On September 24, the Federal Reserve published a proposal that does something most of the market has not priced: it moves payment stablecoin issuance from the perimeter of the banking system into its balance sheet logic. Full reserve backing. Capital charges against operational and certain credit risk. A two-business-day redemption window. And a 120-day decision clock for bank subsidiaries seeking approval under the GENIUS Act.
That is the whole story in four clauses. The rest is commentary, and most of the commentary is wrong.
I want to be precise about what this is and what it is not. This is not a technology document. There is no novel consensus mechanism, no new cryptographic primitive, no clever cryptographic trick to audit. This is regulatory architecture — the deliberate transplant of bank prudential logic onto an instrument that spent a decade avoiding it. Ledger lines reveal what noise obscures, and the lines here are drawn in the shape of a bank charter.
The GENIUS Act was signed in July 2025. This proposal is the down-stream plumbing. Fed rules for bank-subsidiary issuers. OCC rules for nonbank federal issuers. FDIC rules for insured depository institutions. Three agencies, one framework, one intended outcome: payment stablecoins become regulated monetary instruments rather than offshore financial experiments. If you have been trading the narrative of "regulatory clarity" for the past six months, this is the moment clarity stopped being a slogan and started being a compliance schedule.
Here is the technical substance, stripped of narrative.
Full reserve backing eliminates the maturity mismatch that destroyed Terra and nearly destroyed USDC in March 2023. A full-reserve issuer holds high-quality liquid assets equal to its liabilities at par, continuously. No duration arbitrage. No rehypothecation surface. The proposal says what I have said in post-mortems for three years: if a liability is redeemable at par on demand, the asset side must be able to meet that par on demand. It is not a design preference. It is an accounting identity.
The two-business-day redemption window is the anti-bank-run clause. Compare it to money market funds, which operate on T+1 to T+2, and to bank deposits, which legally can take longer. Two days is tight enough to prevent a slow bleed but not tight enough to force a fire sale of reserve assets on day zero. It is the same instinct that produced liquidity coverage ratios in Basel III, scaled to a liability that clears on-chain in seconds. Liquidity is the current of truth, and the Fed has now told issuers exactly how fast that current must run.
Capital charges against operational and certain credit risk are the tell. This is the part the market is ignoring. If the Fed believed full reserves made an issuer risk-free, no capital charge would be necessary. The presence of a charge admits the residual truth: even a fully reserved issuer can lose money through custody failure, operational error, or deterioration in the credit quality of the reserve assets themselves. The phrase "certain credit risks" is doing extraordinary work in that sentence. It implies the reserve basket is not restricted to cash and short-dated Treasuries. There is likely a limited credit surface — repo, reverse repo, maybe agency paper — permitted with a capital buffer against it. I flag this at low confidence because the published text we have is truncated, and anyone telling you they know the exact eligible-asset schedule is guessing.
The 120-day decision clock is the least discussed and most consequential clause. Under the GENIUS Act, bank applications for stablecoin issuance subsidiaries must receive a decision within 120 days. That is a pre-commitment device. It converts the regulator from an unpredictable gatekeeper into a scheduled processor. For a bank treasury committee deciding whether to build a stablecoin business, uncertainty was the binding constraint — not cost, not technology. Removing the indefinite-wait risk is worth more than any single basis point of capital charge. Standardization survives the chaos of collapse, and this clause is standardization of the approval path itself.
Now the part that actually affects P&L, because that is where most people stop reading and where the money is made.
A stablecoin issuer makes money one way. It holds reserves earning interest — almost entirely short-dated US Treasuries — against a liability that pays the holder nothing. That spread is the float. Circle and Tether both live on it. It is a beautiful business because it scales with the liability and the cost of the liability is zero.
The Fed proposal does not kill the float. Full reserve backing preserves it. What the proposal does is attach a cost of capital to the balance sheet that generates the float. Every dollar of required capital is a dollar that cannot be deployed elsewhere, and it earns the risk-free rate rather than the issuer's return on equity. For JPMorgan, Bank of America, and Citi — institutions with deep capital bases and cheap funding — this is a rounding error. For a thinly capitalized issuer, it is an existential margin compression.
I built a version of this model during the 2020 DeFi summer when I ran a $2 million alpha fund on Curve's 3pool. The lesson then was identical to the lesson now: when a regulatory cost is fixed in absolute terms, it is regressive against small players. A capital charge is a fixed cost per unit of liability. It does not care whether you are Circle or a three-person startup in Zug. The result is predictable. Compliance cost rises, capital cost rises, and the industry consolidates toward whoever can amortize both across the largest liability base.
The 2022 Terra collapse taught me the same lesson from the other direction. I liquidated 80% of algorithmic stablecoin exposure inside 48 hours on reserve-anomaly data alone, and the reason I could do that without hesitation is that I had pre-committed to a cost framework before the crisis, not during it. Bear markets demand disciplined forensics, but the forensics have to be designed in the bull market. This Fed proposal is the industry being handed, on a bull-market timetable, the forensic framework it declined to build for itself.
Here is where I part company with the consensus read.
The consensus says this is bullish for stablecoins and therefore bullish for the entire asset class. I think the directional call is right and the reasoning is lazy. The bullishness is not in "regulation." It is in the specific structural fact that the Fed has chosen bank regulation as the template rather than securities regulation. Payment stablecoins are being defined as payment instruments, not investment contracts. Run the Howey test and it fails cleanly on three of four prongs: there is no common enterprise, no expectation of profit from the issuer's efforts, and no profit participation because the GENIUS Act framework bars interest paid to holders. That definitional choice removes the single largest tail risk the sector carried — the risk that a US regulator would one day classify a dollar-pegged token as an unregistered security. The market has not fully repriced that tail risk. It is worth more than any fee schedule.
