At one end of the financial system, an asset manager must deliver securities against several million dollars, provide collateral and ensure that funds arrive at the right moment. At the other, a shopkeeper waits for payment confirmation on a phone before handing over a purchase. The amounts, institutions and technologies differ, but the fundamental question remains the same: how can value be transferred with enough speed, certainty and trust for economic activity to continue?

Fintech has grown within this space. The term now encompasses companies that automate market operations, distribute credit, organise international payments or allow customers to pay a bill through a mobile wallet. Yet this common label conceals profoundly different businesses. Some sell software. Others operate networks. Others lend money and bear the risk of not being repaid. A single application can combine these functions while relying, behind its interface, on several institutions.

Understanding the sector requires following this chain from its most sophisticated functions to its most basic services. This descent is not a ranking of usefulness. A complex trading mechanism may only marginally improve an already efficient market, while a reliable way to receive payments can transform a small business. It does, however, distinguish three forms of power that are frequently confused: owning the customer relationship, controlling the infrastructure and bearing the financial risk.

At the top are technologies involved in organising markets and moving assets. When an institution buys a bond, executing the trade is only the beginning of a sequence of operations. Instructions must be checked, sufficient liquidity made available, securities transferred and their new ownership correctly recorded. Depending on the instrument and market, these tasks involve intermediaries, clearing systems and custody infrastructure.

Part of financial innovation consists of reducing friction within this machinery. Software matches instructions, optimises the use of collateral or provides better visibility over liquidity requirements. The value created is measured in errors avoided, shorter delays and resources released. These companies rarely attract public attention, but their services can become difficult to replace once embedded in an institution’s daily operations.

Tokenisation extends this effort. It represents rights to assets on a programmable ledger, with the aim of coordinating their transfer and payment more directly. The Bank for International Settlements has examined architectures combining central bank money, commercial bank money and tokenised assets. The objective is to modernise the infrastructure through which financial obligations are fulfilled.

The significance of this transformation nevertheless depends on what the token actually represents. A digitised bond remains a claim on an issuer. Its quality depends on the issuer’s ability to pay, the holder’s rights and the possibility of reselling the security. Technology can change the medium and simplify execution; it does not create solvency or liquidity by itself. At this level, legal certainty and institutional trust matter as much as computing performance.

Moving down towards the corporate sector, sophistication takes a different form. The task becomes less about organising a market than knowing where money is, when it will be available and which obligations it must meet. A company may sell in several countries, pay suppliers in another currency and hold balances across different banks. It may also report a profit while running short of cash because customers settle invoices after its own payments fall due.

Treasury fintechs address this mismatch. They consolidate data, automate the reconciliation of invoices and payments, facilitate transfers and improve cash forecasting. Their usefulness does not necessarily depend on a spectacular invention. Making previously fragmented information reliable can be enough to reduce mistakes, administrative costs and reliance on emergency financing.

Visibility must nevertheless be distinguished from the ability to act. Software may show that a subsidiary holds excess cash without having the authority to move it. A payment provider may execute a transfer without extending credit. A financial institution may advance funds or hedge currency exposure. When these services appear on the same screen, their commercial integration does not mean that they sit on the same balance sheet or carry the same responsibilities.

This distinction becomes decisive when fintech finances business activity. Small-business lending, factoring, advances against sales and inventory finance all involve someone making money available today in the expectation of recovering it later. Technology can accelerate underwriting, use invoicing data and make small transactions economical to process. It can also improve monitoring after funds have been disbursed.

But distributing finance and providing the funding are different businesses. A platform may select applications for a bank, connect companies with investors or lend using resources it has raised itself. The first primarily charges for a service; the last must also manage its cost of funding, defaults and losses.

This is where growth indicators become ambiguous. A rapid increase in loan originations may reflect strong distribution, an underserved market or excessively permissive lending standards. Repayment performance is needed to distinguish between these explanations. Volumes, revenues and losses do not always materialise at the same time: a cohort of loans can appear highly profitable before its problems emerge.

Access to more data does not eliminate this timing problem. An algorithm may interpret a merchant’s sales more accurately without anticipating a supply disruption or a collapse in demand. A lender’s quality is revealed by its ability to withstand difficult periods, retain access to funding and recover its claims, as much as by the speed of its approvals.

