On Singles’ Day 2024, Alipay’s payment system processed a peak of 988,000 transactions per second. To put that in context, Visa’s VisaNet, the largest card network in the world, has a tested capacity of 65,000 transactions per second. The gap between these numbers reflects a broader shift in how global transactions are processed. Card networks that dominated payment infrastructure for five decades are now competing with fintech platforms that were built for mobile-first, API-driven, real-time commerce. The global digital payments market will reach $36.09 trillion in total transaction value by 2030, according to Statista, growing at 7.63% annually from $23.64 trillion in 2025. Fintech companies are processing an increasing share of that volume.
The Card Network Model and Its Limitations
Visa and Mastercard built their businesses on a four-party model: the cardholder, the merchant, the issuing bank (cardholder’s bank), and the acquiring bank (merchant’s bank). Every transaction passes through all four parties, with each taking a fee. The total cost to the merchant ranges from 1.5% to 3.5% of the transaction value, depending on the card type, geography, and merchant category.
This model works well for in-person purchases in developed economies where card penetration is high. In the United States, 87% of adults have a debit card and 77% have a credit card. Card acceptance is nearly universal. The infrastructure is mature, reliable, and profitable for all participants.
The Boston Consulting Group projects fintech revenues will reach $1.5 trillion by 2030, with embedded finance and digital lending accounting for the largest share of projected growth.
According to CB Insights’ 2024 fintech report, global fintech funding declined 40 percent between 2022 and 2024, pushing the sector toward consolidation and a sharper focus on profitability over growth at all costs.
The model breaks down in three scenarios. First, in markets where card penetration is low. In India, only 5% of the population has a credit card. In Indonesia, the figure is 3%. In Nigeria, it is under 2%. Second, for small-value transactions where the fixed component of card fees ($0.20 to $0.30 per transaction) makes the economics unfavorable. A $2 coffee purchased with a credit card can cost the merchant $0.37 in processing fees, an 18.5% cost. Third, for cross-border transactions where correspondent banking fees, currency conversion charges, and intermediary markups push total costs to 5% to 8%.
Fintech companies have built alternative infrastructure for each of these scenarios. UPI in India (government-built, free for consumers) bypasses card networks entirely. Digital wallets like GCash and GrabPay serve Southeast Asian markets where cards never achieved mass adoption. Cross-border specialists like Wise and Airwallex route payments through local banking rails instead of correspondent banking chains.
How Fintech Infrastructure Differs from Traditional Rails
The technical architecture of fintech payment systems differs from card networks in several important ways.
Card networks use a store-and-forward model designed in the 1970s. A transaction is authorized in real time (the merchant gets approval within seconds), but the actual money movement happens in batch settlement cycles, typically 24 to 48 hours later. This delay creates float, which banks monetize, and introduces reconciliation complexity for merchants who must match authorizations with settlements.
Real-time payment systems like UPI, Pix, and FedNow use a push-payment model. The payer initiates the payment, the money moves immediately, and settlement is final. There is no authorization-settlement gap. No float. No reconciliation delay. The transaction is complete in seconds.
API-first payment companies like Stripe, Adyen, and Checkout.com add a programmability layer on top of both models. A merchant integrating Stripe’s API can route a transaction to a card network in the United States, to iDEAL in the Netherlands, to Pix in Brazil, and to UPI in India, all through a single integration. The routing logic can be customized: send transactions under $10 through the cheapest rail, send transactions from high-risk geographies through stricter fraud screening, automatically retry failed transactions on an alternative network.
This programmability is what distinguishes fintech infrastructure from traditional rails. A card network offers one path for a transaction. A payment orchestration platform offers dozens, and the selection can be optimized in real time for cost, speed, conversion rate, or fraud risk.