The second consensus error is treating this as a surprise. It is not. The GENIUS Act passed in July. The market has had two months to anticipate implementing rules. I would estimate 60 to 70 percent of the structural impact is already in compliant-issuer valuations. If you are buying the news, you are buying the residual.
The third error — and this is the one I care about — is the assumption that clearer rules are neutral across the stablecoin universe. They are not. They are fundamentally redistributive.
Consider the ecosystem as a flow, because that is what it is. Upstream sits the reserve layer: US Treasuries, custodians, and the Fed as rulemaker. Midstream sits the issuer. Downstream sits everything that touches the dollar on-chain — exchanges, DeFi protocols, payment processors, and cross-border settlement rails.
The proposal reshapes who is allowed to occupy midstream. The bank-subsidiary approval path hands a compliant issuance channel to institutions that already have charters, capital, and deposit franchises. That is not a neutral expansion of access. It is an invitation to a specific category of player. And those players do not need to win on yield, because they do not pay yield. They win on trust and on the ability to bridge tokenized deposits and stablecoin liabilities inside a single balance sheet.
That bridge is the thing to watch. If a bank-subsidiary issuer can move value between a tokenized deposit and a payment stablecoin without leaving its own ledger, the boundary between "bank money" and "stablecoin money" stops being a boundary and becomes a product feature. The regulatory question that follows — is this a deposit, and if so is it insured, and if not what is its resolution regime — is not answered by this proposal. It is deferred to a later rulemaking. Anyone modeling this business as finished is modeling a draft.
Follow the flow, not the headline. The headline is "Fed regulates stablecoins." The flow is that the Fed has built a compliant issuance channel that favors balance-sheet-heavy institutions and imposes a hard, non-negotiable liquidity constraint on everyone. Efficiency is the only permanent alpha, and the efficiency here accrues to scale.
Offshore issuers face the mirror image. The clearer the US framework becomes, the higher the implicit bar for a non-compliant issuer to touch US institutional money. That is not a ban. It is a compliance cliff. The issuer either restructures its reserve composition, its custody, its reporting, and its audit cadence to meet a Fed-equivalent standard, or it accepts that US institutional allocation flows away from it. I have watched enough reserve disclosures to know that the choice is not free, and the timeline is not generous.
Now the contrarian angle, stated plainly, because it is the part most analysts will skip.
Regulatory clarity is not the same thing as regulatory favor, and it is certainly not the same thing as decentralization. The most likely medium-term outcome of this framework is not a thriving open stablecoin market. It is a concentrated one, where issuance permission sits with a small number of chartered entities and the rest of the ecosystem builds on top of them. That is a real tradeoff, and the market is currently pricing only the upside of the trade.
There is also a genuine tension the proposal creates and does not resolve. Compliant stablecoins are attractive precisely because they are permissioned — KYC'd, freezable, sanctionable, auditable. Unpermissioned DeFi is attractive precisely because it is not. The more liquidity migrates into compliant stablecoins, the more the largest DeFi pools are denominated in assets that a compliance officer can freeze. That is not a bug in the framework. It is the framework working as designed. But it means the composability assumptions that DeFi protocols baked into their architecture now rest on top of a regulatory layer that can, under defined conditions, immobilize the base asset. Smart contract risk used to be the tail. Now the tail includes the reserve custodian's compliance posture.
I would also push back on the framing that this is the final rule. It is not. This is a draft entering public comment, and the comment period is where the banking lobby and the crypto lobby will collide. Banks will argue the capital charges are too light relative to the deposit-substitution risk they now carry. Crypto-native issuers will argue the capital charges and the two-day redemption window are calibrated for institutions and will crush non-bank innovation. The final text will move. Treating draft parameters as binding is a common error, and it is an expensive one.
So what do I actually watch over the next quarter, and what should you?
First, the Federal Register entry. Everything I flagged at low confidence — the eligible reserve asset schedule, the calibration of the capital charge, the treatment of operational risk — lives in the full text. If the eligible-asset list permits non-Treasury credit exposure, the float economics of compliant issuers change materially. If it does not, the float becomes an almost pure duration-free Treasury carry, and the business model simplifies toward a fee on scale. Either way, the full document is the first thing to read.
Second, the parallel rulemakings. The OCC and FDIC have their own pieces of the GENIUS Act framework to implement. The shape of the whole — whether the US ends up with a two-track or three-track stablecoin regime — only becomes legible when those land. A single rule tells you the Fed's preferences. Three rules tell you the architecture.
Third, the first bank subsidiary approval. Whoever gets through the 120-day clock first sets the practical template. Their reserve composition, custody arrangement, and disclosed capital position become the de facto benchmark every subsequent applicant is measured against. It is the closest thing to an observable standard a draft can produce.
Fourth, reserve disclosure cadence from the offshore issuer. If the largest non-US issuer adjusts its reserve mix toward short Treasuries and shortens its custody chain in response to this proposal without being legally required to, that is the clearest signal that the compliance cliff is real and the migration is underway.
Code does not lie, only developers do. Regulation is not code, but it behaves like it once it is enforced — it constrains what transactions are possible, and it does so without sentiment. The Fed has just written a constraint that says: to issue dollars on-chain at institutional scale, you must hold dollars in reserve, you must be able to return them in two days, and you must pay for the privilege with capital. That is a clean, legible, and expensive rule. It is also the strongest legitimacy signal the stablecoin sector has ever received.
The question for the next quarter is not whether stablecoins are regulated. They are. The question is who this framework is quietly designed to let issue them, and whether the answer looks more like a bank branch than an open protocol. Watch the black-and-white text, not the narrative. Check the source. Verify the schedule. The graph clarifies what sentiment confuses — and the graph on issuance concentration is about to bend.