Mechanisms, authority, and structure, not one word
A journalist writes that an exchange "froze" a user's stablecoins. A regulator's press release says an issuer "froze" a wallet linked to a fraud network. A different report says a court "seized" the same class of asset. A vendor pitches a token as "unfreezable." All four sentences use freeze-adjacent language, and collapsing them into one word hides a real distinction: some of them describe a technical mechanism (an actual action taken on a token or account), some describe an authority basis (the legal or contractual trigger that causes a mechanism to be invoked), and one describes a structural factor (a limit on whether any mechanism reaches the asset at all, regardless of who's trying). This guide exists because flattening those three categories into "freezing" causes real practitioner mistakes: treating a legal compulsion as if it were itself a mechanically distinct action separate from the mechanisms that carry it out, assuming a burn function can destroy funds sitting at any blacklisted address rather than only tokens a minter already holds, assuming an exchange-level account freeze means the token itself is blocked on-chain, or assuming "unfreezable" marketing language describes an actual legal or technical impossibility rather than a design choice an issuer can reverse.
The approach here separates six technical mechanisms, three authority bases that trigger them, and one structural factor that can defeat any of them, kept in those three tiers throughout rather than forced into one flat list. The central worked example is a real, documented contrast: in July 2024 a US stablecoin issuer froze more than $29 million connected to a Cambodia-based marketplace group called Huione, and two months later that same group launched its own stablecoin, USDH, explicitly marketed as incapable of exactly that kind of freeze. Both halves of that contrast are verified directly against a US Treasury regulatory filing, not a secondary summary of it, and the guide uses it throughout to make the taxonomy concrete rather than abstract.
Work through the three tables below, mechanisms, authority bases, and the structural factor, before describing any real-world event as a "freeze." Then check the Huione/USDH case study and the issuer-terms comparison before assuming any two stablecoin issuers have the same freeze authority, they don't necessarily, and this guide checked Circle's and Tether's own terms and Circle's own Access Denial Policy directly rather than assuming parity or taking a general Terms clause as the whole picture.
Six mechanisms, three authority bases, one structural factor
It's worth stating the shape of this section before the tables, because a flat list of ten items would itself mislead: some of the ten are technical actions, some are legal or contractual triggers that cause an action to be taken, and one is a structural limit on whether any action reaches the asset at all. Treating a trigger as though it were a mechanically distinct action, or vice versa, is exactly the kind of imprecision this guide is built to avoid, so the three categories are kept separate throughout rather than forced into one list.
Mechanisms: the technical or operational actions
Each of these is something that actually happens to a token, account, or asset. Any of them can be invoked by more than one authority basis below, that's the point of separating the two tables.
| # | Mechanism | Who acts | Basis | Not to be confused with |
|---|---|---|---|---|
| M1 | Address denylisting | The issuer, via a smart-contract role built into the token | Circle's own FiatToken "Blacklister" role, address-specific by design, per Circle's own token design documentation: it "prevent[s] all transfers to or from a particular address, and prevents that address from minting or burning" | Stops that address from acting; it does not, by itself, destroy the tokens sitting there (M2) |
| M2 | Token burning | A minter, on tokens the minter itself already holds | Per Circle's own documentation, burn is a minter-only function: "the minter must have a balance greater than or equal to the amount... A minter can only burn tokens which it owns," typically the terminal step of a redemption flow, not an enforcement action against a third party's address | Denylisting an address (M1) stops that address from burning its own tokens; it does not hand the issuer a separate function to destroy tokens sitting at that address. This guide has not found a named, verified source for any issuer having a general capability to destroy funds at an arbitrary third-party address, and does not assume one exists |
| M3 | Redemption refusal | The issuer, at its own fiat off-ramp specifically | A commercial-channel decision, not an on-chain transaction | Affects converting the token to fiat, not whether the token can move on-chain (M1) |
| M4 | Custodian freeze | A third-party custodian holding the asset for a client | The custodian's own client agreement, independent of the issuer entirely | The issuer may take no action at all while this happens |
| M5 | Exchange account freeze | A centralised exchange, at the account level | The exchange's own user agreement, independent of the issuer entirely | Doesn't touch the token contract; the on-chain asset itself may remain fully mobile (M1 not triggered) |
| M6 | Smart contract restriction (pause) | The issuer, via a contract-wide circuit-breaker role | Circle's own FiatToken "Pauser" role, which halts all transfers, minting, and burning for every holder at once, per Circle's own documentation | Not address-specific, unlike M1; a single pause event affects the entire token, not one wallet |
Authority bases: what triggers a mechanism, and under what logic
None of these is itself an on-chain or account-level action. Each is a reason a mechanism above gets invoked, and the same mechanism, most often address denylisting, can be triggered by any of the three.
