I have the proposal text open in another tab. It runs longer than it needs to, the way these things do. Most of the prose is depositor-protection language — bankruptcy-remote treatment, 1:1 backing, segregated custody at a qualified institution, daily attestations. Read it once and you nod. Read it a second time and you start to notice what the text does not say out loud: who can afford to comply, and who cannot.
That second read is what this piece is about. The honest answer to "is this proposal good or bad" is *it depends on who you are*, and the dependency is sharp enough that I want to walk through three composite scenarios — none of them real people, all of them hypothetical illustrations — to show how the same rule lands very differently on three different desks. A startup issuer trying to launch. An incumbent issuer that already dominates the market. A retail trader who just holds the things on a centralized exchange and never thinks about the issuance layer at all. Same rule. Three different outcomes. One pattern underneath.
Scenario 1: The Upstart Issuer Trying to Get to Day One
Imagine a small team — five people, eight months into building — preparing to launch a USD stablecoin. Picture the founder running the math on Day One reserve custody costs. Their plan was to raise a small seed, mint $50 million in initial float against treasuries, and grow from there.
Now they read the proposal. It requires that reserve assets sit with a qualified custodian — operationally, the OCC-chartered or NY DFS Trust Company set. The grounded names in this layer of the market are Coinbase Custody (NY DFS Trust Company), Fidelity Digital Assets (NY DFS Trust), and Anchorage Digital (the first OCC federal trust charter granted to a crypto bank). Three institutions. That is essentially the entire compliant on-ramp at the institutional custody layer.
Walk through what that means for the upstart. They are not the priority client at any of those three. The custody RFP cycle for a new issuer is measured in months, not weeks. The minimum reserve balance to even open an institutional account at these venues is set high enough that a $50M float looks like a courtesy customer. Add daily attestation requirements, segregated-account architecture, the legal opinion needed to satisfy the bankruptcy-remote language, the audit relationship with a firm willing to sign an attestation against on-chain liabilities — and the upstart's pre-launch cost goes from "raise a seed" to "raise a Series A *just to satisfy compliance prerequisites before issuing the first token*."
That is not a regulation. That is a barrier to entry dressed in depositor-protection clothing.
Here is the part that makes me uncomfortable: the depositor-protection rationale is not wrong. After enough collapses where reserves turned out to be commercial paper and undisclosed credit lines, the case for "the reserves must sit with a chartered custodian" is the right case. I will concede that up front. The problem is not the principle. The problem is that the principle, written this way, eliminates the entire competitive layer that would otherwise discipline the incumbents on yield share, redemption speed, and reserve transparency. The upstart is the only force that would have made an incumbent improve. The proposal makes the upstart non-viable.
So on this desk's scoreboard: depositor protection — improved. Market structure — frozen at whoever cleared the bar before the door closed.
Scenario 2: The Incumbent Issuer Reading the Same Text
Now picture a different desk. A general counsel at one of the two or three issuers that already dominate the stablecoin market. Same text in front of them. Same language about qualified custodians, segregated accounts, daily attestations.
Let us say they have been doing most of this voluntarily for the last eighteen months — partly because the market demanded it after enough opaque-issuer blowups, partly because their auditors required it, partly because they were preparing for exactly this regulatory drift. The proposal codifies what their internal compliance team built anyway. The marginal cost of compliance for them is close to zero. The marginal cost for a competitor that does not already have those relationships in place is the cost of standing up the entire stack.
I would not call this a windfall. I would call it a *legible windfall* — they did the work, the regulator wrote the rule that ratifies the work, the competitive set narrows to the players who already cleared the bar. From their seat, the proposal is not a tax. It is a compliance moat with the regulator holding the deed.
Two things follow from this. First, the incumbent's cost of capital improves. Once issuance becomes a permission-gated activity rather than a permissionless one, the perceived risk of the incumbent's float drops — and that shows up in cheaper banking relationships, easier listing on the major venues, lower reserve-yield-sharing demands from exchanges that integrate them. Second, the incumbent gains optionality on which competitors get to exist at all. If the qualified-custodian list is short, and the incumbent has long-standing relationships with all three custodians, the incumbent is informally consulted on which new issuer relationships those custodians take on. That is a soft veto, but it is a veto.
