Fed Docket R-1899 prices stablecoins at 2% capital, 1% past $50bn

The Federal Reserve Board’s proposal under Docket No. R-1899 (RIN 7100-AH29), released on September 24, 2026, sets the first federal formula for payment stablecoin capital, and the rate falls as an issuer grows: 2% of the first $20 billion outstanding, 1.5% of the next $30 billion and 1% above $50 billion, before a 50% surcharge on new entrants.

R-1899 is the Board’s main rule under the Guiding and Establishing National Innovation for U.S. Stablecoins Act (GENIUS Act), and its capital chapter, proposed 12 CFR 247.15 to 247.18, is where the money is. The 392-page notice of proposed rulemaking (NPRM) pairs a tiered operational-risk charge with a levy on non-reserve revenue, a 2% charge on uninsured deposits held as reserves, and forced liquidation if a capital shortfall survives a second quarter-end. Below: the arithmetic, a comparison with the Office of the Comptroller of the Currency (OCC), the EU’s Markets in Crypto-Assets Regulation (MiCA) and the UK Financial Conduct Authority (FCA), and who pays most.

Key facts

• Tiered operational charge: 2.0% of the first $20 billion of outstanding issuance, 1.5% of the next $30 billion, 1.0% above $50 billion; $400 million at $20 billion and $850 million at $50 billion (proposed § 247.18(b)(1), R-1899 NPRM).
• Revenue add-on: 25% of three-year average non-reserve revenue (R-1899 NPRM).
• Reserve credit charge: 2% of reserves held as uninsured deposits, and 2% of any collateral shortfall on reverse repos (proposed § 247.17(a)).
• New-entrant surcharge: an issuer with no loss history gets a loss scalar of 1.5; the floor is 0.8 (proposed § 247.18(b)(2), (c)(2)).
• Wind-down trigger: a capital plan within five business days of a quarter-end shortfall, then liquidation if the shortfall remains at the next quarter-end (proposed § 247.15(c)).
• Transition: state-qualified uninsured depository institutions above $10 billion outstanding move to the Board within 360 days, with a first examination within six months of notice (Board staff memo, September 3, 2026).
• Comment window: 60 days from Federal Register publication, still pending on September 28, 2026 (Federal Reserve press release).

Methodology and sources for this analysis

This piece rests on four Federal Reserve documents released on September 24, 2026: the press release, the R-1899 NPRM, the Board staff memo dated September 3, 2026, and the companion application-procedures proposal, Docket No. R-1900 (RIN 7100-AH30). Every capital figure comes from the NPRM’s tables and rule text. The comparison uses the OCC’s GENIUS Act proposal at 91 FR 10202 (March 2, 2026), Articles 35, 45 and 58 of MiCA (Regulation (EU) 2023/1114) and the FCA’s Policy Statement PS26/12. Dollar illustrations are this publication’s arithmetic on the proposed formula and exclude the revenue component. The window runs from the September 24 release to September 28, 2026; state capital regimes and the OCC, FDIC and NCUA rules are covered only as comparators.

Related:CFTC Letter 26-27 makes mention markets presumptively manipulable

What Docket R-1899 actually charges, line by line

The Board chose a standardised formula over issuer-built models, warning that a non-standardised approach “increases the likelihood that risks could be underassessed.” Regulatory capital is tangible equity: common stock, qualifying perpetual preferred stock, retained earnings and accumulated other comprehensive income, with goodwill and intangibles excluded. The minimum under proposed § 247.15(b) stacks reserve credit charges, a bank-rule charge on non-reserve assets and, largest, an operational-risk requirement.

The Docket R-1899 operational charge is a graduated capital requirement on a Board-supervised permitted payment stablecoin issuer’s average daily outstanding issuance over the quarter. Under proposed § 247.18(b)(1), the issuer holds tangible equity of 2.0% on the first $20 billion, $400 million plus 1.5% of the excess between $20 billion and $50 billion, and $850 million plus 1.0% of anything above $50 billion, with thresholds indexed to nominal GDP. It adds 25% of the three-year average of annual revenue that does not come from reserve assets, such as custody fees. The total is multiplied by a loss scalar driven by the issuer’s five-year operational-loss record, with a floor of 0.8 and no fixed cap. The Board’s stated reason for the taper is that operational losses “may not scale linearly” with issuance, because a large unauthorised mint is detected faster and is harder to launder. An issuer without five years of loss data is assigned a scalar of 1.5.