| # | Authority basis | Who decides | Basis | Not to be confused with |
|---|---|---|---|---|
| A1 | Issuer discretion | The issuer, exercising its own contractual judgment | A decision point, not a mechanism in itself; it results in one or more mechanisms above being invoked. See the issuer-terms comparison below for how differently Circle's and Tether's own terms actually define this discretion | The decision to act (A1) versus the specific mechanism that carries it out (M1-M6) |
| A2 | Lawful-order compulsion | A court or authorised government agency, compelling the issuer or custodian to act | A "lawful order," a term the GENIUS Act now defines in US federal statute as covering an order to "seize, freeze, burn, or prevent the transfer" of stablecoins, four different possible outcomes under one legal trigger, meeting stated particularity and review conditions | Not itself a mechanically distinct tenth action parallel to M1-M6; it's satisfied by executing one of those mechanisms. Don't describe it as "seizure" alone, that's one of four outcomes it can compel, not a synonym for the whole category |
| A3 | Sanctions designation | The issuer, exchange, or custodian, compelled by sanctions law | A compliance obligation triggered by a designation (e.g. an OFAC listing), not a discretionary judgment call | Executed through M1, M3, M4, or M5, not a parallel mechanism of its own; the legal trigger is compulsory where A1 is discretionary |
Structural factor: a limit, not an action or an authority
| # | Factor | Nature | Basis | Not to be confused with |
|---|---|---|---|---|
| S1 | Cross-chain complication | No one decides this; it's a structural limitation on reach | Cross-referenced from FinCrimeRadar's own token-topology work6, already verified there against FATF's cross-chain interoperability finding and Circle's own bridged-token documentation | Any mechanism above (M1-M6), triggered by any authority basis above (A1-A3), may simply not reach a bridged or wrapped copy of the same token on another chain |
A few distinctions in these tables cause most of the real-world confusion, worth stating in prose. Issuer discretion (A1) is a decision, not an action; address denylisting (M1) is the specific technical instruction that most often carries a freeze decision out, though in principle a decision could be executed a different way. Denylisting (M1) and a contract-wide pause (M6) are both smart-contract-level actions built into the same token, but one targets a single address and the other stops the token for every holder simultaneously, confusing the two badly overstates or understates the scope of what actually happened. An exchange account freeze (M5) and a custodian freeze (M4) are both third-party actions independent of anything the issuer does; a user can be locked out of an exchange account while the underlying token remains fully transferable on-chain, and vice versa. And reversibility needs its own caveat: address denylisting (M1) is technically reversible by the same issuer that imposed it, but technically reversible is not the same as legally reversible, where a lawful order (A2) compelled the freeze in the first place, the issuer cannot simply reverse it on its own initiative while that order remains in force; reversal in that case depends on the compelling authority, not the issuer's own discretion.
Huione and USDH: a freeze, then a pitch to make freezing impossible
The clearest documented example of issuer discretion (A1) resulting in a freeze sits inside a US Treasury Financial Crimes Enforcement Network (FinCEN) rulemaking, not a press release or a blog post. FinCEN's own text states: "in July 2024, when one stablecoin issuer froze CVC valued at over USD 29 million that was located in a Huione Group CVC wallet because it was 'associated with activities allegedly linked to fraudulent and transnational criminal operations.'" FinCEN's own notice refers to "one stablecoin issuer" without naming it in the body text; the notice's own footnote cites contemporaneous reporting, The Record, 15 July 2024, "Tether freezes $29 million of cryptocurrency connected to Cambodian marketplace accused of fueling scams", identifying the issuer as Tether. That's an important distinction to state precisely: this is FinCEN's own citation chain naming Tether via a named press source, not this guide importing an untethered secondary claim on its own initiative. FinCEN's own body text still doesn't name Tether directly, so treat the identification as resting on FinCEN's own footnoted source, not on FinCEN's own prose. This isn't a one-off in a superseded draft either: FinCEN's own final rule on Huione, issued 15 October 2025, was checked directly against its own text and repeats the identical pattern at its own footnote 98, the same anonymised "one stablecoin issuer" body text, the same citation to the same 15 July 2024 Record article naming Tether. The identification and the $29 million figure carry through unchanged from the proposed rule to the final rule.
