Two percent of the first twenty billion dollars. That is the number the market grabbed. It is not the number that matters.
Buried in the Federal Reserve's stablecoin proposal — issued under the authority of the GENIUS Act — sits a second capital charge that has almost nothing to do with reserves: a requirement equal to 25% of an issuer's three-year average revenue from activities outside its reserve portfolio. Read it twice. The Fed is not pricing the risk of the asset backing the token. It is pricing the risk of the business wrapped around the token.
I spent four weeks in 2020 auditing Uniswap v2's initial contracts by hand, and the lesson that survived the process was not about invariant math. It was that the exploitable vector is almost never the component everyone is staring at. It is the adjacent function with sloppier assumptions. In this proposal, the reserve whitelist is the audited contract. The 25% charge is the adjacent function. And the yield presumption is the reentrancy.
The GENIUS Act gave the United States its first comprehensive federal framework for payment stablecoins. It also handed the Federal Reserve a mandate it had never held: supervisory authority over a class of dollar-denominated instruments that settle on public ledgers. The proposal that followed — now in a 60-day comment window — fills in the blanks the legislation left open. Reserve composition. Capital. Redemption. And, quietly, the boundaries of what an issuer is allowed to pay its holders.
The scope is narrower than the headline suggests. The Fed's direct reach covers insured state member bank subsidiaries that issue stablecoins, plus uninsured state-licensed issuers once outstanding balances cross the $10 billion line. Compliance dates land on January 18, 2027, or 120 days after a final rule — whichever arrives first. The OCC has proposed parallel language, and the two documents track each other closely enough on the yield question that the alignment appears coordinated rather than coincidental.
Note the timing. The proposal arrives while the Clarity Act — the market-structure companion that would settle whether stablecoin rewards are a security question or a banking question — remains stuck in the Senate, with banks lobbying to restrict rewards and the industry lobbying to preserve them. The Fed has effectively moved first. When a regulator pre-commits ahead of a legislature, the legislative debate stops being a debate and becomes an argument about ratifying a fait accompli.
There is a dissent worth logging. Governor Michael Barr endorsed the reserve and capital architecture but objected to the anti-money-laundering standard, which would let the Fed intervene only when an issuer's AML failures are material or systemic. That is a threshold, not a rule. And thresholds are where supervisory discretion lives — which is precisely why the dissent exists.
This is the third time in five years that I have watched a regulatory document get read primarily through its price implications. In 2024, ahead of the spot Ethereum ETF decisions, I led a five-scenario simulation for my firm and we modeled the approval window at roughly 60% by Q3. The number mattered less than the structural insight: regulatory clarity is a narrative driver, not a valuation driver. It changes who is allowed to compete. It does not change what the asset does.
Start with the reserve whitelist, because it is the cleanest part of the document.
Eligible backing is confined to cash, balances held at the Federal Reserve, demand deposits, Treasury securities with remaining maturity of 93 days or less, overnight repurchase and reverse repurchase agreements, and funds that invest exclusively in those instruments — including tokenized versions of them. Corporate bonds, commercial paper, mortgage-backed securities: excluded.
That list is not novel. It is the post-Terra consensus on high-quality liquid assets, written into federal text. When UST broke in May 2022, I built a 40-slide deck in Istanbul tracing the depeg mechanics and the contagion path into Anchor deposits, and I delivered it three days before the major outlets picked up the story. The structural lesson from that episode was that an algorithmic stablecoin's failure mode is not the algorithm — it is the assumption that a risk asset can be sold into a falling market at par. The Fed's whitelist is a formal rejection of that assumption. Nothing on the list needs a buyer.
The interesting inclusion is the last one. Tokenized Treasury funds — the on-chain vehicles holding short-dated government paper — now have a plausible path into the reserve stack of a federally supervised issuer. That is not a small concession. The compliance infrastructure for tokenized government debt and the collateral infrastructure for stablecoins are converging on the same rails.
Now the capital tiers. The first $20 billion of outstanding stablecoins carries a 2% charge. From $20 billion to $30 billion, 1.5%. Above $50 billion, 1%.
The descending schedule encodes a familiar regulatory belief — that scale reduces systemic risk through diversification and operational maturity. It also encodes something less flattering. The tiered structure is a subsidy for incumbents. An issuer at $60 billion faces a materially lower blended capital burden than an issuer at $15 billion, and the gap compounds with every dollar of growth. If you wanted to design a rule that accelerated consolidation in a market with three meaningful participants, this is roughly how you would do it.
Circle's compliance posture, its bank relationships, and its pending public listing all become assets under this schedule. Tether, which sits outside the Fed's direct perimeter, faces a different problem: the cost of not being inside it. And PYUSD, issued through a bank-subsidiary path, gets the clearest runway of the three.
The architecture is worth naming precisely, because it is not a single regime. Insured state member bank subsidiaries sit under direct Federal Reserve supervision — examination, reporting, capital adequacy review. Uninsured state-licensed issuers above the threshold sit under a rules-based regime, constrained by capital and reserve requirements rather than continuous examination. Offshore issuers sit outside the perimeter entirely. Three layers, three cost structures, one dollar. That is not regulatory clarity. That is a market segmentation scheme with a dollar sign attached to it.
Then the 25% charge, which is where the document stops being about stablecoins.
