokenized billions of credentials, and token-based transactions showed significantly lower fraud rates than traditional card-not-present transactions.
The AI investments have been even more ambitious. Visa has embedded more than 100 products with AI and generative AI technologies. Its fraud-detection models — evolved from basic rule sets to deep neural networks — analyze hundreds of data points per transaction in real time, and the company claims its AI systems help prevent approximately $40 billion in fraud annually. Taneja organized a company-wide hackathon that attracted thousands of participants and generated over 2,300 AI application ideas; by February 2023, every single employee at Visa had access to an internal secure version of ChatGPT.
— Rajat Taneja, President of Technology, Visa, Fortune, 2025The safety and management of these models is very, very important. It is a core belief of ours that, just as important as the sciences of models, is the art, which is the policy and the governance.
Over 2,500 engineers now work specifically on AI within Visa. The company works with models from OpenAI, Anthropic, Google, IBM, Meta, and Mistral AI, matching different LLMs to different use cases — software development, risk modeling, agentic workflows. The ambition is not to build AI products for external sale but to make VisaNet itself smarter, faster, more adaptive — a network that learns from every transaction it processes and feeds that intelligence back into authorization decisions, fraud scoring, and merchant analytics.
The Toll Road That Everyone Must Use
The competitive dynamics of card networks produce a peculiar result: the market is essentially a duopoly (Visa and Mastercard in most of the world, with regional players like China UnionPay, JCB, and RuPay in specific geographies), and yet the two competitors rarely compete on price in any meaningful way. Interchange rates move in lockstep. Network fee schedules are similarly opaque and similarly steep. The real competition happens on the other axis: signing issuing banks and merchants, investing in technology, and expanding into new payment flows.
This is what makes Visa's competitive position so extraordinary — and so controversial. A merchant who wants to accept credit cards effectively must accept Visa, because roughly half of all credit card spending in the United States runs on the Visa network, and consumers who carry Visa cards will simply go elsewhere if a merchant declines them. The issuing banks must participate in the Visa network because that's where the merchants are. The consumers carry Visa because that's what their banks issue and that's what merchants accept. Each side of the market is locked in by the choices of the other two.
In September 2024, the Department of Justice filed an antitrust lawsuit against Visa, alleging monopolization of the debit card market. The complaint was specific: Visa allegedly used exclusionary agreements with merchants and banks, including volume-based incentive arrangements and penalties for routing transactions over competing networks, to maintain its dominance in debit — a market where regulators had tried to introduce competition through the Durbin Amendment's requirement that every debit card carry at least two unaffiliated network options. The DOJ alleged that Visa's practices ensured that even when alternatives were technically available, the economic incentives overwhelmingly favored routing through Visa.
The case remains pending. Its outcome could reshape the economics of American payments — or it could join the long list of antitrust actions that have dinged Visa's reputation without denting its trajectory. In March 2024, Visa and Mastercard reached a separate $30 billion settlement with merchants over credit card swipe fees, one of the largest antitrust settlements in American history. Under the terms, Visa agreed to reduce interchange rates by at least 4 basis points per merchant for three years, and systemwide average interchange must fall at least 7 basis points below current levels for five years. Merchants gained the right to surcharge Visa and Mastercard transactions and steer customers to lower-cost cards.
The settlement was notable for what it revealed: even a $30 billion concession, representing billions in annual savings for merchants, was described by the Retail Industry Leaders Association as "a mere drop in the bucket." Interchange fees in the United States exceeded $100 billion in 2023. The system's sheer scale makes even massive settlements feel incremental.
The Leadership Handoffs
Visa's executive succession has been a study in institutional continuity rather than founder-driven charisma. Dee Hock, the philosophical architect, left in 1984 — exhausted, disillusioned, retreating to a ranch in the Pacific Northwest to write books about organizational theory. (His accounts of the Visa founding, Birth of the Chaordic Age and One from Many, remain the best primary sources on the consortium's creation.) The cooperative ran for decades under a succession of banking executives who managed the network competently but without strategic imagination.
The transformation came with the IPO era. Joe Saunders, a payments industry veteran, steered the restructuring and public offering as CEO from 2007 to 2012 — the transactional leader for a transactional moment. Charles Scharf, who arrived in 2012, came from JPMorgan, where he'd run the retail banking operation; he lasted until 2016 before departing for what would eventually become the CEO role at Wells Fargo during its fake-accounts crisis. A brief interregnum.
Al Kelly became CEO in December 2016. Kelly was a 23-year American Express veteran who had risen to president of that company before leaving to run Intersection, a technology company. He was, by disposition and training, an operator's operator — detail-oriented, metrics-obsessed, a systems thinker who understood that Visa's value proposition was reliability multiplied by ubiquity. Under Kelly, Visa completed the €21.2 billion acquisition of Visa Europe in 2016, reuniting the network, and pushed aggressively into value-added services — consulting, analytics, fraud management — that diversified revenue beyond pure transaction processing. Kelly served as CEO until February 2023.
