Stablecoin settlement for corporate treasury

For most of a decade, the treasury conversation about stablecoins ended at the first question from the controller: what is this on the balance sheet, and who stands behind redemption. In 2026 those questions finally have statutory answers on both sides of the Atlantic, and that changes the shape of the discussion. The question is no longer whether stablecoins are legitimate infrastructure; it is where, if anywhere, they earn a place in a specific company’s settlement stack.

The US GENIUS Act, enacted in July 2025, creates the first federal framework for payment stablecoins: digital assets designed for payment or settlement, with a redemption obligation at fixed value. It takes effect at the earlier of 18 months after enactment or 120 days after final implementing rules, which absent earlier final rules lands on 18 January 2027. In the EU, MiCA’s e-money token category covers stablecoins referencing a single official currency, regulated along the lines of electronic money, which makes EMTs the relevant instrument for euro-corridor settlement. Two regimes, converging on the same idea: a regulated, redeemable, reserve-backed payment instrument. Not identical regimes, though, and the differences still require jurisdiction-by-jurisdiction legal review.

What the statutes actually permit, and refuse

The load-bearing concept in GENIUS is the permitted payment stablecoin issuer: a subsidiary of an insured depository institution, a federally qualified nonbank supervised by the OCC, or a state-qualified issuer. Permitted issuers hold 1:1 reserves in high-quality liquid assets, cash, demand deposits, short-term Treasuries, Treasury-backed repos, with monthly public disclosure and independent examination. Treasury’s 2026 proposed implementing rule treats them as financial institutions for Bank Secrecy Act purposes, pulling issuance fully inside the AML perimeter.

The negative space matters more for treasurers than the permissions. A stablecoin from a non-permitted issuer shall not be treated as cash or a cash equivalent for accounting purposes, is not eligible as cash-equivalent margin or collateral, and is not an acceptable settlement asset for wholesale payments between banking organisations. And even permitted-issuer tokens are not legal tender, not deposits, not securities, and issuers get no automatic access to Federal Reserve master accounts. The settlement stack still runs through intermediaries, and your counterparty analysis should assume so.

The same 2026 rulemaking wave produced something arguably more consequential for conservative treasuries: the FDIC’s proposal that tokenized deposits meeting the statutory definition of a deposit are treated no differently under the Federal Deposit Insurance Act than any other deposit. A tokenized deposit is a bank liability on a ledger, which keeps it inside the accounting, insurance, and counterparty framework treasury already understands. For flows that stay within one banking relationship, that is usually the shorter path than a third-party stablecoin.

Where the rail actually earns its place

Strip away the enthusiasm and the credible treasury use case is narrow and specific: short-duration settlement, not storage. The design pattern that follows from redemption-at-par and reserve rules is prefund, move, reconcile, redeem. Value enters the token for hours or days, does its job, and exits into local currency. Holding stablecoins as a strategic reserve asset solves no treasury problem that a Treasury bill does not solve better.

Within that pattern, the corridors that already work:

• Intercompany settlement. Moving value from a US parent to an EU or LATAM subsidiary without correspondent-banking cutoffs, then sweeping into local currency where permitted. The win is timing certainty, and it shows up directly in how much buffer liquidity subsidiaries must carry.
• Marketplace and platform payouts. Mass disbursements to vendors, contractors, or creators, with settlement finality in minutes and reconciliation drivable from the ledger itself.
• 24/7 liquidity moves. Weekend and after-hours transfers between bank accounts, exchange venues, and payment platforms while fiat rails are closed. For any business with meaningful weekend flows, this is the corridor that converts sceptics.
• Cross-border trade settlement. Short invoice-to-payment cycles where FX conversion can be centralised at the treasury centre rather than executed badly at each endpoint.
• Treasury-to-exchange settlement. Funding digital-asset operations where a regulated payment stablecoin beats cards and wires on both speed and reconciliation.

The honest filter for each candidate corridor: does the flow cross platforms, time zones, or operating hours in a way that bank rails handle poorly, and are the counterparties, custodians, and venues on both ends already inside the regulated perimeter. If either answer is no, the corridor is a pilot in search of a problem.

The risk register, written down before the pilot

Treasury adoption fails on unmanaged edge cases, not on grand strategy, so enumerate them.

Classification and accounting risk. Only permitted-issuer tokens get the favourable treatment; issuer status is therefore a standing due-diligence item, not a one-time check. Redemption risk. Par redeemability is only as good as reserve quality, redemption windows, and issuer solvency; read the monthly disclosures the way you read money-market fund holdings. Banking-access risk. Issuers depend on intermediary banks, so an issuer’s banking arrangements are part of your counterparty analysis. AML and sanctions risk. Transferability creates wallet-screening, travel-rule, and counterparty obligations that scale with corridor count. Custody risk. Key compromise, settlement-finality errors, and internal fraud are material wherever governance is thin. Fragmentation risk. GENIUS and MiCA rhyme but do not match; a corridor legal opinion is per-jurisdiction work.

Building it: custody, controls, and the operational spine

Expect the operating model to be split across four roles rather than bought from one provider: the issuer, a qualified custodian (GENIUS-era proposals point clearly at regulated banking custody rather than informal wallet services), an on-chain transfer agent, and your own ERP and controls layer where policy actually lives.

The controls translate directly from payment operations. Segregation of assets at the custodian. Whitelisted addresses so value can only reach pre-approved counterparties. Role-based dual authorisation with limits, mirroring existing payment mandates. Sanctions and wallet screening on every transfer. Daily reconciliation of on-chain positions against the general ledger, which is easier than bank reconciliation in one way, the chain does not close overnight, and harder in another, because the ledger never stops and your close process assumes things do.

Redemption governance is the piece teams skip and regret. Decide in advance who can redeem, through which channel, in what size, on what notice, and test the path with real value before an urgent Friday evening forces the first attempt. A settlement rail you have never exited under pressure is not yet a settlement rail. Run the first corridor small, measure cycle time and reconciliation effort against the wire-based baseline, and let those numbers, not the technology, decide whether corridor two exists.