This is counterintuitive. Let me be exact about what I'm claiming and what I'm not.
Not that banks are more innovative. Not that decentralized tokenization is irrelevant. Rather: the structural preconditions for tokenization at scale — regulatory licenses, compliance infrastructure, custody capability, institutional relationships — already exist for banks. Startups spend years and millions building them from zero. For banks, they're existing overhead. For financial instruments specifically, the incumbents' advantage is permission. Not technology.
I've worked through over a hundred tokenization concepts. The failure pattern is consistent: regulatory stage. Teams hit a wall trying to satisfy the legal preconditions that make a token legally enforceable. Banks and exchanges already satisfy those preconditions. Licensed to issue securities. Licensed to hold deposits, custody assets, serve qualified investors. Tokenization for them is a technology upgrade. Not a new permission question.
What Banks Have Already Built
Production, not concept. Not pilots announced last year that faded. Real flows, real clients, real capital movement — at varying stages of maturity.
JPMorgan: Kinexys processes roughly $2B daily in tokenized deposits between JPMorgan clients, on blockchain infrastructure JPMorgan operates. Separately, in November 2025, JPMorgan extended JPM Coin to Coinbase's Base network for a narrow set of institutional clients — B2C2, Coinbase, Mastercard among the first. That's a scoped product extension, not a migration of the Kinexys platform itself. In January 2026, JPMorgan announced Canton Network deployment. The trajectory points toward expanding multi-chain institutional presence.
BNY Mellon launched Digital Cash on January 9, 2026 — demand deposit claims, on-chain, against a private permissioned blockchain. Six initial participants: ICE, Citadel Securities, DRW Holdings, Ripple Prime, Baillie Gifford, Circle. The participant list signals institutional commitment, though production volume data is not yet publicly available.
Custodia Bank completed the Avit pilot in March 2025. Eight test transactions with Vantage Bank on Ethereum mainnet. The experiment: minting, transferring, and redeeming ERC-20 stablecoins inside a regulated banking environment. 1:1 backed by deposits. The result demonstrated that tokenized money can operate within banking compliance frameworks. Not a commercial launch. Not yet at scale. But it showed the path exists for smaller banks that build via infrastructure providers.
State Street announced Digital Asset Platform on January 15, 2026. Covers tokenized money market funds, ETFs, deposits, and stablecoins. The platform is under development with 2 or 3 initial clients reported. This is the announcement phase. Production at scale hasn't arrived.
Citi Token Services integrates with Citi's 24/7 USD Clearing infrastructure. The tokenized deposit component operates within Citi's own system; the clearing network — which reaches 250+ banks across 40 jurisdictions — settles outward on behalf of counterparties. This is internal tokenization powering external settlement, not interbank token movement.
The UK Regulated Liability Network ran two phases. Experimentation (2024): 11 entities including Barclays, Citi UK, HSBC UK, Lloyds, Mastercard, NatWest, Nationwide, Santander UK, Standard Chartered, Virgin Money, Visa. It concluded successfully. Then came the live pilot phase — tokenized sterling deposits, active through mid-2026. The pilot narrowed to 7 banks: Barclays, HSBC, Lloyds, Monzo, NatWest, Nationwide, Santander. The concept is validated. Scaling beyond pilot conditions is the open question.
What's absent from most of these: new licenses. In most cases, banks tokenized using permissions they already held. A tokenized deposit is a deposit under banking law. A tokenized fund is a fund under securities law. KYC, AML, sanctions screening, regulatory reporting — the entire compliance apparatus that applies to traditional instruments applies identically to tokenized versions. Custodia Bank is the exception, operating under a Wyoming SPDI charter designed explicitly for this purpose.
Why Startups Face a Structural Disadvantage
A startup trying to issue a tokenized security runs a different gauntlet entirely. Regulatory license or exemption in every jurisdiction where it operates. Investor qualification infrastructure. Custody that regulators will accept. Legal counsel in each relevant jurisdiction. Capital to build all of this before the first token ever moves.
I've watched this pattern across over a hundred projects. Roughly 90% fail at the regulatory stage. Not engineering. Not business model. Regulatory stage. Teams can't satisfy the conditions that convert a token into a legally enforceable financial instrument. No license means no enforceability.
Banks don't face this problem because the preconditions are already met. When a bank tokenizes a deposit, it's exercising a permission it already holds. When a startup tries to tokenize an investment product, it's asking the regulator for a new permission — in a framework that may not grant it.
This isn't a statement about startup innovation or capability. For financial instruments — deposits, securities, funds, structured products — the licensed institution moves faster. The gap isn't technical. It's regulatory permissions, compliance infrastructure, and institutional trust.
HKEX Synapse: The Model That Explains the Pattern
HKEX Synapse shows exactly why this matters. The Hong Kong Exchange's integrated post-trade platform connects Hong Kong and mainland equity markets through the Stock Connect program. It uses DAML smart contracts to automate settlement.
Not in the crypto sense — no ERC-20 tokens, no DeFi, no permissionless access. But functionally: financial instruments as programmable digital objects. Automated settlement. Distributed infrastructure.
HKEX didn't seek new regulatory permission. They implemented a technology upgrade on infrastructure they already operated. The regulatory conversation was about implementation standards. Not foundational legality.
This pattern repeats across institutional tokenization. Banks view it as a technology upgrade to settlement, custody, clearing, and distribution. Regulators ask implementation questions. Startups building from scratch encounter the foundational question first: do you have permission to do this at all?
The "Cash Island" Problem
Institutional tokenization leaves one major problem unsolved. And it's significant.
