Suit Up · Asia's institutional desk on tokenisation

The thesis

Stablecoins emerged across 2017-2022 with an explicit pitch: bearer-style, programmable dollar exposure that operates outside the bank deposit-and-payment perimeter. The regulatory response that crystallised across 2024-2026 (GENIUS, MiCA EMT, the Hong Kong Stablecoins Ordinance, the Japan PSA EPI amendments, the MAS SCS framework) did not reject the form. It absorbed it. The five regimes converge, drafted independently and from different starting positions, on a recognisable structural shape: a regulated issuer, segregated reserves of high-quality liquid assets, par redemption on demand, supervised custody, and capital and liquidity requirements scaled to the redemption obligation. That is a new tier of regulated money supervision under a new licensing wrapper, and structurally it sits much closer to the bank-money discipline covered in Bank money, Part 1 than the original stablecoin pitch let on. The branding is doing work that the substance does not support.

The setup

When Tether launched in 2014 and USDC in 2018, the implicit contract with the holder was that the instrument would behave like a dollar without travelling through the deposit franchise. The customer relationship was not a deposit relationship. The reserves were not subject to the supervisory perimeter that surrounds a bank balance sheet. Bearer-style transferability on chain replaced the account-based world where the bank knows who holds what at every moment. For the early growth period, the absence of a regulated wrapper was not a bug; it was the product.

The institutional adoption story since 2022 changed that pitch in two directions at once. The instruments survived. The regulatory perimeter around them moved decisively toward bank-money discipline. The simplest reading of why is structural rather than political: an instrument that promises par-value redemption with broad public holding has the same systemic-risk profile as a bank deposit, and supervisors converge on what works for the deposit franchise because it has been litigated through generations of bank failures. The five-regime convergence is not a coincidence and it is not regulatory overreach. It is recognition that the underlying object is closer to bank money than the original positioning admitted.

The argument

The five regimes' convergent structure

Compare the regimes point by point and the family resemblance is hard to miss.

On regulated issuer status: GENIUS confines issuance to subsidiaries of insured depository institutions, federally chartered non-bank financial institutions supervised by the OCC, or state-qualified issuers below USD 10 billion outstanding. MiCA EMT restricts issuance to credit institutions or authorised electronic money institutions. The Hong Kong Ordinance issues a standalone HKMA licence with banks eligible to apply, the licence shape itself non-bank. Japan's PSA EPI regime runs three routes (bank-direct, fund-transfer service provider, or trust company), each inheriting the parent licence's supervisory perimeter. The MAS SCS framework requires a Major Payment Institution (MPI) licence as the issuer of record. Five different statutes, drafted on five different timelines, converge on the same answer: the issuer must be a supervised entity, and a thin-wrapper offshore vehicle cannot stand at arm's length from the obligation.

On reserve composition, redemption, and custody: each regime restricts reserves to a narrow set of HQLA in the pegged currency (cash, short-dated government securities, repo against the same, qualifying money-market funds), each gives the holder a statutory par-redemption right on a deposit-equivalent timeline (MiCA "any time, free of fees", Hong Kong's one-business-day window the tightest of the five), and each requires segregated custody at a supervised third party with bankruptcy-remote treatment. The asset side is as close to the safest dollar (or yen, HKD, SGD) as each regime can specify; the redemption mechanic maps the holder's relationship onto a deposit rather than a security; the custody logic is borrowed directly from how deposit-taking institutions ring-fence customer assets.

On capital and liquidity: each regime imposes capital sizing scaled to the redemption obligation. MiCA EMT issuers face EMI capital plus stablecoin-specific add-ons; GENIUS bank-route issuers carry parent prudential capital and OCC-chartered NBFIs face capital sized to operating risk plus reserve buffers; the Hong Kong Ordinance requires HK$25 million paid-up share capital plus 12 months of operating-expense coverage; MAS SCS layers stablecoin-specific add-ons onto the MPI base capital. The numbers differ. The shape is the same: capital sized to the run-and-redemption profile of an instrument the customer treats as cash.