The text never says this. It does not have to. The structure does the work.
Scenario 3: The Retail Holder on a CEX Who Never Reads Proposals
The third persona is the one who does not read FDIC proposals at all. Imagine a retail trader who keeps a working balance in USDT or USDC, parks it on a centralized exchange, and rotates between spot and perps depending on the week. They do not care which qualified custodian holds the issuer's reserves. They care whether the stablecoin pegs, whether withdrawals clear, and whether their funding-rate math still works.
Here is where I want to do the actual math — because the retail position looks neutral until you trace the second-order effects through the venues this trader actually uses.
Start with the volume layer. Binance does $18.5 billion in daily spot-and-derivatives volume across 1,850 listed pairs. Bybit does $9.2 billion across 970 pairs. Bitget does $6.1 billion across 830 pairs. OKX does $4.9 billion across 720 pairs. MEXC does $3.8 billion across 2,400 pairs — the longest-tail venue of the five. Sum the daily volumes: $18.5B + $9.2B + $6.1B + $4.9B + $3.8B = $42.5 billion per day across the five venues a typical retail trader actually uses.
Sum the pair counts: 1,850 + 970 + 830 + 720 + 2,400 = 6,770 pairs. Divide volume by pair count: $42.5B / 6,770 = roughly $6.28 million in average daily volume per pair. That is the average — the top pairs are larger by orders of magnitude, the long-tail pairs are smaller by orders of magnitude — but $6.28 million per pair is the rough denominator.
Now apply the stablecoin pass-through. A conservative read is that at least half of these 6,770 pairs are quoted in a stablecoin — most BTC, ETH, and altcoin pairs settle into USDT or USDC on these venues, not into fiat. Half of 6,770 is 3,385 pairs. Multiply by $6.28 million daily average: $21.25 billion per day in stablecoin-mediated CEX volume across the five venues alone. Annualized at 365 trading days: roughly $7.76 trillion in stablecoin throughput per year just on this slice of the market.
That is the number that matters for the retail holder. $7.76 trillion a year of stablecoin throughput passes through five venues. The proposal does not regulate the venues. It regulates the *issuers* whose tokens those venues quote pairs in. Narrow the issuer set from "any compliant issuer" to "any incumbent already grandfathered into qualified-custodian relationships", and you have funneled the entire $7.76 trillion through whichever two or three names cleared the bar before the door closed.
The retail trader does not pay this as a fee. They pay it as concentration risk. If the qualified-issuer list is two or three names, every stablecoin pair on Binance, Bybit, Bitget, OKX, and MEXC ultimately depends on those two or three issuers continuing to operate without incident. The CER reserve-status field reads "verified" for four of these five venues and "partial" for MEXC, and the proof-of-reserves audits all happened in early 2025 — and the venues themselves are auditable. The issuers behind the tokens those venues quote are now the single point of failure the depositor-protection language was supposed to prevent.
A reserve rule that creates depositor protection at the issuer layer and concentration risk at the market-structure layer has not eliminated systemic risk. It has moved it one floor up.
What All Three Share
I keep coming back to a pattern across the three scenarios. The upstart sees a wall they cannot climb. The incumbent sees a moat they did not have to dig. The retail holder sees nothing — and that is the problem, because the second-order effects on stablecoin issuer concentration land on the retail holder's withdrawal flow, on the venues they trade on, on the funding-rate math that depends on consistent stablecoin liquidity.
What all three share is that the proposal's *stated purpose* and its *structural effect* run in different directions. Stated: protect depositors from a Tether-style or Silvergate-style blowup. Structural: ratify whoever already dominates, eliminate the competitive layer, concentrate the flow, and reframe the resulting concentration as a feature of supervision rather than a bug of the rule.
I do not think this was malicious. Regulators write the rules they can defend, and "qualified custodian with daily attestation" is the easiest rule to defend in a hearing. The harder rule to write — the one that would actually preserve a competitive issuer layer — is the one where compliance costs scale with float size, where small issuers have a graduated bar to clear, where the custodian universe is expanded with a clear charter path rather than restricted to whoever the OCC and NY DFS have already blessed. That rule exists in the EU's MiCA framework in a partial form. It does not exist in this proposal.