Most tokenized deposit systems operate within a single institution. JPMorgan tokens move between JPMorgan clients. BNY between BNY clients. Citi between Citi counterparties. As of this writing, the US has no production-grade interbank tokenized deposit system. Each bank mints its own on-chain money. No bridges exist. A JPMorgan tokenized deposit has no fungibility with a BNY one — even though both are US dollar demand deposits at regulated banks.
This is the cash island problem. Institutional tokenization fragments liquidity into pools that don't connect.
The UK RLN pilot comes closest to a working model for multi-bank coordination. SWIFT has been integrating Chainlink CCIP as a cross-chain connectivity layer — an effort that moved into pilot stages in late 2025. The direction is clear: connecting SWIFT messaging to blockchain infrastructure so that existing bank networks can interact with on-chain settlement. But regulated interbank settlement over blockchain remains pre-production. Standards are forming. Production-scale deployment hasn't arrived.
Until then, bank-issued tokenized money reproduces traditional banking fragmentation — just on different infrastructure. Settlement speed and programmability improve within each island. The unified global liquidity the RWA narrative promises remains ahead.
What This Means for Startups
The evidence points one direction: banks will tokenize financial instruments first. Which means the strategic calculus for startups changes.
Head-to-head competition on issuance is structurally disadvantaged. A startup issuing tokenized securities against banks issuing tokenized deposits under existing banking licenses is competing on terrain the bank already controls.
Where startups win is infrastructure. Banks need the technology to move tokenization forward. Custody solutions. Smart contract platforms. Compliance automation. Token issuance infrastructure. Secondary market execution. Reconciliation and settlement. This is where startup technical depth meets the bank's regulatory standing. White-label tokenization platforms work because they handle security token issuance, investor onboarding, cap table management, distribution logistics, and secondary markets — all deployed under the bank's brand, powered by startup engineering.
Where the longer-term value lives: DeFi integration. Banks can tokenize their existing operations. What they cannot do — regulatory constraint, cultural constraint, both — is plug those tokenized instruments into permissionless DeFi. AMMs, lending protocols, yield vaults, stablecoin collateral systems. That financial layer institutional infrastructure cannot replicate on its own. The liquidity layer thesis explores this in more depth.
The adoption sequence flows like this. Banks tokenize first using existing licenses and compliance infrastructure. They solve issuance. They solve custody. They solve regulation — within their own walls. They create cash islands, not open markets.
Startups and DeFi solve the liquidity problem later. Secondary markets. Cross-institution interoperability. Composability with broader infrastructure. This happens later because the regulatory framework needs to catch up. When it does, value flows to whoever built the connective tissue between islands.
The GENIUS Act Divide
One regulatory development illustrates the pattern.
The GENIUS Act (July 2025) created a federal stablecoin framework. The key provision: no interest allowed on stablecoins. These are payment instruments, not deposits. Full 1:1 reserve backing required.
Tokenized deposits occupy a different regulatory category. They're deposits under banking law. They can earn yield. Standard fractional reserve rules apply. Where applicable, existing deposit insurance frameworks — including FDIC coverage — extend to them, subject to the same conditions as any other bank deposit.
The distinction matters. Non-bank stablecoin issuers (Circle, Tether, others) hold strict reserves, pay zero interest. Banks hold fractional reserves, offer yield, operate under established deposit insurance regimes.
For institutional users the choice clarifies itself. Tokenized deposit at a licensed bank: potential deposit insurance, interest income, a regulatory framework compliance teams already understand. Stablecoin: no insurance, no interest, a newer framework still being interpreted.
Stablecoins aren't obsolete. Cross-border payments work on stablecoins. DeFi integration requires stablecoins. Emerging market infrastructure relies on them where traditional banking doesn't reach.
But for institutional treasury management, interbank settlement, corporate cash reserves — the tokenized deposit has a structural advantage. Only banks can issue it.
The Sequence
Banks tokenize first. Not faster innovation. Preconditions already satisfied.
Deposits get tokenized for internal settlement. Securities for operational efficiency. Money market funds for faster distribution. Existing licenses. Existing compliance. Existing clients. The technology upgrade layer is tractable.
What happens next is harder. The cash island problem. Someone needs to solve that — through standards like the UK RLN model, through cross-chain infrastructure connecting traditional banking to on-chain settlement, through regulatory mandate. The DeFi integration question compounds it. How do tokenized institutional instruments interact with permissionless financial systems? That framework doesn't exist yet.
Startups that position as issuers will find the market occupied. But infrastructure providers — custody, compliance, issuance automation, interoperability technology that banks actually deploy — will find demand accelerating.
The real opportunity is narrower and higher-value: founders, DeFi protocols, infrastructure teams that bridge institutional cash islands to open DeFi liquidity. That connective layer doesn't exist yet. The islands are being built right now. The demand to bridge between them is forming.
Key Takeaways
- Banks will tokenize financial instruments at scale before startups do — because they already hold the regulatory licenses, compliance infrastructure, and custody capabilities that most startup projects spend years trying to build
- JPMorgan Kinexys ($2B/day in tokenized deposits), BNY Digital Cash (6 institutional participants at launch), Custodia Avit (pilot demonstrating feasibility), State Street (platform under development), UK RLN (7-bank live pilot), and Citi Token Services (internal tokenization powering external clearing) demonstrate various stages of institutional tokenization — from production flows to proofs-of-concept
- The "cash island" problem limits institutional tokenization: each bank's tokens work within its own client base, and no production-grade interbank system exists yet
- The GENIUS Act creates a regulatory distinction: stablecoins cannot pay interest; tokenized deposits can — giving banks a structural advantage for institutional use
- Startups should position as infrastructure providers (custody, compliance, issuance platforms) rather than competing as issuers
- The highest-value opportunity is connecting institutional cash islands to open DeFi liquidity — a layer that doesn't exist yet but will define the next phase