Two percent, on the first twenty billion. That's the number that should be flashing on every stablecoin issuer's terminal this week โ not the price of Bitcoin, not the ETF inflows, not whatever the perpetual funding rate is doing at 3 a.m. in Istanbul.
The Federal Reserve, working under the GENIUS Act, published a proposal that does something deceptively simple: it takes the balance-sheet discipline of a commercial bank and bolts it onto the issuance of dollar tokens. 100% reserves. Tiered capital. A two-business-day redemption promise. And โ this is the part the market is still mispricing โ a presumption that issuing yield through affiliates or relevant third parties is a prohibited payment of interest.
I've spent the last three months building reserve-liquidity models for cross-border settlement desks. My first read of the tier structure produced a file I did not expect to write: a cash-flow statement for a stablecoin issuer that looks less like a crypto company and more like a narrow bank with a marketing department.
Let me set the plumbing before the plumbing sets the narrative.
The GENIUS Act is the first comprehensive U.S. federal framework for payment stablecoins. It designates who supervises whom. Insured state member bank subsidiaries go through a Federal Reserve application. Uninsured state-chartered issuers holding more than $10 billion in outstanding tokens fall under direct Fed supervision. Smaller players sit in a patchwork. The compliance date lands on January 18, 2027, or 120 days after the final rule is published โ whichever comes first.
This proposal does not stand alone. The OCC runs a parallel track, and the two are consistent on the yield presumption. That consistency matters more than any individual clause. When two regulators independently converge on the same interpretation, you are not looking at a draft. You are looking at the future.
Sixty days of comment. That is the entire window in which the yield business gets to argue for its life. Governance grandee Michael Barr, notably, has already lodged a dissent โ not on reserves, not on capital, but on anti-money-laundering enforcement, arguing the Fed should act only when an issuer's AML failure is "material or systemic." Read that carefully. The internal fight is not about whether to regulate. It is about how much discretion the Fed keeps for itself.
Start with the reserve whitelist, because that is where the real architecture lives. Eligible reserves: cash, Federal Reserve balances, demand deposits, Treasuries with maturities of ninety-three days or less, overnight repos and reverse repos, funds that invest solely in those instruments โ and tokenized versions of the same.
Read that last item again. Tokenized versions of the same.
For three years I've argued the RWA narrative was a solution hunting for a problem, that most tokenized treasury products were collateral plumbing dressed up as innovation. I was half right. The problem just arrived, and it has a compliance department. Tokenized money-market exposure โ BUIDL, OUSG, and whatever the next entrant ships โ now has a legally defined lane into the reserve stack of every regulated dollar token in the United States. That is not a bull thesis on RWA. That is a structural demand curve being legislated into existence.
Now the capital tiers, which is where the math gets interesting. First $20 billion of outstanding tokens: 2% capital charge. The next $30 billion: 1.5%. Above $50 billion: 1%. Scale reduces the rate โ classic marginal-risk logic, and a quiet endorsement of concentration.
Run the numbers. A $20 billion issuer holds $400 million in high-quality liquid assets against its float. At a 4% Treasury yield, that is roughly $16 million a year in foregone interest โ the opportunity cost of capital that cannot be lent, invested, or deployed. For a mid-cap issuer, that is not a rounding error. For Circle, it is a line item.
Then there's the second capital requirement โ 25% of the three-year average of non-reserve revenue. Follow the logic. The Fed is not merely capping the reserve business. It is pricing the non-reserve business too. Lending, advisory, any side venture that generates revenue outside the reserve wrapper gets taxed at a quarter of its trailing earnings, held as capital. The message is surgical: you cannot arbitrage your way out of the reserve box by building a shadow finance arm and calling it a product.
Redemption sits at two business days, with safe-harbor flexibility the proposal does not fully define. USDC runs closer to T+1. USDT sits nearer T+2. So the mandate mostly ratifies current practice โ which is exactly why it is dangerous. It freezes today's norm as tomorrow's floor, and floors are only ever tested downward under stress.
And then the yield presumption itself. Issuer pays an affiliate or a "relevant third party." That party pays the holder. Presumed prohibited. The covered universe explicitly includes service providers and white-label partners. Rebuttal is possible, but only in writing, and the burden lands on the issuer.
I've been tracing liquidity ghosts through the ICO fog long enough to recognize what this is. It is not a ban. It is a cost-of-defense mechanism. The Fed is forcing every yield-bearing stablecoin model to prove its innocence, one filing at a time, while the legal meter runs.
Here is where the consensus read gets lazy.
The street is telling itself a clean story: the Fed killed yield-bearing stablecoins, DeFi loses a leg, compliance wins, move on. I don't buy it. Not because the presumption is weak โ it is brutal โ but because of where the yield actually migrates.
Look at the covered chain again. The prohibition targets payments routed from the issuer through an intermediary to the holder. It does not obviously reach yield generated natively inside a protocol, distributed transparently on-chain, funded by borrower demand rather than by the issuer's reserve income. That distinction is the whole ballgame, and almost nobody is pricing it.
Watch what happens over the next four quarters. Third-party distribution models โ the white-label and brand-partner structures โ die or restructure. Protocol-native models, where the yield emerges from lending markets and the issuer never touches the payment, inherit a competitive moat built by federal rulemaking. The Fed thinks it is enclosing the yield business. It may be handing the most durable yield architectures an exemption by accident, and pushing risk into exactly the venue it cannot audit.
There is a second blind spot, and it is the one that keeps me up. The supervision perimeter is selective. Uninsured issuers under $10 billion โ outside the direct line. Offshore issuers โ outside entirely. Tether keeps its offshore dominance and its higher risk appetite. What you get is not one stablecoin market. You get three: a compliant U.S. dollar rail, a small-issuer gray zone, and an offshore dollar system that answers to nobody in Washington.
Add the Clarity Act, which stalled in the Senate this month precisely over stablecoin rewards, with banks lobbying hard to restrict them. Yield-bearing stablecoins are now caught between a rule that presumes them illegal and a statute that has not decided whether to permit them. That is not ambiguity. That is a pricing gap.
Capital charges favor scale-advantaged incumbents. A whitelist blesses tokenized treasuries. A yield rule pushes risk into the one jurisdiction regulators can't see. This is liquidity geography being redrawn, and the map favors whoever already owns the coastline.
January 18, 2027 is a date, not a forecast. The sixty-day comment window is the real event โ it decides whether the yield presumption narrows or hardens, whether "relevant third party" gets defined precisely or left as a weapon, and whether the $10 billion supervision threshold becomes a ceiling small issuers engineer themselves under.
My question for the next cycle: when the compliant dollar becomes a utility with a capital charge and a redemption clock, who captures the spread on the world's dollar demand โ the chartered issuer, the tokenized treasury fund sitting inside its reserves, or the offshore token that never asked permission? The plumbing is being built right now. Watch it, not the price.