So we get the easy rule, written for the easy hearing, with the inadvertent side effect of locking in market structure.
Which Scenario Is You
If you are building a stablecoin issuer right now, you are Scenario 1, and the honest read is that your runway calculation needs to be redone. The compliance prerequisites for launch are now closer to the prerequisites for a chartered bank than for a fintech product. Plan accordingly, or pivot to a stack where the regulatory perimeter does not include issuer registration in the U.S.
If you are a treasurer or risk officer at a counterparty that holds stablecoin reserves, you are Scenario 2-adjacent. The diligence question shifts from *"is this issuer audited"* to *"is this issuer one of the two or three names that will still be issuing under the new rule."* The answer narrows your acceptable issuer list, which narrows your effective counterparty exposure to a handful of names. Plan your concentration limits with that in mind.
If you are a trader who keeps working capital in USDT or USDC on a CEX, you are Scenario 3, and the action item is the boring one: think about which two issuers your effective stablecoin balance ultimately depends on, and decide whether you are comfortable holding that concentration when the regulatory framework has, in effect, made the concentration permanent.
FAQ
Does the FDIC proposal actually ban small stablecoin issuers from operating?
Not directly — the text does not contain a prohibition. The effect is indirect, through the qualified-custodian requirement and the daily-attestation language. A small issuer is technically allowed to operate, but the operational stack required to comply costs more than the float of a small issuer can support. The ban is economic, not legal. That distinction matters in litigation but not in practice.
Why does qualified-custodian language favor incumbents specifically?
Because the universe of qualified custodians at this layer is genuinely small — Coinbase Custody, Fidelity Digital Assets, and Anchorage Digital are the names with the right charters. Each of them has a finite number of institutional client slots, prioritized by float size and existing relationship. An incumbent issuer is already a client. A new issuer joins a queue with no guarantee of acceptance. The bottleneck is the custodian's RFP cycle, not the regulator's approval process.
Could a startup issuer launch outside the U.S. instead?
Yes, and many will. The MiCA framework in the EU has a different compliance structure with graduated reserve and capital requirements based on float size. Singapore's MAS framework is also more startup-tolerant on the registration side. The cost of going offshore is loss of U.S. market access for redemptions and CEX listing in venues that prioritize U.S.-registered issuers — which is a real cost, just smaller than the U.S. compliance bill.
Does the proposal address Tether-style reserve composition risk?
Indirectly. The 1:1 backing language and the qualified-custodian language combined would prevent the commercial-paper-heavy reserve structures that were common in 2021-2022. That part of the proposal is genuinely improving the risk profile of compliant issuers. The criticism is not that the proposal does not improve safety — it does. The criticism is that the safety improvement comes packaged with a market-structure consolidation that was not necessary to achieve it.
How exposed is the average CEX user to issuer concentration risk?
Quite exposed, in ways most users do not see. Five major venues (Binance, Bybit, Bitget, OKX, MEXC) process roughly $42.5 billion per day in volume, and the majority of pairs settle into stablecoins rather than fiat. If the proposal narrows the issuer universe to two or three names, every working balance held on those venues depends on those two or three issuers remaining solvent and compliant. The diversification is illusory — the underlying issuer concentration is the binding constraint.
Are self-custody users affected by this proposal?
Less directly, but the second-order effects reach them. Self-custody users still rely on stablecoins to enter and exit positions, and the stablecoins they use are issued by the same handful of compliant names. Hardware-wallet users — whether on Ledger, Trezor, or GridPlus Lattice1 — are insulated from venue risk but not from issuer risk. The reserve composition behind the USDC or USDT in their cold-storage address is the same reserve composition the proposal regulates.
Is there any read of this proposal where it is not a moat?
There is a read where the moat is incidental rather than deliberate, and the regulator is solving for "next blowup prevention" without modeling competitive structure at all. That read is plausible. It does not change the structural outcome. Whether the moat was intended or accidental, the effect on the issuer universe is the same — and the question that nobody in the public comment record has answered with data yet is whether the regulator was offered an alternative drafting that preserved depositor protection without concentrating issuance, and what happened to that alternative. If you have seen that drafting in the comment file, write.