One calibration detail matters more than the headline. The NPRM derives the 25% revenue figure by assuming reserves earn about 4%, matching the 4.07% average three-month Treasury bill rate in 2025. On that assumption, 25% of revenue “corresponds approximately to a requirement of one percent of assets, which is the marginal capital requirement that would apply to the largest” issuers. The Board’s own benchmark is the 1% top tier; the 1.5% and 2% tiers are surcharges on smaller balance sheets. The $5 million de novo floor, which applies for three years after approval, only binds below $250 million outstanding.

Source: The Industry Spread calculation from proposed § 247.18(b)(1)-(2) of the R-1899 NPRM. Excludes the revenue component, reserve credit charge and GDP indexation. The $75 billion and $180 billion rows approximate the late-2025 USDC and USDT supply ranges cited in the NPRM; neither issuer is a Board-supervised issuer.

Why the regressive curve favours incumbents

A $1 billion entrant with no loss history faces an effective 3% charge. An issuer at the NPRM’s own USDT scale, with five clean years of data, would face 0.96%. The threefold gap favours existing concentration. The NPRM itself says network effects make it “harder for smaller competitors to break through,” then proposes a curve that charges them more per dollar.

The taper echoes Circle’s November 2025 letter to Treasury, which the Fed cites. But the Board concedes there “is currently insufficient available data” on stablecoin operational losses and that the schedule rests “mostly” on qualitative analysis. The regressive shape is a judgment, not a measurement, and Question 157 invites commenters to argue for a flat 1% to 2% charge plus a fixed $1 million to $25 million.

“As such, a flat, per-unit capital charge can perversely incentivize risk-taking in technology or operations to offset costs that are not captured in the fixed cost approach, for example, offering payment stablecoins on untested blockchains or failing to adequately account for cybersecurity risk.”

— Circle Internet Group, comment letter to the US Treasury, November 4, 2025 (Regulations.gov, TREAS-DO-2025-0037-0259)

Circle also asked for “supervisor-approved internal models,” which the Board rejected.

The 2% uninsured-deposit charge pushes reserves into Treasury bills

Proposed § 247.17(a)(1) applies 2% to reserves held as uninsured deposit claims. Insured deposits and Treasury bills maturing in 93 days or less carry no charge. The Board benchmarked it to the 1.6% bank-rule equivalent and cites the March 2023 episode in which Circle had about $3.3 billion of USDC reserves at Silicon Valley Bank when it failed.

With insurance capped at $250,000 per depositor per bank, nearly every deposit dollar of a large issuer is uninsured. An issuer with $20 billion outstanding that keeps $4 billion of reserves in bank deposits adds $80 million of capital, 20% on top of its $400 million operational baseline. Reverse repos carry the same 2% on any shortfall after haircuts of 0% for Treasuries of 93 days or less, 0.35% to one year, 1.41% to five years and 2.83% beyond.

The incentive moves reserves from bank deposits into Treasuries, against the grain of the tokenised-deposit push; see our report on the Dallas Fed’s $700bn tokenised-deposit estimate. Question 150 floats a graded alternative of 2% for investment-grade banks, 3% for speculative grade and 4% below that.

A wind-down trigger, not a capital buffer

Bank capital is designed to be drawn down. R-1899’s is not. Under proposed § 247.15(c), an issuer below its minimum at a quarter-end files a capital plan within five business days that must be “feasible for returning” to compliance “within one quarter.” If the shortfall remains at the next quarter-end, it must, starting the next day, begin liquidating all reserve assets, redeem outstanding coins without fees and stop issuing.