Two months later, per the same FinCEN filing, Huione Group launched its own answer to that freeze: "In September 2024, Huione Group launched USDH, a stablecoin it explicitly advertised as 'unfreezable' and 'not restricted by traditional regulatory agencies.'" That "unfreezable" language is FinCEN's own primary-source restatement of Huione's own marketing claim, not FinCEN's editorialising, keep the two separate. FinCEN's own characterisation goes further, in its own voice rather than quoting Huione: it states that Huione Group, "by offering USDH, which is 'unfreezable,' even upon a lawful request from law enforcement," facilitates money laundering benefiting transnational criminal organisations and DPRK actors. That sentence is FinCEN's own regulatory conclusion, an application of its authority to the facts, distinct from both the underlying Huione marketing claim and from the earlier freeze event itself.
Read against the tables above, USDH is best understood as a stablecoin explicitly designed to decline issuer discretion (A1) over its own token, and specifically to forgo building in an address-denylisting mechanism (M1), as a marketed feature. Whatever technical claims underlie "unfreezable" haven't been independently verified in this guide, and shouldn't be taken at face value merely because they're widely reported; what is verified is that FinCEN's own text treats the marketing claim as real enough to build a regulatory finding on, in both the proposed and now the confirmed final rule.
US federal law now defines a "lawful order" to include an order to "seize, freeze, burn, or prevent the transfer" of a stablecoin, four different possible outcomes under a single legal trigger, not a single action synonymous with "seizure." It separately requires that "a permitted payment stablecoin issuer may issue payment stablecoins only if the issuer has the technological capability to comply, and will comply, with the terms of any lawful order," per the GENIUS Act, enacted 18 July 2025. That is a real, verified, statutory authority basis (A2, lawful-order compulsion), executed through whichever of mechanisms M1-M6 fits the order, not a parallel mechanically distinct mechanism of its own, and distinct from any issuer's own contractual freeze clause (A1). It is not, however, currently the operative legal standard: the Act's own effective-date clause sets its operative date at the earlier of 18 months after enactment, roughly 18 January 2027, or 120 days after final implementing regulations are issued. A Federal Register entry titled "GENIUS Act Regulations on Payment Stablecoin Issuance, Offer, and Sale" appeared on 18 August 2026; checked directly against its own text, it's headed "Proposed Rules," a notice of proposed rulemaking, not a final rule, so it does not by itself start the shorter 120-day clock. Treat "USDH would violate the GENIUS Act" as premature until a final rule issues or the 18-month date passes, and treat "the GENIUS Act imposes no freeze-capability requirement" as equally wrong, both statements collapse an enacted-but-not-yet-operative statute into something simpler than it is.
Comparing issuer terms directly, not assuming parity
Circle's own live USDC Terms of Use carry two clauses worth separating, not blending. A general "Blocked Addresses & Forfeited Funds" clause states Circle "reserves the right to 'block' certain USDC addresses and, if such addresses are Circle custodied addresses, freeze associated USDC (temporarily or permanently) that it determines, in its sole discretion, may be associated with illegal activity or activity that otherwise violates these Terms." Read on its own, that clause sounds like open-ended discretion. A separate, specifically named "Blocklisting" clause immediately follows it: "USDC is issued and redeemed in accordance with Circle's blocklisting policy. Circle reserves the right to block the transfer of USDC to and from an address on chain as permitted under the blocklisting policy," and that clause hyperlinks, in Circle's own live Terms, directly to Circle's Stablecoin Access Denial Policy, a separate document this guide fetched and checked directly rather than assuming its contents from the general clause alone. That Policy's own text is much narrower than the general clause reads on its own: "Circle will not deny access to individual addresses, other than in circumstances that strictly conform to the requirements set forth under Part 3, Policy Exceptions," and Part 3 lists exactly three: a determination, in Circle's sole discretion, that a threat to the security, integrity, or reliability of the Circle Stablecoin network exists; a law, regulation, or legal order from a recognised US or French authority or court; or certain urgent law-enforcement, sanctions-related, national-security, or Mutual Legal Assistance Treaty-process requests pending a court order. Reversal isn't unconditional either: the Policy states Circle may lift an access denial "upon formal confirmation" from the relevant authority that the underlying obligation no longer applies, or that a security incident is resolved, not simply whenever Circle itself chooses to. This guide cannot fully resolve from the Terms' text alone whether the broader "Blocked Addresses" clause is itself constrained by the same three-exception Policy or operates as a separate, wider basis; state both clauses precisely rather than assuming either answer. Separately, on a distinct legal basis from either clause above, Circle's Terms state "Circle may also be required to freeze USDC and/or surrender associated USD held in Segregated Accounts in the event it receives a legal order from a valid government authority requiring it to do so." One more scope point worth stating plainly: Circle's own Terms open with the note that "these USDC Terms... only apply to holders of USDC located outside of the European Economic Area," EEA holders are directed to a separate USDC White Paper; this guide has not checked that separate document.
