The Permitted Issuer

What the GENIUS Act Actually Does

The Guiding and Establishing National Innovation for U.S. Stablecoins Act — GENIUS, as nobody outside the legislative-staff office calls it — was signed into law on July 18, 2025. It establishes a federal framework for "payment stablecoins" with three load-bearing provisions and one effective date.

From the effective date forward, only a "Permitted Payment Stablecoin Issuer" — a PPSI in the regulatory shorthand — may issue payment stablecoins in or to the United States. The three load-bearing provisions:

1. Reserve backing. Every dollar of outstanding stablecoin must be backed 1:1 with assets from a narrow approved list: U.S. dollar cash, deposits at insured depository institutions, short-term Treasuries, overnight Treasury repos (tri-party, clearing-agency, or highly-creditworthy counterparty), and money-market funds. That's it.
2. No yield to holders. A permitted stablecoin issuer may not offer any form of interest or yield to the stablecoin holders. The float income from the reserves accrues entirely to the issuer.
3. Permitted-issuer gate. To qualify as a PPSI, an entity must be either (a) a federally or state-chartered bank issuing through a subsidiary supervised by the bank's primary federal regulator, (b) a federally licensed nonbank under OCC oversight, or (c) for issuers with less than $10 billion outstanding, a state-licensed nonbank under a regime "substantially similar" to the federal framework.

Anything that does not fit those three constraints — and a great deal of the existing U.S. stablecoin ecosystem does not — must cease U.S. issuance on the effective date.

What Got Quietly Excluded

The most consequential part of any reserve-backing regulation is rarely what's on the approved list. It's what's not.

The approved list is essentially the asset side of a transaction-banking deposit account. By construction, it caps the yield the reserves can generate at the short end of the Treasury curve — roughly five percent in mid-2026, somewhere between two and four percent across any reasonable medium-term forecast. The issuer's gross revenue per dollar of float is bounded by what a tri-party repo desk can deliver. The issuer's expenses are operational and modest. The arithmetic of the business is the arithmetic of a money-market fund whose investors are required to take zero yield.

Who Gets to Issue, by Statute

The fourth box is where the analysis lives. Roughly half of the on-chain stablecoin float as of mid-2026 — and a much larger share of the DeFi-native stablecoin design space — falls into the fourth box. Not because the projects are obviously unsafe (some are, some aren't), but because they were designed around legal vocabulary the GENIUS Act explicitly does not recognize: distributed governance, algorithmic mint/burn mechanisms, yield as a feature, overcollateralized crypto-backed mints.

What Each DeFi-Native Stablecoin Loses, Specifically

Who Captures the $10 Billion

The total U.S.-facing stablecoin float as of mid-2026 is approximately $200 billion. At a short-Treasury yield of roughly 5 percent, that float generates approximately $10 billion in annual gross revenue — entirely to the issuer, under the no-yield rule.

The qualifying issuers, and their approximate market positions, are:

• Circle (USDC). Already aligned with the approved reserve list. Federal trust charter in place. The model-citizen PPSI. Float income proportional to USDC market share — likely the second-largest survivor.
• Tether (USDT). Largest by float ($120B+), but historically based outside the U.S. with a reserve mix that includes commercial paper, gold, secured loans, and other assets not on the GENIUS approved list. To remain a U.S.-facing issuer, Tether must either incorporate as an OCC-licensed nonbank and restructure reserves into the approved list, or accept that U.S. distribution ends. The economic incentive points strongly toward restructuring. The political incentive is more complicated.
• Bank-subsidiary issuers. The new entrants. JPM Coin's successor product, BofA's payments-rail stablecoin, Citi Token Services. These already exist in pilot form. The GENIUS Act gives them the legal cover to scale to retail. Banks now have a fee-free deposit product: customer's dollars sit in 1:1-backed stablecoin reserves, the bank captures the full T-bill yield, and pays the customer zero. This is structurally better economics than a no-interest checking account, because there's no FDIC deposit insurance premium to net out.

The float economics also create a new competitive dynamic. Before GENIUS, the implicit revenue from stablecoin float was contested — some DeFi protocols distributed it to holders (via savings rates, staking yields, or token buybacks), Tether retained most of it as undisclosed profit, Circle retained most of it as disclosed profit. After GENIUS, all of it accrues to the issuer, and the issuer's relevant scale economy is the cost of compliance per dollar of float. That favors the largest existing issuers and the bank subsidiaries.

The Capture Question

The public framing of the GENIUS Act through the 2025 legislative cycle was consumer protection. The justifying narrative was the 2022 collapse of TerraUSD — a $40 billion algorithmic-stablecoin failure that wiped out a substantial portion of retail crypto exposure. The argument went: algorithmic stablecoins are unsafe; users were harmed; Congress must act.

That narrative is true and incomplete. TerraUSD was indeed catastrophic for the retail users who held it. The GENIUS Act does prevent a similar product from being marketed in the U.S. It also does several other things that consumer protection does not require:

• Prohibits yield-bearing stablecoins, including those with conservative 1:1 reserve backing and yields no higher than what a money-market fund pays. The yield prohibition is not a TerraUSD response. TerraUSD's collapse was an algorithmic-design failure, not a yield-on-1:1-reserves failure.
• Excludes from the approved-reserve list any asset that yields meaningfully above short Treasuries — even when that asset is itself low-risk (agency MBS, longer-duration Treasuries). This is a margin-protection rule for issuers, not a consumer-protection rule.
• Requires entity-level chartering and supervision, which has the practical effect of excluding distributed-governance projects entirely. A DAO whose members vote on parameters cannot be served process; it cannot be a PPSI; it cannot operate.

What to Watch Between Now and January 18, 2027