The point is not that the five regimes are identical. They differ at the margin, and the stablecoin licensing decision tree walks through the margins in detail. The point is that they converge structurally on what a bank treasurer would recognise as the supervisory architecture surrounding a deposit-equivalent instrument. The convergence is, in substance, bank-money discipline applied to non-bank issuers.

The bank-money-vs-stablecoin distinction collapses operationally even as it persists legally

The legal distinction between a USDC holder and a JPMorgan deposit holder is real and we do not contest it. The USDC holder has a contractual claim on Circle, secured (in practice and in posture) by Circle's segregated reserves at supervised custodians. The JPMorgan deposit holder has a deposit liability against JPMorgan's full balance sheet, with the bank's prudential capital, the FDIC insurance overlay up to the cap, and the Federal Reserve's lender-of-last-resort framework standing behind it. The resolution regimes are different. The deposit-insurance posture is different. The relationship to the wholesale settlement layer (covered in Bank money, Part 4) is different. In tail-risk scenarios, those distinctions matter acutely.

In normal-state operations, they collapse. Both instruments deliver the same operational service: dollar exposure, on-demand par redemption, from a regulated issuer holding HQLA reserves under supervisory cover. For most institutional treasury use cases (operational cash, FX settlement, intraday liquidity, cross-border B2B payment), the operational equivalence is what matters. All five stablecoin regimes and the tokenised-deposit construction (covered in Tokenised deposits, Part 1 and the tokenise-deposits playbook) clear those operational tests in normal-state. The two products compete on price (yield, fees), programmability, and distribution, not on structural fitness. The deposit-stablecoin arbitrage window is the cleanest current articulation of that competition in the US perimeter: GENIUS prohibits stablecoin issuers from paying interest, while tokenised deposits inside the existing banking licence can pay the prevailing deposit rate.

The tail-risk asymmetry is real. A bank failure unwinds through resolution; a stablecoin issuer failure unwinds through whatever segregated-reserve recovery mechanism the regime contemplates. Deposit insurance covers the first up to the cap; nothing comparable covers the second at the holder level. Our read is that the asymmetry is currently underpriced in the spread for retail and small-business holders and roughly correct for institutional balances that sit far above any insurance cap anyway. That second observation is part of why institutional adoption of regulated stablecoins has scaled at the pace it has: the insurance gap that bites for retail is structurally less binding for the corporate-treasury use case where most of the early flow lives.

Implications for the institutional product landscape

Two consequences follow from the convergence.

The first is that the operational competition between bank-issued tokenised deposits and stablecoins intensifies. Banks are realising they can offer the same operational fit as a bank product, capturing the deposit franchise that stablecoin distribution might otherwise erode. JPMorgan's Kinexys Digital Payments is the most-developed worked example. DBS Token Services in Singapore, Wells Fargo Digital Cash in the US, and HSBC's Orion-side tokenised-deposit work in the UK and Hong Kong are all recognisably the same play: build the bank-money rail inside the existing licence, capture institutional flow, and route around the structural gap that a stablecoin construction would impose. Non-bank stablecoin issuers, in parallel, are realising the licensing burden makes them quasi-banks anyway. Circle's OCC trust-bank charter route, Paxos's NYDFS limited-purpose trust framework migrating toward OCC chartering, and the supervisory architecture each issuer has built around its reserves are all examples of non-bank stablecoin operations converging toward bank-equivalent supervisory posture. The freedom-from-bank-regulation pitch that animated the early growth is structurally hollow under the post-2024 regimes; the licensing burden you accept under GENIUS, MiCA, the Hong Kong Ordinance, the PSA EPI regime, or MAS SCS is, by design, comparable to what a bank carries for an equivalent deposit-substitute product.

The second is that the regulatory perimeter for both products is converging upward, not downward. The regulatory floor under stablecoins is rising as the licensing matures: each successive iteration of supervisory guidance under each of the five regimes tightens reserve composition, redemption discipline, and disclosure cadence. The regulatory ceiling on tokenised deposits is being explored as the wholesale-settlement use cases scale; Basel SCO60 Group 1a treatment, supervisor expectations on chain governance, and reconciliation discipline are all areas where the live cohort is getting more guidance, not less. The two perimeters are meeting near the deposit-equivalent supervisory baseline rather than at some intermediate "lighter than bank, heavier than stablecoin" level. That meeting point is the structural conclusion of the convergence.

