Who makes money in Web 2.5? The middle layer rakes in more than financial giants

The fusion of traditional finance and cryptocurrencies has given rise to a new “middle layer,” and Visa, SWIFT, and Chainlink are all competing to control the translation layer of fund movement.
(Background: Traditional finance giants have jumped into web3, and Chainlink held SmartCon 2024 in Hong Kong)
(Additional context: LINK surged by over 20% — Chainlink, together with the U.S. clearinghouse (DTCC) and major banks including JPMorgan, successfully introduced fund data onto the blockchain)

Table of contents

Toggle

• Crypto infrastructure: the rise of Web 2.5
• Middle-layer battle: 50 banks betting on Chainlink
• The value of the translation layer: SWIFT’s sanctions power and the $1.5 trillion market
• Floating return mechanism: Visa’s $5.3 billion intent to acquire Plaid
• Commoditizing blockchains: Web 2.5 replacing decentralized utopias

For most of financial history, moving money has been the hard part. The challenge is how to pay from point A to point B, which requires passing through a series of banks, and each bank takes a commission. Sometimes, it even involves cross-border fund transfers.

Over the past decade, cryptocurrencies and stablecoins promised to reduce these frictions to the maximum extent through crypto applications and wallets. But if those funds can’t be used in a broader economic system, then these fast, low-cost transfers are meaningless. Dollars locked in a single crypto wallet are worth less than their real value. That is why crypto is now playing a more complete infrastructure role—moving existing traditional assets.

The fusion of old and new financial systems has created a new middle layer, and value is being accumulated in this layer. In today’s article, I’ll break down who is capturing value in this new layer.

For more than ten years, the crypto industry has been working to persuade people to download wallets, route assets across different blockchains, and store funds in new applications. But people won’t simply abandon systems they’ve been familiar with and used for decades just to try something new. No provider wants to connect payment flows via a blockchain link and then watch as the money just sits in their wallet, waiting for them to figure out how to convert it back into a bank account that can be used for everyday expenses. Moving funds from a wallet to a bank account requires fees, and in most cases, it also requires compliance checks to be carried out.

The problem has never been that cryptocurrencies can’t transfer money instantly. The issue is that their architecture requires people to give up the systems they already use—such as bank accounts, credit cards, and payroll systems—and instead use an entirely new system. Connection points, exit points, and bridging solutions are all hidden friction points, not features to show off. People always adopt the new technology that can move their existing funds to their existing accounts faster and cheaper.

The ideal infrastructure is one where cryptocurrencies serve as an efficient, invisible helper and the underlying carrier for traditional finance. We call this best state “Web 2.5.” Although this term may sound a bit awkward, the idea behind it is to combine the advantages of both. We keep the essence of traditional finance—regulation, licenses, verification, and the user interfaces and user experiences people have already trusted and are accustomed to. Then we combine that with the low-cost, programmable, always-online settlement approach brought by cryptocurrencies. The two don’t need to replace each other. Banks remain banks, while cryptocurrencies inject new vitality into the outdated, sluggish infrastructure of fund flows.

Crypto infrastructure: the rise of Web 2.5

But if cryptocurrencies become an invisible underlying layer, while traditional finance remains the familiar front-end surface, where does value accumulate in the new world of Web 2.5?

The value of the layer connecting two major financial systems has historically exceeded the value of most of the institutions it connects. Visa’s operating profit last year (fiscal year ending as of September 2025) reached $24 billion, while transaction fees per network transaction account for less than 1%. Even so, its operating margin is still as high as 60%. The Depository Trust & Clearing Corporation (DTCC), which is currently building its own on-chain settlement system, handled securities trades worth $4.7 trillion in 2025 and earned $2.9 billion from them.

Both institutions are now building a conversion layer that allows banks to convert ISO 20022 instructions into on-chain settlement while retaining their infrastructure.

On June 23, Chainlink, together with a consortium of more than 50 European and Korean banks (with total assets of about $10 trillion), announced the launch of the Pangea project to test real-time settlement for foreign exchange transactions.

The goal is to transition FX settlement infrastructure from the traditional T+2 cycle to a real-time T+0 model.

Middle-layer battle: 50 banks betting on Chainlink

Chainlink’s run-time environment (CRE) serves as an orchestration layer, connecting blockchains and other external payment systems without manual routing or bridging. It converts each standard instruction into on-chain atomic swaps and returns the results to the bank system for reading.

Chainlink is a relatively new technology. However, DTCC—an institution with a 50-year history and located at the center of the U.S. market (it handled roughly $4.7 trillion in securities trades last year)—has chosen the same Chainlink run-time to provide support for its collateral application chain.

In terms of traditional institutions, SWIFT is a clear example. Based on early predictions about cryptocurrencies, SWIFT was the institution that blockchain was supposed to replace. Many people predicted stablecoins would bypass this information-transmission monopoly. Eight years ago, this bank-facing global information transmission network described blockchain as “not ready for mainstream adoption.” Yet SWIFT is now building a blockchain-based shared ledger with more than 40 banks.