Investment services occupy a different position. They provide market access, automate portfolio construction and reduce certain administrative or commercial barriers. For customers, the improvement is tangible when investing regularly, understanding fees and monitoring savings become easier. For the operator, revenue may come from subscriptions, commissions or fees linked to assets under administration.

Ease of use must not, however, be confused with investment quality. An interface can make a product accessible without making its risks understandable. It can also encourage excessive activity if the business model rewards frequent transactions. The question is therefore not only how much an order costs, but what behaviour the service as a whole encourages.

Digital insurance presents a comparable separation between distribution and risk-taking. An intermediary may simplify applications and document submission without bearing claims. An insurer that assumes the obligations must assess risks, maintain sufficient resources and arrange reinsurance where appropriate. In both cases, the service can only be fully judged when the customer needs it: when a loss occurs and compensation must be assessed and paid.

Behind these visible businesses lies an industry of components. Some companies provide account-management software; others supply connections to banking data, card issuance, identity verification or suspicious-transaction detection. They sell institutions and platforms the tools needed to offer financial services without building every element themselves.

Banking-as-a-service illustrates this structure. One company may manage the brand and customer relationship while a licensed institution provides certain banking capabilities. The World Bank describes the model as connecting the systems of authorised financial or payment institutions with non-bank businesses. The BIS notes that these partnerships reshape the value chain and alter dependencies between banks and technology companies.

This specialisation lowers barriers to launching new services. It also creates chains of responsibility that users do not always see. The brand answering their messages may differ from the company maintaining the technical ledger and the institution holding their funds. If records diverge or an intermediary stops operating, accurate reconciliation and clear responsibilities become crucial. US banking agencies have issued a joint statement addressing the risks of arrangements between banks and third parties delivering deposit products.

Access to data is another source of competition. Open banking allows certain banking information to be shared with other providers, subject to the applicable framework and customer permission. Open finance extends this approach to areas such as investments, insurance and pensions. The ambition is to enable new services using data previously confined within separate systems.

A durable advantage belongs less to whoever displays the information than to whoever makes it usable, reliable and difficult to replace within a customer’s operations. An invisible supplier embedded in thousands of processes can occupy a stronger position than a well-known application whose users can easily switch.

This logic is particularly evident in payments. The consumer’s gesture is brief; the industrial operation behind it is considerably more involved. A request must be transmitted, authorisation obtained, fraud signals checked, the transaction recorded and settlement to the merchant arranged. Everything must operate with sufficient availability to prevent payment from becoming an obstacle to the sale.

Stripe reports that businesses using its platform generated $1.9 trillion in volume in 2025, an increase of 34%. This measures the activity passing through its infrastructure. It represents neither Stripe’s revenue nor wealth created entirely by the company.

The distinction between flows and revenue is essential. In the first half of 2025, Adyen reported €649 billion in processed volume and approximately €1.09 billion in net revenue. The figures describe different things: the value of transactions on one side, and the revenue recognised by the operator under its business model on the other.

In this industry, small charges applied to large volumes can support a substantial business. Profitability nevertheless depends on infrastructure and partner costs, fraud and service delivery. It also depends on improving outcomes for merchants. Declining fewer legitimate payments while controlling fraud can be worth more than a modest reduction in the advertised fee.

At the everyday account level, fintechs seek a different kind of centrality. Receiving a salary, paying expenses and keeping emergency savings within the same service creates a more durable relationship than occasionally using a card. The commercial objective is therefore often to move from being a secondary account to becoming the primary one.

The word neobank introduces an ambiguity here. It describes a way of presenting and distributing services but does not, by itself, identify the operator’s legal status. Depending on the country and product, the provider may be a bank or rely on other institutions. Understanding the model requires examining who holds the funds, which activities are authorised and what protections apply.

The number of accounts opened provides an imperfect measure of the relationship’s strength. A customer attracted by a promotion may remain inactive. A user who directs income into the account, pays regularly and maintains a balance generates a different pattern of activity. Trust is then built in ordinary situations: a declined card, a delayed transfer, a disputed transaction or an account that needs to be unlocked. Customer service is part of the financial product.

