Large technological substitutions are rarely announced. They accumulate quietly in the data while the incumbent industry continues to report growth, conduct conferences, and publish white papers about digital transformation. By the time the substitution is visible enough to be undeniable, it has usually been underway for a decade. The history of print advertising, travel agencies, and physical retail all follow this pattern. The question worth asking about banking is not whether the pattern applies — it clearly does — but how far along the substitution curve we currently are.
The honest answer, based on Q1 2026 data, is further than most people inside the banking industry are prepared to acknowledge, and not as far as the more enthusiastic crypto analysts would have you believe. The migration is real, the numbers are large, and the mechanism is now understood. The timeline is genuinely uncertain.
The numbers that matter
These are not projections. They are Q1 2026 figures drawn from Bessemer Venture Partners' stablecoin infrastructure report, Visa's own investor disclosures, and the Stablecoin Insider quarterly report. The stablecoin supply crossed $273 billion in March 2026 — a 40-fold increase from $6.8 billion in March 2020. That six-year trajectory is steeper than the growth curve of mobile payments in its equivalent period.
What matters about these numbers is not their size but their composition. The common dismissal of stablecoin volume data is that it is dominated by arbitrage bots, exchange settlement, and DeFi protocol activity — activity that looks like payments but is not. This dismissal is partially correct, and the more careful analysts distinguish between gross transaction volume (which includes automated activity) and adjusted volume (which attempts to strip it out). The adjusted figure is the meaningful one, and even the adjusted figure is now approaching Visa scale.
More telling than the aggregate volume is the real-world payments subset. Real-world stablecoin payments volume — merchant acceptance, remittances, payroll, B2B invoicing — doubled in 2025 to $400 billion on an annualised basis. The growth in this category is coming from a low base, which makes it easy to dismiss as percentage-driven noise. The absolute number is not noise. $400 billion in real payments routed through stablecoin rails in a single year is a banking product with a meaningful market share.
Why this time is structurally different
Crypto has promised to disrupt banking several times before, and the promises have consistently exceeded the delivery. The 2017 ICO wave produced no payment infrastructure worth using. The 2021 DeFi summer produced extraordinary yield but almost no real-world spending. Each cycle attracted capital and attention, and each cycle receded without having materially changed how most people manage their money.
The current moment is structurally different for three reasons that were not present in previous cycles.
The first is the maturity of the stablecoin layer. Previous cycles lacked a credible, liquid, widely-accepted unit of account for on-chain spending. Spending in volatile assets (ETH, BTC) creates tax events and psychological friction. Spending in stablecoins does not — and USDC and USDT now have sufficient liquidity, regulatory clarity (in progress), and institutional acceptance to function as genuine payment instruments rather than trading intermediaries.
The second is the emergence of fiat-to-on-chain IBANs. Monerium, Bridge, and a small number of EU-licensed institutions now provide IBAN accounts that map directly to on-chain stablecoin addresses. This closes the loop between the traditional banking system and the on-chain stack in a way that previous infrastructure did not. You can now receive a salary in fiat, have it arrive as USDC on-chain within minutes, and spend it via a Visa card without ever touching a bank account in the traditional sense. The translation step has been removed.
The third is regulatory momentum. The EU's MiCA framework, the UK's crypto asset regime, and the US GENIUS Act (pending at the time of writing) are all moving toward frameworks that legitimise stablecoin issuance and on-chain payment infrastructure. This is significant not because regulation accelerates adoption directly, but because it removes the institutional hesitation that has kept large-scale employer payroll, B2B payments, and merchant acceptance off the table. Regulatory clarity is the unlock for the $400 billion real-payments figure to grow by an order of magnitude.
The substitution mechanism
The pattern of the migration is not the wholesale replacement of banks. It is the systematic unbundling of banking services, with each service being replaced by a superior on-chain alternative as the infrastructure for that specific service matures.
Savings accounts were the first service to be meaningfully substituted. A stablecoin position in a curated DeFi vault has been returning 5–7% annually for the better part of three years. A European savings account returns 2–3% on the best available product. The substitution here is not theoretical — it is the primary reason stablecoin supply growth has been so consistent across cycle conditions. People park money in on-chain yield products the way previous generations parked it in money market funds.
