The collapse of Silvergate Bank in March 2023 did not merely remove a lender from the market. It exposed the structural dependency that every major stablecoin issuer had quietly accepted as operational cost. Within seventy-two hours of the FDIC takeover announcement, Circle publicly disclosed $3.3 billion in deposits tied up at Silvergate. Tether moved quickly to diversify, but the episode revealed something the marketing decks never contained: the entire stablecoin ecosystem — all $180 billion of it — operates on a handful of rubber-stamped IOUs from institutions that can be closed overnight by a regulator's pen.
That is the uncomfortable reality the "stablecoins need banks" thesis forces into the open. The argument is not subtle. It asserts that no matter how sophisticated the smart contract layer becomes, no matter how elegant the on-chain settlement logic, stablecoin scaling cannot proceed without institutional-grade banking infrastructure anchoring the rails. Institutions will not trust reserves held in crypto-native custody. They will not settle bilateral positions through decentralized bridges. They require the familiar architecture of bank accounts, federally insured deposits, audited balance sheets, and regulators with phone numbers.
The thesis has gained traction as institutional interest in stablecoins has accelerated. PayPal's PYUSD, JPMorgan's JPM Coin, and a growing roster of tokenized deposit products from traditional lenders have moved from whitepaper to pilot. But the infrastructure connecting these pilots to production-scale operation remains thin. The bottleneck, as the argument correctly identifies, is not technology. It is the absence of a regulated settlement layer that institutions can build internal controls around.
Understanding why requires stepping back from the on-chain narrative and looking at the balance sheet mechanics that actually govern this market.
The Balance Sheet Problem Nobody Talks About
Stablecoins are not cryptocurrency. They are short-term liability instruments issued against reserve assets. USDC holds Treasuries and cash. USDT holds a mix of commercial paper, Treasuries, and cash equivalents. The moment an institution receives a stablecoin, it holds a claim against the issuer's reserves — not against a blockchain. The value of that claim depends entirely on the credibility of the reserve audit and the solvency of the custodian holding those assets.
For retail participants, this model has worked adequately. Price stability has held. Redemption queues, when they appear, have cleared. But for institutional counterparties — the treasurers, the payment processors, the hedge funds running cross-border margin — the model introduces a problem that blockchain cannot solve: counterparty risk lives outside the chain.
When a corporate treasurer at a Fortune 500 company considers using stablecoins for cross-border supplier payments, the decision tree does not begin with gas fees or block confirmation times. It begins with a question that sounds mundane but carries enormous weight: who holds the reserves, and can I verify their existence without trusting the issuer's self-reported numbers? The answer, under the current architecture, is no. Reserve attestations are point-in-time snapshots produced by accounting firms selected by the issuer. They are not real-time. They are not on-chain. They do not satisfy the internal audit requirements of a public company's treasury function.
Bank custody solves this at the infrastructure level. When a regulated bank holds reserves, the institution inherits the bank's existing audit framework, its Federal Deposit Insurance Corporation backstop, its OCC supervision, and its correspondent banking relationships. The stablecoin becomes a digital representation of a bank deposit rather than a claim on a corporate balance sheet operating outside the banking system. That distinction matters enormously to the institutional buyer who needs to justify the decision to a compliance committee.
The implications for the current competitive landscape are severe. Tether's dominance — roughly 70% of the stablecoin market by supply — rests on a reserve structure that relies heavily on offshore banking relationships and non-traditional custodians. Circle's USDC has pursued a different path, with reserves held at BlackRock's Treasury fund and custodied through regulated banks. The 2023 banking disruptions pushed Circle toward increasingly conservative reserve management, but the underlying structural advantage of the bank-anchored model was made visible rather than theoretical.
The market knows this. Post-Silvergate, Circle executed banking agreements with multiple regulated institutions, and its public communications increasingly emphasized the distinction between "bank-grade" reserve custody and the self-custodied alternatives. This is not coincidence. It is competitive positioning that reflects the real-world requirements of the institutional customers both companies are courting.
The Regulatory Architecture Is Converging on This Conclusion
The legislative trajectory in the United States has moved decisively toward requiring bank-level oversight for stablecoin issuance. The GENIUS Act, which advanced through Senate consideration in 2025, proposes a federal licensing framework that would restrict stablecoin issuance to federally or state-chartered banks and non-bank entities meeting strict capital and reserve requirements. The Lummis-Gillibrand Payment Stablecoin Act, which preceded it, contained similar provisions. The pattern is consistent: regulators want an identifiable, licensable entity on the other side of the stablecoin liability.
The European Union's Markets in Crypto-Assets regulation has already operationalized this requirement. MiCA's electronic money token framework requires stablecoin issuers operating in EU markets to obtain an e-money license, maintain 1:1 reserve backing in liquid assets, and comply with anti-money laundering obligations equivalent to those applied to payment institutions. The infrastructure is not theoretical. It is live. As of mid-2025, several European stablecoin issuers have obtained e-money licenses and are operating within the framework.