The Scale of Fintech Transaction Processing
The volume of transactions processed by fintech companies has reached a scale that rivals traditional financial institutions.
| Company | Annual Payment Volume (2024) | Founded | Valuation/Market Cap |
|---|---|---|---|
| Stripe | $1T+ (estimated) | 2010 | $65B (private) |
| Adyen | €939B processed | 2006 | ~$45B (public) |
| Block (Square) | $210B gross payment volume | 2009 | ~$40B (public) |
| PayPal | $1.53T total payment volume | 1998 | ~$70B (public) |
| Wise | $118B cross-border volume | 2011 | ~$10B (public) |
Sources: Company earnings reports, Statista
PayPal alone processes more transaction volume than the GDP of every country except the United States, China, Japan, and Germany. Stripe’s estimated $1 trillion in annual processing volume means roughly 3% of all global e-commerce transactions flow through its infrastructure. These companies are no longer startups competing at the margins. They are systemically important payment processors.
Cross-Border Transactions: Where Fintech Has the Largest Advantage
Cross-border payments represent roughly $150 trillion in annual B2B flows and $800 billion in consumer remittances. This is where fintech companies have created the most value, because it is where traditional infrastructure is most expensive and slowest.
The correspondent banking model for cross-border payments involves multiple intermediaries. A payment from a business in São Paulo to a supplier in Shanghai might pass through four banks: the sender’s bank in Brazil, a correspondent bank in New York, a correspondent bank in Hong Kong, and the recipient’s bank in China. Each intermediary charges a fee ($15 to $50) and adds processing time (one to five days). The total cost can reach 5% to 8% of the transaction value.
Fintech alternatives reduce both cost and time. Wise uses a matching model that pairs opposing currency flows, avoiding international wire transfers entirely. A customer sending dollars to a euro account is matched with another customer sending euros to a dollar account. The money stays domestic on both sides, and Wise handles the balancing. This model brings the average fee to 0.62%, roughly one-tenth of the correspondent banking cost.
Airwallex takes a different approach, maintaining local bank accounts and licenses in over 20 markets. A business using Airwallex can collect payment in Australian dollars from an Australian customer, hold the funds in an Australian account, and disburse them in Indonesian rupiah to an Indonesian supplier, all through a single API. No correspondent banks. No SWIFT messages. Settlement in hours rather than days.
Thunes connects mobile money networks, bank accounts, and digital wallets across 130 countries, enabling transactions between systems that do not natively interoperate. A GCash wallet in the Philippines can send money to an M-Pesa account in Kenya through Thunes’ network, a corridor that traditional banking infrastructure does not serve at all.
What Determines Who Wins
The fintech payment infrastructure market is consolidating around a few competitive dimensions.
Geographic coverage matters most for companies serving international merchants. A payment processor that supports 10 markets is useful. One that supports 40 markets is significantly more valuable because it eliminates the need for the merchant to maintain relationships with multiple providers. Adyen’s single-platform model (one integration for 37 countries) has been a key differentiator in winning enterprise clients like McDonald’s, Spotify, and Microsoft.
Approval rates are becoming the primary competitive metric. The payment processor that approves the highest percentage of legitimate transactions while blocking fraud generates the most revenue for the merchant. A 1% improvement in approval rate on $1 billion in volume represents $10 million in additional revenue. This is why Stripe, Adyen, and Checkout.com invest heavily in machine learning models that analyse transaction patterns across their entire network to optimize authorization decisions.
Developer experience determines which processor captures new merchants. Stripe’s documentation, SDKs, and developer tools set a standard that competitors are still trying to match. The first payment processor a developer integrates during a company’s startup phase often remains in place as the company scales, because re-integrating payments touches every part of the business from checkout to accounting to tax compliance.
The $36 trillion in digital payment transaction value projected for 2030 will flow through infrastructure built by companies that did not exist 20 years ago. The card networks will remain important, particularly in developed markets. But the growth is happening in real-time systems, API-first platforms, and cross-border corridors where fintech companies have built infrastructure that is faster, cheaper, and more programmable than anything the traditional banking system offers.