The Board’s worked example: short on March 31, liquidation begins on July 1 if capital is not restored by June 30. That is one quarter from the first failed test to a forced exit, automatic rather than discretionary. A reserve-backing breach moves faster: a plan within 24 hours, then liquidation the next business day unless the Board directs otherwise.

The logic: an issuer holds little but reserves and stablecoin liabilities, so equity exists to absorb operational losses while redemptions continue at par. The cost is cliff risk: holders may redeem early once a first shortfall is public. Question 173 asks whether the cure period should be two quarters.

“Stablecoins will only be stable if they can be reliably and promptly redeemed at par in a range of conditions. This includes during market stress, when pressure can be put on the value of even otherwise liquid government debt, and during episodes of strain on the individual issuer or its related entities.”

— Michael S. Barr, Governor, Federal Reserve Board (statement, September 24, 2026)

How the Fed, OCC, EU and UK price the same risk

Sources: Federal Reserve R-1899 NPRM; OCC, 91 FR 10202; Regulation (EU) 2023/1114; FCA PS26/12. Last updated: September 28, 2026.

The regimes split on whether the biggest issuer should hold proportionally more capital or less. MiCA says more, lifting significant tokens from 2% to 3%. The Fed says less, on the theory that operational losses scale sub-linearly. The FCA settled on a flat 1% after deciding 2% overstated operational risk. The OCC set no number, preferring firm-by-firm capital plus a year of expenses in liquid assets.

For a US charter choice, that is arbitrage with money attached: the Fed’s schedule, an OCC negotiation, or a state regime below $10 billion. See our coverage of the GENIUS Act and MiCA split and the FCA’s final crypto rules.

Docket R-1900 decides who gets priced at all

Docket R-1900 is the Board’s companion rule setting how an insured state member bank applies for approval to run a subsidiary that issues payment stablecoins, and it decides who ever faces the R-1899 capital schedule. Under proposed 12 CFR 247.30, the application is a letter with a business plan, financial information, policies and procedures, capital-structure documents where the subsidiary is not wholly owned, biographical reports and certifications. Under section 5 of the GENIUS Act, the Board has 30 days to say whether an application is substantially complete and 120 days from that submission date to decide; if it misses that deadline, the application is deemed approved. A denied applicant can request a hearing within 30 days. A material change, such as weaker finances, a revised business plan or new principal shareholders, can restart the clock.

The gate is narrow. The NPRM counts 703 insured state member banks at December 31, 2025, 647 of them community banks. Everyone outside that population and the $10 billion state transition is priced by another regulator; see our report on the OCC’s crypto trust charters.

Enforcement context: why the Board wants an automatic exit

The clearest precedent is In the Matter of Tether Holdings Limited et al., CFTC Docket No. 22-04, settled on October 15, 2021. The Commodity Futures Trading Commission (CFTC) ordered Tether to pay a $41 million civil monetary penalty for misrepresenting that USDT was fully backed. The order found Tether held enough fiat reserves to back tokens in circulation on only 27.6% of days in a 26-month sample from 2016 to 2018, and that it commingled reserves with Bitfinex’s operational and customer funds.

That order came years after the misstatements, against an issuer with no capital requirement. R-1899 turns the lesson into rule text: segregated reserves, weekly confidential reports on issuance, redemptions and reserves, and a fixed exit timetable. The Board also closed a gap at parent level. Because the GENIUS Act bars it from making a bank parent hold capital beyond the issuer’s own requirement, the proposal has the parent deconsolidate the issuer and deduct the issuer’s full minimum from its common equity tier 1 capital, so the same dollar cannot count twice.

What this means for issuers, banks and compliance teams

Prospective bank-affiliated issuers. Model equity at the 1.5 scalar unless you hold auditable pre-approval loss data. At $5 billion that is $150 million of tangible equity, deducted in full from the parent’s common equity tier 1. Start logging loss events above $20,000 now.

State-qualified issuers near $10 billion. Notice is due within five calendar days of crossing, a capital analysis within 270 days and full transition within 360, unless it halts net issuance or wins a waiver. FinCEN’s PPSI rules add parallel anti-money-laundering duties.