This regulatory convergence creates a structural moat for bank-anchored stablecoins and represents an existential compliance challenge for issuers operating without banking licenses or their functional equivalent. The argument that "without banks, stablecoins cannot scale" is increasingly being rewritten as "without banking licenses, stablecoins cannot operate legally at scale in regulated markets."
The forensic reality is this: the 2023 banking crisis did not merely inconvenience stablecoin issuers. It revealed that the upstream infrastructure — the regulated banking relationships that make fiat-to-stablecoin on-ramps and stablecoin-to-fiat off-ramps functional — is the actual constraint on scaling. Supply can be minted on-chain in seconds. Redemption requires a bank account on the other end. That asymmetry is the true bottleneck, and it has nothing to do with consensus algorithms or rollup throughput.
The Bull Case the Banking Thesis Gets Wrong
The argument, however, contains a structural blind spot that its proponents rarely acknowledge: it assumes institutional trust requires institutional banking architecture, and that assumption is being challenged by the emerging infrastructure of on-chain compliance technology.
The counter-argument is not that decentralized stablecoins will replace bank-anchored ones. It is that the compliance and audit requirements driving institutions toward banks can, in principle, be satisfied by non-bank infrastructure that achieves the same verification outcomes through different mechanisms. Real-time on-chain reserve proofs, implemented by entities like Chainlink and several academic research groups, can provide continuous, cryptographic verification of reserve backing without requiring a chartered bank as custodian. Regulated custodians that are not banks — state-licensed trust companies, for example — can perform the same reserve oversight function under different supervisory frameworks.
The MiCA framework itself demonstrates this possibility. The regulation does not mandate that only banks issue stablecoins. It mandates that issuers obtain an e-money license, which is accessible to non-bank payment institutions meeting capital and compliance requirements. Several non-bank issuers have navigated this process successfully. The "bank is mandatory" narrative therefore overstates the regulatory reality in at least one major jurisdiction.
Furthermore, the banking thesis is constructed primarily from the perspective of developed-market institutions operating within established regulatory frameworks. The use case driving the most explosive stablecoin growth — informal remittance corridors in Southeast Asia, Latin America, and Sub-Saharan Africa — operates on a fundamentally different infrastructure model. Users in these markets do not need FDIC-insured reserves. They need stable value transfer without banking intermediation, often because they are underbanked or unbanked by design. The banking-centric thesis cannot account for this segment without contradicting itself: if banks are the prerequisite, the unbanked are excluded by definition.
The practical implication is that the stablecoin market may bifurcate rather than consolidate around a single infrastructure model. Institutional-facing stablecoins — used for treasury management, securities settlement, and B2B payments — will likely converge on bank-anchored, regulated structures. Consumer-facing and remittance-focused stablecoins serving underbanked populations may continue operating through non-bank channels, whether or not that satisfies the requirements of the GENIUS Act's drafters.
What the Market Has Already Priced In
The competitive dynamics are shifting in ways that the banking thesis implies but does not fully articulate. Circle's trajectory — from a crypto-native startup to a company with explicit banking partnerships, regulated reserve structures, and public statements about institutional adoption — reflects the market's implicit verdict on the infrastructure question. Tether's continued dominance despite its non-bank structure suggests that network effects and first-mover brand recognition retain significant value, but the compliance pressure from regulators in the EU, UK, and increasingly the US is systematically increasing the cost of operating outside the regulated banking perimeter.
The entities positioned to benefit most from this shift are not necessarily the stablecoin issuers themselves. The compliance infrastructure surrounding regulated stablecoins — real-time reserve monitoring tools, chainalysis and compliance analytics platforms, KYC/AML infrastructure designed for on-chain transactions, and the banking-as-a-service providers that enable non-bank companies to access regulated rails — represents a growth vector that is less visible but potentially more durable than the stablecoin supply numbers themselves.
This is the part of the thesis that deserves more attention than it receives. When a regulated bank becomes the custodian of stablecoin reserves, it does not merely provide a service to the issuer. It creates a new product category: programmable digital deposits that carry bank credit risk rather than corporate credit risk. The revenue model shifts from issuer interest income on reserves to bank fee income from custody, issuance facilitation, and transaction settlement. The stablecoin does not replace the bank. The bank absorbs the stablecoin's margin structure and adds its own on top.
The blockchain remembers every transaction. But the auditors forget that the reserve audit is always one quarter behind the current balance sheet. Banking infrastructure does not solve this problem perfectly. It relocates it to an institution with a stronger balance sheet, better regulatory standing, and a government backstop. For institutions managing nine-figure treasury positions, that relocation is not optional. It is the entire ballgame.
The stablecoin market will continue to grow. The next $100 billion in supply, however, will be denominated in instruments that look substantially more like regulated bank deposits than the crypto-native liability structures that defined the first decade of this market. Whether that transition happens through explicit bank issuance, bank-adjacent licensing arrangements, or non-bank entities operating under e-money equivalents is a tactical question. The strategic conclusion is not in dispute: the infrastructure layer is where the real competition will be fought, and that layer runs through institutions with balance sheets, charters, and regulators sitting across the table from them.