Cross-border payments are the second service to be substantially substituted. SWIFT payments between non-correspondent banking relationships take two to five business days and cost 2–5% in fees. USDC transfers settle in seconds and cost a fraction of a cent. The $400 billion in real-world stablecoin payments is disproportionately cross-border remittance, because that is where the substitution benefit is largest. This is also where the 80% of stablecoin activity occurring outside the United States comes from — the populations with the most to gain from escaping costly cross-border banking infrastructure are leading the adoption.
Spending — the service this publication covers — is the third. It is also the newest, and the least mature of the three. But the infrastructure now exists, the cards work, and the 460% year-on-year growth in Visa's stablecoin-linked card spend is the early signal of a category entering its growth phase. The pattern matches the early data from mobile payments, which also looked like a curiosity until it did not.
What banking retains
It is worth being precise about what the on-chain stack does not currently replace, because the substitution is partial and the parts that remain are not trivial.
Credit underwriting based on traditional income and credit history is not substituted. On-chain lending requires overcollateralisation — you need to already hold the asset you are borrowing against. This excludes the largest and most profitable segment of consumer banking: unsecured credit to people who do not already have significant assets. The credit card business, mortgage lending, and personal loans depend on a bank's willingness to extend credit based on future earning potential. No on-chain protocol currently prices that risk or has the legal standing to enforce repayment.
Deposit insurance is not substituted. Smart contract risk is real. Protocol risk is real. The FDIC-equivalent protection that makes traditional deposits genuinely safe for ordinary people — protection from institutional failure by a solvent government backstop — does not exist on-chain. This matters most for people with modest savings who cannot absorb the loss of a DeFi protocol failure. The on-chain stack is currently optimised for people with enough capital to absorb smart contract risk as a portfolio cost. Extending it safely to people who cannot bear that risk remains unsolved.
Regulatory compliance infrastructure is not substituted. Businesses that need to operate within existing financial regulations — payroll tax, VAT remittance, AML/KYC compliance — still route through traditional banks because the compliance layer sits in traditional banking infrastructure. This is changing, but slowly, and the regulated stablecoin frameworks coming online in 2026 and 2027 will take years to build the same depth of compliance tooling that traditional banking has accumulated over decades.
The timeline question
The honest answer is that nobody knows. Technology substitution timelines are among the least predictable things in economics, because they depend on regulatory decisions, incumbent response, and the accumulation of consumer trust — none of which follows a predictable schedule.
What the data supports is a narrower claim: the substitution of the savings, cross-border payments, and consumer spending layers of banking by on-chain infrastructure is already occurring, is accelerating, and is unlikely to reverse. The $10.9 trillion in adjusted stablecoin volume, the $400 billion in real-world payments, and the $18 billion in crypto card spending are not speculative projections. They are 2025 and early 2026 actuals.
The question of when on-chain infrastructure becomes the primary layer for most consumer financial activity — rather than an alternative layer for a sophisticated minority — is genuinely open. Five years is probably too aggressive. Thirty years is probably too conservative. The current trajectory suggests something in between, with the pace depending heavily on whether the regulated stablecoin frameworks currently in progress in the US and EU produce workable infrastructure or produce the kind of compliance overhead that slows adoption to institutional timescales.
For the readers of this publication — people already operating within the on-chain stack — the macro trajectory is less important than the present-day question of what works. The stack works now. The migration is already available to you, if you want it. The rest of the world will catch up on its own schedule.
A note on what this publication is for
This is the only piece in the Off the Rails archive without affiliate links. That is intentional. The macro argument above stands or falls on its own — it does not need to be selling anything to be true, and mixing a commercial interest into a thesis piece undermines both.
The practical guides in this publication — the card reviews, the comparisons, the how-tos — exist because the infrastructure described in this piece is real and usable and most people do not know how to set it up. The macro piece exists to explain why any of that matters. They are different kinds of writing for different purposes, and this publication will continue to publish both.
If this piece was useful and you want the practical guides in your inbox on Sundays, the newsletter link is below.