Legal and compliance teams. Build the quarter-end capital test into the close with a five-business-day plan template, and write liquidation playbooks around a fixed date. Foreign-issuer questions remain open, as covered in our July report on foreign issuers.

What’s next: the forward view

The comment deadline is 60 days after Federal Register publication, and a Federal Register search on September 28, 2026 found neither R-1899 nor R-1900 published. The GENIUS Act takes effect on the earlier of January 18, 2027, or 120 days after final rules, so the regime may start before a final capital schedule exists.

The NPRM lists the contested parameters: a flat 1% to 2% charge plus a fixed amount (Question 157), revenue rates of 10%, 15% or 30% (Question 159), a 12% to 18% rate for custody revenue (Question 160), graded deposit charges (Question 150) and a two-quarter cure (Question 173). The UK regime starts on October 25, 2027, at a flat 1% and MiCA’s 3% significant-token rate already applies: three regimes pricing size three ways.

TL;DR

The Federal Reserve’s Docket R-1899 (RIN 7100-AH29), proposed on September 24, 2026, prices stablecoin issuance for Board-supervised issuers at 2% of the first $20 billion, 1.5% of the next $30 billion and 1% above $50 billion, plus 25% of non-reserve revenue. New issuers without loss data pay 50% more through a 1.5 loss scalar. Uninsured bank deposits held as reserves draw an extra 2% charge, steering reserves into short Treasury bills. A shortfall that survives two quarter-ends forces liquidation. The NPRM puts the top two issuers at about 58% and 24-25% of supply, and the curve charges smaller rivals up to three times as much per dollar. Comments close 60 days after Federal Register publication, still pending.

FAQ

What is Federal Reserve Docket R-1899?

Docket No. R-1899, RIN 7100-AH29, is the Federal Reserve Board’s proposed rule implementing its GENIUS Act responsibilities, released on September 24, 2026. It covers reserve assets, capital, permissible activities, the yield ban, risk management, custody and the Act’s tying prohibition. Its capital chapter applies to subsidiaries of insured state member banks and to state-qualified uninsured depository institutions above $10 billion outstanding that transition to Board supervision.

Why do smaller stablecoin issuers pay a higher rate?

The Board argues operational losses do not grow in line with issuance because large unauthorised mints are detected faster and are harder to launder, so the tiers fall from 2% to 1%. Separately, an issuer without five years of loss data gets a loss scalar of 1.5. A new $1 billion issuer therefore faces an effective 3% charge, while a large issuer with a clean record could fall below 1%.

What happens if a Board-supervised issuer misses its capital minimum?

Under proposed § 247.15(c), it files a capital plan within five business days of the quarter-end showing a return to compliance within one quarter. If it is still short at the following quarter-end, it must begin liquidating all reserve assets the next day, redeem outstanding stablecoins without fees and stop issuing. The Board has asked whether the cure period should be two quarters.

Does the rule apply to Tether or Circle?

Not directly as proposed. The capital schedule applies only to Board-supervised issuers: subsidiaries of insured state member banks and covered state-qualified issuers that transition after crossing $10 billion. Issuers chartered through the OCC, FDIC, NCUA or states follow their own regulator’s rules. Scale comparisons to USDT or USDC in this article illustrate the formula only and do not describe either firm’s actual obligations.

When is the comment deadline for R-1899 and R-1900?

Both close 60 days after Federal Register publication. As of September 28, 2026, neither had appeared in the Federal Register, so no calendar deadline exists yet and any date quoted elsewhere is an estimate. Comments on the capital rule cite Docket No. R-1899 and RIN 7100-AH29 on the Board’s proposals website; comments on the application rule cite Docket No. R-1900 and RIN 7100-AH30.

This article is informational analysis only and does not constitute legal, regulatory, tax, or investment advice. Regulatory frameworks change frequently and interpretation depends on facts and circumstances; primary documents and official regulator guidance always supersede summaries. Firms should consult qualified legal counsel and their relevant supervisory authority before taking any action based on the analysis above.