The second capital requirement is calculated against three-year average revenue from activities outside the reserve portfolio. The design logic becomes transparent once you name it: the Fed is preventing issuers from escaping capital discipline by routing profitability through non-reserve business lines. Lending. Advisory. Payment services. Anything that generates income but does not sit inside the reserve wrapper.
This is a balance-sheet constraint on a business model, not a reserve constraint on a token. It tells you what the Fed believes an issuer should be: a narrow, well-capitalized, low-margin utility that holds Treasuries and mints dollars. Not a fintech with a yield product bolted to the side.
There is a second-order effect that deserves more attention than it is getting. Capital requirements only bind if the capital is actually held. A 2% charge on $20 billion is $400 million in high-quality liquid assets sitting on a balance sheet doing nothing. At a 4% Treasury yield, the opportunity cost of that idle capital runs near $80 million annually before the second charge is applied. Scale that across the industry and the capital stack stops being a compliance line item and becomes a competitive moat. Auditing the hype for structural integrity means asking who can afford to hold capital that earns nothing. The answer is: the largest issuer, the bank subsidiary, and the incumbent already sitting on a Treasury portfolio.
Redemption gets two business days. That is roughly where USDC and USDT already operate, so the requirement is codification rather than reform. The safety-harbor language provides flexibility, but the proposal does not spell out the trigger conditions — which means the operative rule is currently unwritten.
And then the yield presumption, the provision that will generate the most comment letters.
The proposal presumes a prohibited interest payment has occurred when an issuer pays an affiliate or a related third party, and that party then pays the holder. The framing covers service providers and white-label partners explicitly. A rebuttal is available in writing. But the burden sits with the issuer, and the definition of related third party is broad enough that a compliance officer's rational response is to shut down anything resembling a payment chain.
Last year I worked with two core developers from Polygon on zero-knowledge proof circuits, trimming verification costs by roughly 15%. The exercise taught me something that transfers directly here: verification is always cheaper than assumption, and regulators know it. The rebuttable presumption is a verification requirement dressed as a prohibition. The Fed is not saying the payment is illegal. It is saying the issuer must prove it is not. That distinction is where the legal costs live.
Here is the sentiment-reality gap. Scan the timeline and the read is uniform: clarity is bullish, regulation unlocks institutions, stablecoins finally get a rulebook. The volume is real. The conclusion is not. The proposal does not ban yield. It bans the intermediary through which yield is delivered — a different instrument with a different blast radius. Protocol-native distribution, where the reward accrues on-chain with no counterparty in the loop, sits in a materially different position than a branded partnership where a partner pays the holder. One is a smart contract. The other is a payment the Fed has already labeled in advance.
That distinction is not a technicality. It is the entire competitive landscape of the next eighteen months.
The consensus interpretation is that the Fed is being cautious about stablecoin risk. I think the more accurate reading is that the Fed is protecting the deposit franchise.
Consider the direction of travel across the whole document. Reserves are narrowed to instruments banks already hold. Capital is calibrated to bank-like ratios. Redemption is aligned to bank-like settlement windows. Non-reserve revenue is penalized. Yield distributed through third parties is presumed illegal. Each provision is independently defensible. Together they describe an instrument that can move dollars but cannot pay for the privilege of holding them.
That is not a stablecoin framework. That is a demand-deposit framework, applied to a token.
The consequence is that the profit pool migrates. If an issuer cannot earn on spread differentials, cannot pay for distribution, and faces a hard capital charge on non-reserve income, the rational issuer becomes a payment company. Fees on transfer. Fees on settlement. Enterprise contracts. The stablecoin stops being a product and becomes a piece of infrastructure — and the margin moves to whoever controls the last mile: exchanges, wallets, custodians, and the payment networks sitting between the token and the merchant.
I have watched this pattern before. In 2023, while most of my colleagues were writing about AI narrative rotation, I was pulling API call data from early agent marketplaces and interviewing founders directly rather than through their PR channels. The number that mattered was not the token price. It was the 300% growth in API calls — usage arriving before valuation. We hunt the signal in the noise of consensus, and the signal here is not the capital ratio. It is where the revenue line relocates.
There is a structural hole worth naming. The Fed's direct perimeter starts at $10 billion in outstanding balances for uninsured state-licensed issuers. Below that line, supervision is lighter. That is a cliff, and cliffs create behavior. Small issuers gain an incentive to stay small. Large issuers gain an incentive to get larger. The middle thins out. Collateral damage is a feature, not a bug — the schedule is designed to produce a market of a few very large, very plain, very supervised issuers.
Meanwhile the Clarity Act sits stalled, with stablecoin rewards as one of its central disputes. Banks pushed to restrict them. The Fed has now effectively pre-committed to a position while the legislature has not. That asymmetry matters more than any single clause in either text.
Watch three things over the comment window and the eighteen months that follow.
First, whether the final rule narrows the definition of related third party. If it stays broad, expect a wave of preemptive product shutdowns that never surface in any enforcement action — compliance by attrition, invisible in the data.
Second, where the tokenized Treasury funds land. If on-chain government paper becomes standard reserve collateral under federal supervision, the RWA narrative stops being a narrative and becomes a plumbing requirement.
Third, whether issuers begin reporting non-reserve revenue as a line item at all — because the moment that number becomes a capital input, it becomes a disclosure, and disclosure is where narratives go to die.
The tether does not snap when the price drops. It snaps when the structure holding the price in place gets rewritten. That is what happened here. The market read the first twenty billion dollars. The document was about everything that comes after.
