The arbitrage exists because the US GENIUS Act explicitly prohibits issuers of qualified payment stablecoins from paying interest to holders, while placing no such restriction on tokenised deposits. Tokenised deposits are bank liabilities under existing banking law and the Federal Reserve retains prudential jurisdiction; banks can pay the prevailing deposit rate on tokenised deposits as a routine matter. At ~4.5% Fed funds (April 2026), the structural yield differential is roughly 450 basis points for institutional treasury use cases where yield matters and DeFi composability does not. For a corporate treasurer holding $1B in operational cash, the arithmetic is $45M per year. The window is open from H2 2026 through approximately H2 2027, with material economics in the July 2026 to December 2027 zone, before custodian reward programmes (permitted under GENIUS for stablecoin custodians to distribute reserve yield to end users) operationalise and Fed rate-cut paths shrink the absolute spread. For an institutional tokenisation operator routing US-dollar settlement, the question is which bank captures the second-mover position behind JPMorgan and Kinexys and on which chain.
Why the window is wider than it looks
At first glance, 18 months of a 4-percent yield advantage is a marginal arbitrage. Three reasons it is not.
Second-mover institutional settlement flow is sticky. The bank that operationalises tokenised deposits second (after JPMorgan) captures corporate-treasury and inter-bank settlement flow that does not want to run on JPMD specifically, for competitive concentration reasons or relationship preferences. Once those flows are wired they are hard to dislodge. Whoever lands the second-bank slot among State Street, BNY, Citi, and the European peers (Deutsche Bank, BNP Paribas, HSBC) captures a 5-10 year positioning advantage rather than an 18-month one.
The window is also the opportunity to build public-chain infrastructure at lower competitive intensity. Once custodian reward programmes commoditise the headline-yield advantage, the moat reverts to distribution: which chain the deposit token sits on, which wallets can access it, which DeFi protocols integrate natively. JPMorgan's choice to deploy JPMD on Base (Coinbase's L2) is the leading indicator for why public-chain deployment during the window is structurally correct rather than incidentally so. A second-bank entrant choosing Canton or a private Kinexys-style chain instead of a public L2 cannot recover the public-chain distribution moat later without re-issuance.
Tokenised deposits as the operational settlement asset unlock Wave-2 RWA adoption. The settlement-asset problem has been the rate-limiter on the broader institutional RWA thesis: with no neutral, interest-bearing, banking-licensed settlement asset, allocators have stayed in account-based clearing. An 18-month window where deposits become the default institutional-settlement rail has disproportionate second-order effects on tokenised equities (DTCC's H2 2026 launch trajectory), tokenised private credit (Apollo's ACRED, Centrifuge's Janus Henderson product), and institutional DeFi (Aave V4 RWA Spokes, Anchorage-fronted Morpho institutional vaults).
The candidate set for the second-bank slot
The named candidates and the reasoning behind them (probabilities are operator estimates, not market consensus):
- State Street. Digital-asset division actively publishing through 2025-2026; global custodian scale; SAB 121 rescission was directly unblocking; but slower internally than JPMorgan.
- BNY Mellon. Bitcoin ETF custody live; institutional gold standard for custody brand; conservative on public-chain deployment so far.
- Citi. Has the global treasury-services footprint; participates in Canton Network; internally fragmented and historically slower to ship.
- DBS. Already has DDEx and a USDG reserve manager role through Paxos Digital Singapore; Singapore jurisdiction is favourable; but scale is APAC-only and deposit-token distribution outside Singapore would require additional architecture.
- HSBC and Standard Chartered. HSBC has Orion as the tokenisation platform; Standard Chartered has Zodia Custody as a vehicle and has co-led Partior with DBS and JPMorgan; neither has a live tokenised-deposit product comparable to JPMD.
- BNP Paribas and Deutsche Bank. Public Ethereum deployment in motion (BNP through Chainlink integrations); MiCA jurisdiction clarity; but the EU wholesale CBDC narrative competes for institutional mindshare.
The probability that one of these six lands within the window is high. The probability that the entrant chooses a public L2 over a private or permissioned chain is closer to 50-50; JPMorgan's Base choice was contrarian and competitive imitation could go either way.
Why stablecoin issuers cannot directly counter
The structural asymmetry that gets less attention than it should: stablecoin issuers have no direct counter-move during the window. Circle's path is the custodian reward programme, which takes 12-18 months to operationalise and benefits the custodian distribution partners more than Circle itself. Paxos's USDG partner-economics model (the Global Dollar Network distributing 90%+ of reserve yield back to participating partners) is competitive on distribution but cannot close the headline-yield gap. Tether is structurally unable to compete for GENIUS-compliant institutional flow at all because it sits outside the comparability perimeter.
The stablecoin complex is largely ceding the institutional-settlement vertical to tokenised deposits during the window, with the hope of recapturing it post-2027 through yield-distribution workarounds. This asymmetry is underappreciated in narratives that frame stablecoins versus deposits as a neck-and-neck race.
What infrastructure providers capture regardless
Whoever wins the second-bank slot, the horizontal infrastructure layer captures margin. Securitize (NYSE digital transfer agent, BlackRock investor, Vault Registrar EIP candidate) and Zerohash (OCC charter filed, embedded at Morgan Stanley, BlackRock, Stripe, Interactive Brokers) are positioned to provide tokenisation and settlement rails that the second-bank entrant will need. The more fragmented the bank-entry picture, the more valuable a horizontal infrastructure provider becomes. The two firms together cover most of the regulated US perimeter for tokenised-asset issuance and institutional crypto-settlement embedding.
Allocator implications
For banks, the window is the highest-leverage 18 months in the institutional digital-money race. Second-bank entry choices (chain, issuance model, interop partner) outlive the arbitrage by a decade.
For asset managers, the settlement-asset unlock from tokenised-deposit scaling is the quiet catalyst for Wave-2 RWA product adoption. Plan for acceleration rather than deferral.
For corporate treasurers, the yield arbitrage is real and legal; JPMD is the only live option as of April 2026, but the second-bank entrant likely launches within 12 months and a treasury programme structured to take advantage should be in design now.
For DeFi protocols, the window is when public-chain tokenised deposits become the institutional bridge asset. Native integration with tokenised deposits is structurally more valuable than native integration with any specific stablecoin.
APAC angle
The arbitrage is a US-perimeter phenomenon, but the cross-border consequences are real for APAC banks. DBS, Standard Chartered, and HSBC sit on the Partior interbank rail with JPMorgan; if JPMD on Base develops corridor connections to APAC tokenised-deposit rails, the cluster (see Asia institutional cluster vs US fragmentation) extends into a multi-currency tokenised-deposit settlement layer faster than the GENIUS Act timeline alone would suggest. The MAS SCS framework is a separate perimeter for SGD-pegged stablecoins; the SCS regime does not produce the GENIUS-style yield gap because it does not prohibit interest in the same way, but the structural choice between SGD-pegged stablecoin and SGD-denominated tokenised deposit is similar in shape.
Open questions
- Which bank lands the second-mover slot, and on which chain.
- Whether custodian reward programmes operationalise faster than the 12-18 month estimate, in which case the window narrows.
- Whether DBS or another APAC bank issues a USD-denominated tokenised deposit that competes for the same window from a non-US perimeter.
- Whether the second entrant's chain choice triggers a cascade (others follow), or whether the market remains multi-chain with weak interop.