Tether's own live USDT Terms of Service, checked directly rather than assumed to mirror Circle's, use noticeably broader language, and "Prohibited Use" is worth naming precisely rather than shorthanding as a sanctions clause: it's a single defined term covering a long, heterogeneous list, money laundering and sanctions evasion sit alongside tax evasion, using illegally obtained funds, providing false registration information, operating multiple accounts or "throwaway" wallets, circumventing jurisdiction restrictions with a VPN, and more, all bundled under one "Prohibited Uses" definition. That clause states that where Tether "determines or suspects" a Prohibited Use, it "may address such Prohibited Use through an appropriate sanction, in its sole and absolute discretion," and that such a sanction "may include... freezing or confiscation of any Fiat, funds, property, proceeds, Tether Tokens or any Digital Tokens in any Tether Tokens Wallet or other User Wallet; blacklisting any Digital Tokens Address which holds Tether Tokens." A separate general clause states Tether "may suspend or terminate your access to the Site or any of the Services, freeze any Tether Tokens held by you, or terminate your Tether Token Wallet, as required by applicable Law or where Tether, in its sole discretion, determines it is prudent to do so." A further clause states Tether "may, at its sole and absolute discretion, seize and deliver your property to any applicable Government, law enforcement, or other authorities." One scope point applies here too: Tether's own Terms state they "apply to users of tether.to and app.tether.to," and that a user agrees to them "by purchasing or redeeming a Tether Token through the Site or transferring Tether Tokens into or from a Tether Tokens Wallet." That's a contractual scope tied to direct users of Tether's own platform and Services, not, by itself, proof of contractual authority over every secondary-market holder who has never touched tether.to directly; the on-chain blacklisting mechanism itself, being enforced by the token contract's own code rather than by contract law, is a separate question from who has agreed to these Terms, and operates on whichever address the blacklister role targets regardless of contractual privity.
The practitioner point isn't that one issuer is stricter than the other in practice, this guide hasn't independently verified freezing frequency or operational behaviour for either issuer, only their published terms and policies. The point is textual and structural: Tether's broad discretionary language sits in one place, a wide "Prohibited Uses" definition plus a standalone "prudent" clause, both scoped contractually to defined users of its own platform. Circle's discretionary language is split across a broad general clause and a specifically named, materially narrower policy document that its own Terms link to directly, with an EEA carve-out on top. Don't read either as simply "broader" or "narrower" than the other in the abstract, the comparison depends on which specific clause is actually in play. Don't assume any third issuer's terms fall somewhere on this spectrum without checking them directly.
Reporting exists on a New York bankruptcy court ordering Circle to maintain a blacklist tied to the Multichain Foundation liquidation (The Block, 31 October 2025); on a disputed roughly $12.6 million USDC freeze connected to a DeFi protocol called Zama (CCN, June 2026); and on public statements by Circle's CEO contrasting Circle's and Tether's freezing patterns (reported April 2026). None of these three have been checked against their own primary documents or court filings in this guide, they're noted here as illustrative of live practice, not cited with the same confidence as the FinCEN filing, either issuer's own terms, or Circle's own Access Denial Policy above.
A customer's exchange account is locked after a fraud complaint, and the exchange's notice says their "USDT has been frozen." You're drafting a case note and need to state precisely what happened.
This isn't the failure and resolution question
Every mechanism and authority basis in the taxonomy above is a targeted intervention against a specific asset, address, or account, because of suspected illicit conduct or a legal compulsion tied to it. That's a different question from what happens when a stablecoin issuer itself can't continue, insolvency, a capital breach, an operational collapse. FinCrimeRadar's own guide on systemic stablecoins covers that failure-and-resolution question directly, including the Bank of England's draft recovery and wind-down Code of Practice and its "normal and stress times" no-suspension expectation for redemption, and this guide doesn't duplicate that material. Read "What happens when the issuer itself fails"7 for that separate track; it remains, in its own words, a live regulatory development area, the Code of Practice is still draft, consultation-stage material.
One connection is worth flagging rather than developing further here: redemption refusal (M3 in the taxonomy above) is a targeted, case-specific action by a functioning issuer. A wind-down plan's own "complete redemption for all holders" strategy, discussed in the failure-and-resolution guide, is a different thing again, an issuer-wide, ordinarily permanent process triggered by the issuer's own collapse, not a targeted refusal aimed at one suspect wallet. Don't read one as a variant of the other.