This is not a replacement for the SWIFT network; it is an orchestration layer built on top of it. The flow of funds on-chain itself has never been the threat. What SWIFT worries about is being excluded from the layer that decides how funds flow on-chain. As long as it can participate and have a voice in the decision-making process, it can keep playing. Therefore, it is building this layer itself.

Even sovereign nations are rushing into this space to capture value. The Bank for International Settlements (BIS) convened seven central banks and more than 40 private institutions to jointly launch the “Agorá project,” aiming to test atomic settlement using tokenized central bank reserves.

The value of the translation layer: SWIFT’s sanctions power and the $1.5 trillion market

But is the real value in building bridges between two financial giants and/or major banking-industry giants?

A translation layer that merely enables both sides to talk is likely more valuable than the players themselves.

Visa and Mastercard originally were routing networks between banks and merchants. Even today, they don’t hold deposits, issue no cards, and assume no risk. However, Visa’s market cap exceeds that of all global banks except JPMorgan.

The value brought by an operating translation layer goes far beyond money. Those who decide where funds flow also get the power to decide when to shut the pipeline.

The SWIFT system was created in 1973, initially just a way for banks to transmit standardized information to one another. Fifty years later, it has acquired powerful authority to impose sanctions on countries. Over the past decade, SWIFT has played a key role in economic warfare, such as imposing sanctions on Russia due to its war in Ukraine. It even imposed EU sanctions on Iranian banks to curb the country’s nuclear program, and eased sanctions after progress was made on the nuclear deal.

Floating return mechanism: Visa’s $5.3 billion intent to acquire Plaid

Chainlink’s current pilot work with Project Pangea on permissioned addressable fund pools for real-time FX transaction settlement is significant.

Cross-border payments range annually from $150 trillion to $190 trillion and are expected to exceed $250 trillion by 2030. If Chainlink and its alliance of 50 partner banks can secure even 1% of that share, its potential market size (TAM) would exceed $1.5 trillion. Even charging only 0.1% in fees, Chainlink could generate $1.5 billion in revenue by building a bridge between traditional finance and on-chain settlement.

But there’s a caveat here. Both SWIFT and Visa became dominant standards in their respective fields, so the entire system ultimately has to adopt them. Each domain has only one winner, and over decades it has continuously consolidated its position.

Now we have four different models—protocols, market utilities, bank cooperatives, and central bank clubs—all competing for the same single translation layer to connect the financial worlds of Web 2.0 and Web 3.0.

The economic mechanism driving the value of this layer has long existed. As fund-flow technologies advanced, transaction processing itself gradually became a commodity. As fund-flow costs fell, the accessible value mainly concentrated in two areas. First is authorization: the power of those who decide whether a transaction is possible and under what conditions. Second is floating returns: the interest generated while funds sit idle waiting to be transferred.

We previously wrote about how payments work when carried out between AI agents (see here and here). Now, the same logic also applies to interbank settlement.

Commoditizing blockchains: Web 2.5 replacing decentralized utopias

This is exactly why the middle coordination layer is worth fighting for. It creates a two-way network effect. The more banks connect on one side, the more attractive it becomes for the settlement entity on the other side, and vice versa. With each additional institution, the cost for existing institutions to exit increases. Although competition exists between banks and between blockchains, the institutions that coordinate the conversion layer can provide services to all banks and blockchains—and charge fees for it.

Stripe adopted the same strategy in the credit card payments space. It enables businesses of all sizes to easily accept and manage online payments through a simple, developer-friendly API, thereby hiding the complexity of payment processors, acquiring institutions, and payment networks. Then it charges all users a fee, eliminating transaction friction and burying it in the backend.

That is why the connecting layer becomes a hot target for acquisition. Once someone builds this connecting layer, others would rather buy it directly than rebuild it from scratch. Five years ago, Visa agreed to acquire Plaid for $5.3 billion, and we saw exactly that. Although the deal eventually fell through due to the U.S. Department of Justice’s antitrust lawsuit, the intent behind it was clear. Visa attempted to acquire the market share of Plaid’s operating connecting layer, responsible for connecting thousands of fintech applications to bank accounts.

The Web 2.5 world is more promising than a fully decentralized Web 3.0 utopia vision, because it doesn’t require capital to flee existing participants in search of the services that crypto enables. Instead, it treats crypto as a more efficient underlying infrastructure for moving funds and assets within existing ecosystems.

Although bank-side projects including Pangea, DTCC’s AppChain, and Agorá are still in pre-production stages, we like the push direction of participants such as Chainlink. For a long time, internal debates in the crypto space have been ongoing, with people arguing about how to build better crypto applications to get users to abandon traditional payment methods. Developers have also been debating which blockchain has the lowest gas fees and which token is best for storing funds. Web 2.5 makes these arguments unnecessary by removing all the jargon and hiding the infrastructure in the backend.

The internet we use is really just information data packets transmitted over a global computer network. That’s certainly useful knowledge, but if you just want to get online, there’s nothing to brag about. No one cares whether the technology behind those lightning-fast, low-cost transactions is cryptocurrency or something else.

Blockchains are increasingly being commoditized, becoming an interchangeable, intangible, and thinly profitable component in transactions. Its value today is reflected in business models of fund flow—and gives people a say in how funds move and whether they move at all.