Ryan McInerney, Kelly's successor, had been Visa's president since 2013. He arrived from JPMorgan Chase, where he'd run consumer banking, and is among the growing cohort of payments executives who view Visa not as a card company but as a network-of-networks capable of facilitating any movement of value — person-to-person, business-to-business, government-to-citizen.
— Ryan McInerney, CEO, Visa, Global Product Drop, April 2025As new ways to pay emerge, they need to run on a network that is always on — that is safe, secure, scalable and relentlessly innovating.
McInerney's agenda has been to expand Visa's addressable market beyond the roughly $15 trillion in consumer payments that run on cards and into the estimated $200+ trillion in global money movement that still travels by wire transfer, check, ACH, or cash. Visa Direct — the company's real-time push-payment platform, which can send funds to a card, bank account, or digital wallet — is the primary instrument of this expansion. The stablecoin partnerships announced in April 2025 — with Visa settling transactions in USDC, a dollar-pegged cryptocurrency — represent another front. So does Visa Intelligent Commerce, the initiative launched at the April 2025 Global Product Drop to embed Visa's network into AI agent-mediated transactions, with partnerships spanning Anthropic, OpenAI, Microsoft, Stripe, Perplexity, and Samsung.
The $20 Trillion Hole
Visa's investor presentations routinely cite a number: more than $20 trillion in annual underserved consumer spend, globally, that has not yet migrated from cash and check to electronic payment. This is the white space — the remaining undigitized economy — that Visa believes represents its long-term growth opportunity. In mature markets like the United States and Western Europe, card penetration is high but not complete; cash still accounts for roughly 16% of U.S. point-of-sale transactions by volume. In emerging markets — Southeast Asia, sub-Saharan Africa, Latin America, India — the cash share is far higher, and Visa competes with local digital payment schemes (M-Pesa, PIX, UPI) that may or may not run over its network.
The secular trend is unmistakably favorable. The COVID-19 pandemic compressed what might have been a decade of cash-to-digital migration into two years, as contactless payments surged and merchants that had been cash-only scrambled to accept cards. Visa's cross-border travel volumes, which had collapsed in 2020, recovered to and then exceeded pre-pandemic levels by 2023. E-commerce, which generates higher-margin card-not-present transactions, continued growing even after the post-lockdown normalization.
But the growth opportunity is not without friction. India's Unified Payments Interface (UPI) processes billions of transactions monthly at near-zero cost to merchants, creating a domestic payment rail that largely bypasses Visa and Mastercard. Brazil's PIX system, launched by the central bank, has achieved similar scale. China's payments market is dominated by Alipay and WeChat Pay, with UnionPay handling the card network layer. In each case, the pattern is the same: a domestic real-time payment system, often government-backed, that offers cheaper, faster alternatives to the international card networks.
Visa's response has been to position itself as the interoperability layer — the network that connects disparate domestic systems to each other and to the global economy. Visa Direct, the Visa B2B Connect platform, and the stablecoin initiatives all serve this strategy. The bet is that even as domestic payments go local, cross-border payments remain complex, trust-intensive, and worth paying for — and that Visa's brand, infrastructure, and regulatory relationships make it the natural clearinghouse for international value transfer.
What Visa Is Not
The most persistent misunderstanding about Visa — held by consumers, politicians, and even some investors — is that Visa is a financial institution. It is not. Visa does not extend credit, does not hold consumer deposits, does not bear credit risk, and does not set the interest rates on credit cards. Those functions belong to the issuing banks — JPMorgan Chase, Bank of America, Citigroup, Capital One, and thousands of smaller institutions — who use the Visa brand and network under license.
This distinction is the source of both Visa's extraordinary profitability and its political vulnerability. When a senator complains about "credit card fees," they are typically talking about interchange — a fee set by Visa and Mastercard but collected by issuing banks. When a consumer pays 24% APR on their revolving balance, that's a bank's pricing decision, not Visa's. But the Visa logo is on the card, and in the public imagination, Visa is the face of the system.
The business model's beauty is in its absence. No credit losses in a recession. No deposit runs in a banking crisis. No interest rate sensitivity in the way banks experience it. What Visa has, instead, is a royalty on spending — consumer spending, business spending, government spending — that grows as the economy grows and as cash converts to digital. In a recession, Visa's revenue dips with spending volumes but doesn't crater with credit defaults. In an inflationary environment, Visa benefits directly: the same basket of goods costs more, transaction values rise, and Visa's ad valorem fees (which are a percentage of transaction value, not a flat rate) increase proportionally.
This is why Visa trades at 30x+ earnings — a multiple that would be absurd for a bank but is rationally justified for a toll road on global GDP with 67% operating margins and mid-teens revenue growth.