A colleague tells you a US-issued "unfreezable" stablecoin like USDH is now flatly illegal, "because the GENIUS Act requires every issuer to have freeze capability, and that law already exists."
The structural factor in practice: does the freeze even reach the token you're looking at
FinCrimeRadar's own guide on unhosted wallets and the visibility gap already works through this in depth, and this guide cross-references it rather than repeating it: before assuming any mechanism or authority basis above (M1-M6, A1-A3) actually applies to a given wallet, establish whether you're looking at the issuer's native token or a bridged or wrapped representation of it, which contract and chain actually hold the funds, who controls the bridge, and whether the issuer's own freeze or blacklist mechanism reaches that specific representation at all. That guide's own worked example, drawn from Circle's own bridged-token documentation, makes a sharper point than "the mechanism might not exist yet" on a bridged token: a standards-compliant bridged copy is typically code-identical to the native token from the moment it's deployed, meaning the freeze and blacklist mechanism, including the Blacklister and Pauser roles described in the taxonomy above, is already built in. What changes is who controls it, often a bridge operator, not the issuer, until an optional control-transfer step occurs. Read the full six-question topology checklist and the underlying FATF and Circle sourcing at "Questions to ask before assuming a freeze will work."
Four patterns, four cards
FAQ
Is Tether's freeze authority the same as Circle's?
Does the GENIUS Act make "unfreezable" stablecoins illegal in the US right now?
If an exchange freezes my account, has the stablecoin itself been frozen?
Can a frozen stablecoin be un-frozen, and is that the same as burning it?
Does a stablecoin issuer's freeze reach a bridged version of the same token on another chain?
At a glance
Four patterns, the risk that makes each one look routine, the signal that gives it away, and the response that fits.
Primary sources
The Huione/USDH quotes, both issuers' terms, and the GENIUS Act text were checked directly against the primary documents below on 30 August 2026, not against secondary summaries alone, except where a specific entry says otherwise. Where a fact rests only on secondary corroboration, that's stated in the text at the point it's used.
- FinCEN, Special Measure Regarding Huione Group, as a Foreign Financial Institution of Primary Money Laundering Concern (Notice of Proposed Rulemaking), 90 FR 18934, 5 May 2025.
- Circle, USDC Terms of Use, dated 12 December 2025, retrieved 30 August 2026.
- Tether, USDT Terms of Service, retrieved 30 August 2026.
- GENIUS Act (Guiding and Establishing National Innovation for U.S. Stablecoins Act), Public Law 119-27, 18 July 2025.
- Circle (stablecoin-evm repository), Token Design documentation, accessed 30 August 2026.
- FinCrimeRadar, Stablecoin Financial Crime: Unhosted Wallets, P2P Transfers and the Visibility Gap, cross-referenced for the token-topology questions and the FATF/Circle bridged-token sourcing behind them, not independently re-verified in this guide.
- FinCrimeRadar, Systemic Stablecoins: The Joint Bank of England and FCA Regime, cross-referenced for the failure-and-resolution material, not duplicated in this guide.
- Circle, Stablecoin Access Denial Policy, fetched and checked directly 30 August 2026, confirmed as the document Circle's own USDC Terms hyperlink to for its "blocklisting policy" reference.
- FinCEN, Imposition of Special Measure Regarding Huione Group, as a Foreign Financial Institution of Primary Money Laundering Concern (Final Rule), 90 FR [pagination not separately recorded], 16 October 2025, checked directly to confirm the anonymised-body/named-footnote pattern and the $29 million figure carry through unchanged from the proposed rule.
- Department of the Treasury, GENIUS Act Regulations on Payment Stablecoin Issuance, Offer, and Sale (Proposed Rule), 91 FR 53368, 18 August 2026, checked directly to confirm it is headed "Proposed Rules," not a final rule.
Secondary-sourced, not independently verified this session: the New York bankruptcy court's Multichain Foundation blacklist order (The Block, 31 October 2025); the disputed Zama pool freeze (CCN, June 2026); and public statements attributed to Circle's CEO on Circle's and Tether's relative freezing patterns (reported April 2026).
Name the mechanism, not the metaphor.
"Freeze" is a convenient word and a genuinely imprecise one. Several different actors, under several different legal and contractual bases, can each do something a headline would call "freezing" a stablecoin, using one of a handful of underlying mechanisms, and a structural limitation can defeat any of them regardless. Telling the mechanism, the authority behind it, and the limits on both apart is most of what separates a useful case note from a misleading one. FinCrimeRadar's screening tool checks sanctions, PEP, and adverse media exposure for the entities behind cases like this one, free, no signup required.