Implications for practitioners

For bank treasury and digital-asset desk leads: tokenised deposits are not a defensive play against stablecoins. They are an offensive play to capture the deposit franchise that stablecoin distribution might otherwise erode. The relevant competitor for most institutional flow is not the next stablecoin issuer in your jurisdiction; it is the bank across the street that has shipped its tokenised-deposit rail first and locked in the corporate-treasury relationship. Sequencing matters. The tokenise-deposits playbook walks through the seven design decisions; the load-bearing one is whether to ship intra-bank first or stake an in-house cross-bank build, and the answer is intra-bank first for almost everyone except the GSIB cohort that already has one in production.

For non-bank stablecoin issuers: the licensing-burden moat is asymmetric. You need to comply with bank-money-equivalent supervision; the bank does not need to comply with anything additional to issue a tokenised deposit. The freedom-from-bank-regulation pitch that anchored the early growth is no longer available. The durable moat is rail integration and brand, not regulatory arbitrage. Circle and Paxos have built the brand position; JPYC is building the FTSP-route brand in Japan; the next-tier issuers will need to compete on distribution, programmability, and rail integration rather than on structural innovation, because the structural envelope is now defined by the regulator.

For allocators choosing between the two products as a treasury cash equivalent: the operational equivalence in normal-state is real; the tail-risk asymmetry is real; price the difference accordingly. For balances below the insurance cap and use cases where deposit insurance is meaningful, the bank-money option is cheaper at the structural level. For institutional balances above the cap, the spread should reflect resolution-regime differences and the absence of a holder-level insurance overlay on the stablecoin side, but the spread should not assume a category gap that the regulatory work has substantially closed.

For founders considering a stablecoin launch in 2026: the regime is set. The licensing pathway is clear in each of the five jurisdictions. Differentiation will come from rail integration and distribution, not from structural innovation. A new stablecoin issuer that does not have a credible distribution thesis (a corporate partner, a wallet integration, a cross-border corridor with under-served flow) is competing against incumbents who already cleared the licensing bar and built the supervisory relationships. The launch question is no longer "what reserve composition" or "what redemption mechanic"; it is "what flow does this instrument route that an incumbent does not".

Where this could be wrong

Two scenarios would force a meaningful update.

The first is a fully-reserved non-bank issuer scaling to GSIB-equivalent flow at a meaningfully lower cost-base than tokenised-deposit-issuing banks. If a Circle, a Paxos, or a not-yet-emerged competitor demonstrates that the non-bank licensing perimeter can be operated at a structural cost advantage versus the in-bank tokenised-deposit construction, the convergence partially reverses on the cost side even if the supervisory shape continues to converge. We do not currently see strong evidence for this, but the cost-base data on each product is opaque enough that the case is not closed.

The second is a major stablecoin issuer event (run, depeg, fraud) that causes regulators to either further harden the regime or partially deregulate. Both directions are possible. A hardening response (tighter reserve rules, deposit-insurance-equivalent overlays, prudential capital scaled toward bank levels) deepens the convergence and accelerates the structural absorption into bank-money supervision. A deregulatory response (loosening reserve composition, permitting interest-bearing constructions, easing custody requirements) reverses the convergence and reopens the structural gap between the two products. Our read is that the hardening direction is more likely under any plausible supervisory consensus, but the deregulatory direction has enough political support in some jurisdictions that it cannot be ruled out across a 2026-2028 window.

Read across both scenarios, our directional view is that the convergence persists and probably deepens through 2028, with the operational competition between tokenised deposits and regulated stablecoins as the dominant dynamic in institutional digital-money. The branding distinction will outlive the structural distinction by years, but the structural distinction is already mostly closed.

[Suit Up]

Stablecoins are bank money in disguise; the regulators have noticed