International transfers address a narrower need but reveal the same tension between apparent simplicity and operational complexity. Sending money requires connecting currencies, payment systems and different control procedures. Improvements can come from better liquidity management, local connections and more transparent pricing.

Wise reported an average cross-border take rate of 0.50%, or 50 basis points, in the first quarter of its 2027 financial year. This aggregate ratio does not represent the price of every transfer. It nevertheless illustrates how lower unit charges and growing volumes can interact within a specialised business model.

Stablecoins are also seeking a role in this circulation. They can support certain transfers over programmable networks, but the complete journey often includes buying the token, holding it, converting it and accessing the currency the recipient ultimately needs. A fast transaction between two wallets therefore does not necessarily mean an inexpensive transfer all the way to the final beneficiary. The BIS also highlights the questions of trust and monetary sovereignty associated with their expansion.

As the chain approaches everyday purchases, finance becomes almost invisible. A payment link sent by a tradesperson, a terminal installed in a shop or a QR code presented to a customer makes collection possible. The operator can then add invoicing, sales tracking and management tools. Payments become an entry point into a broader commercial relationship.

Buy now, pay later nevertheless demonstrates the limits of any ranking based on ease of use. For the buyer, a few confirmations are enough. Behind the transaction, someone must fund the advance, pay for those resources and absorb potential defaults. The function remains lending. A merchant may accept a fee to facilitate a sale, but the economics depend on what remains after costs and losses.

At the end of the chain, financial services come down to fundamental needs: depositing cash, receiving money, paying a bill and making a withdrawal. Mobile money addresses these uses by combining digital infrastructure with local distribution. Agents handling deposits and withdrawals must have sufficient liquidity. The quality of the physical network remains decisive.

According to the GSMA, the sector exceeded $2 trillion in transactions in 2025 and reached 2.3 billion registered accounts. These accounts do not represent an equal number of distinct individuals, and not all are active. Their number nevertheless demonstrates the scale achieved by a service whose use can remain very simple.

The World Bank estimates that 79% of adults worldwide held an account with a financial institution or mobile money provider in 2024, compared with 74% in 2021. Published in the Global Findex 2025, these findings measure progress in access; they do not, by themselves, establish that services are affordable, sufficiently used or suited to every need.

For a household or small business, the difference lies in practical details: the distance to a withdrawal point, the fee, the availability of cash and the ability to correct an error. Even the most basic service therefore requires robust infrastructure. Its value lies in reliably solving a recurring problem.

What emerges from this journey is an industry far less homogeneous than its name suggests. Software companies must retain customers and control costs. Payment networks must maintain availability and efficiency at scale. Lenders must survive losses. Distributors must preserve trust and access to the partners that make their products possible. Comparing these companies solely through user numbers, transaction volumes or valuations means confusing different economic models.

Artificial intelligence may alter each of these layers, particularly in document processing, cash-flow analysis and customer support. Its contribution should be assessed through results: time saved, fewer errors, better detection and a service that actually improves. It does not remove the need to identify who makes a decision and who answers for its consequences.

Fintech therefore redistributes financial functions as much as it modernises them. It can reduce the cost of a service, open a market and shift the customer relationship. It can also add an intermediary or transfer a dependency to a less visible supplier. Its importance is measured by what it changes in the fulfilment of financial obligations, rather than by the smoothness of the screen alone.

From the bond market to the corner shop, power belongs to those who make a transaction possible and reliable enough to be repeated. At the top and at the bottom, the technological promise ultimately faces the same test: the funds must be available, the payment must arrive, and someone must answer when it does not.

Principal sources

Bank for International Settlements, The next-generation monetary and financial system, Annual Economic Report 2025; Anchoring trust in money: innovation beyond stablecoins, Annual Economic Report 2026; A two-sided affair: banks and tech firms in banking; executive summary on key considerations for open finance.

World Bank, Global Findex 2025, survey data covering 2024; overview of banking-as-a-service.

Federal Reserve and US banking agencies, Joint Statement on Banks’ Arrangements with Third Parties to Deliver Bank Deposit Products and Services, 2024.

GSMA, State of the Industry Report on Mobile Money 2026, data covering 2025.

Stripe, 2025 annual update; Adyen, first-half 2025 results; Wise, first-quarter financial year 2027 results. Company-reported figures cover different periods and scopes.