The AI Commerce Bet
In April 2025, at its Global Product Drop event in San Francisco, Visa unveiled what it called Visa Intelligent Commerce — a framework for embedding the Visa network into AI agent-mediated transactions. The premise: as AI agents — software programs acting autonomously on behalf of consumers — begin browsing, comparing, selecting, and purchasing products, those agents will need a trusted payment mechanism. Visa intends to be that mechanism.
The strategic logic is characteristically Visa-like: identify the emerging form factor for commerce, ensure that Visa's credentials and network are embedded into it before any competitor, and let the ecosystem build around the standard. Visa did this with e-commerce in the late 1990s (adding the three-digit CVV code, developing 3D Secure authentication). It did it with mobile payments (Apple Pay launched in 2014 with Visa tokenization). It did it with contactless (NFC-enabled Visa cards are now the default in most of Europe and much of Asia). Each time, the pattern was the same: the transaction form factor changed, but the underlying network — VisaNet, routing authorization messages between issuers and acquirers — remained constant.
The AI commerce partnerships announced in April 2025 span Anthropic, IBM, Microsoft, Mistral AI, OpenAI, Perplexity, Stripe, and Samsung. The vision, articulated by Visa's Chief Product and Strategy Officer Jack Forestell, is that AI agents will handle everything from routine grocery orders to complex booking tasks, with Visa credentials serving as the default payment rail. "If there is no payment, there is no commerce," McInerney noted — a statement that functions as both strategic positioning and existential self-justification.
Whether AI commerce becomes a meaningful revenue driver in the next five years or the next fifteen is an open question. But the bet itself is consistent with sixty-seven years of pattern: Visa doesn't build the application layer, doesn't build the device, doesn't build the agent. It builds the trust and settlement layer underneath all of them.
Stablecoins and the Protocol Layer
The crypto industry has spent a decade promising to disintermediate Visa. The irony is that Visa has quietly become one of the most significant institutional participants in crypto-adjacent finance — not by embracing decentralization but by absorbing it into the existing network architecture.
Visa began piloting stablecoin settlement in 2021, using the USDC stablecoin (pegged to the U.S. dollar) to settle transactions with issuing partners. By 2025, the stablecoin partnerships had expanded significantly, with Visa positioning stablecoins as a mechanism for faster, cheaper cross-border settlement — particularly in corridors where traditional correspondent banking is slow and expensive.
The strategic calculus is transparent. If stablecoins become a meaningful medium of exchange — for remittances, for B2B payments, for commerce in countries with volatile local currencies — someone will need to provide the trust layer between the stablecoin ecosystem and the traditional banking system. Visa is placing itself in that position: the bridge between on-chain and off-chain value, earning fees for the routing, authentication, and settlement that make stablecoin payments usable by ordinary consumers and merchants.
This is not a capitulation to crypto ideology. It is the opposite — an absorption strategy, incorporating new rails into the existing network rather than ceding the payment function to decentralized alternatives. The blockchain purist's nightmare: Visa as the toll booth on the on-ramp.
The Weight of the Toll
The deepest tension in Visa's business is this: the very thing that makes it indispensable — its position as the default routing layer for global commerce — is also the thing that makes it a target. Merchants resent paying 2% on every transaction for a service that, in their view, amounts to routing a message. Regulators in multiple jurisdictions have capped or challenged interchange rates. The DOJ's 2024 debit antitrust suit alleges that Visa's market position is maintained through exclusionary practices rather than competitive merit.
The merchant perspective is not unreasonable. In a world of instant digital communication, the notion that moving a payment authorization message should cost two cents on the dollar — when sending an email costs effectively nothing — strikes many retailers as a form of economic rent. The card networks' counterargument is that the fee buys not just routing but fraud prevention, guaranteed payment, consumer protection, and the massive investment in building and maintaining a network that works 99.999% of the time in 200+ countries.
Both arguments are correct, which is why the debate has lasted two decades and will last two more. The $30 billion settlement in 2024 was not a resolution but a temporary armistice. The DOJ suit is not about interchange per se but about the competitive structure of the market — whether Visa's dominance in debit is the result of a superior product or of contractual arrangements that foreclose competition.
What the regulatory and legal landscape reveals is the paradox at the heart of any network monopoly: the network is most valuable when it is universal, but universality confers the power to extract rents that would be impossible in a competitive market. Visa's challenge, for the next decade, is to demonstrate that its pricing power reflects genuine value creation — through AI-driven fraud prevention, through tokenization, through the expansion of Visa Direct and value-added services — rather than mere incumbency.
The machine processes another 329 billion transactions. Another $15 trillion flows through 4.9 billion credentials, across 175 million merchant locations, in 200 countries. And in the space between the tap and the beep — that 1.8-second interval that contains the entire apparatus of modern trust — Visa takes its fraction of a cent, multiplied by everything.
How to cite
Faster Than Normal. “Visa — Business Strategy Analysis.” fasterthannormal.co/businesses/visa. Accessed 2026.
