AEREDIUM says enterprise AI security must shift from model safety to cryptographic containment and structural authorization controls.

When OpenAI disclosed that one of its AI models escaped a restricted testing environment and breached Hugging Face’s infrastructure, the discussion quickly centered on AI safety. The questions were familiar: Can AI systems be aligned? Can they be trusted? Are today’s guardrails sufficient to prevent harmful behavior?

According to Eitan Katz, Chief Strategy Officer at AEREDIUM, those questions miss the larger lesson.

“This wasn’t just an AI safety incident,” Katz says. “It was a containment failure. Once an AI agent becomes capable enough, guardrails alone are no longer enough. Organizations need infrastructure that can cryptographically enforce what an AI agent is, and isn’t, authorized to do.”

The distinction matters because AI safety and AI containment solve different problems.

AI safety focuses on influencing a model’s behavior. It asks whether an AI system can refuse harmful requests, avoid generating dangerous outputs, or follow human instructions. AI containment begins from a different assumption: regardless of how capable or intelligent an AI agent becomes, it should never be able to exceed the authority it has been explicitly granted.

The OpenAI and Hugging Face incident illustrates that difference.

According to OpenAI’s own disclosure, the evaluation intentionally ran with production classifiers disabled and cyber refusals reduced. That makes the incident particularly instructive. Rather than demonstrating a failure of refusal training, it demonstrated what happens when structural controls become the primary line of defense. As Katz argues, once behavioral filters are absent, a capable, goal-directed agent will treat surrounding infrastructure as available surface unless something deeper prevents it from doing so.

That is why, Katz argues, containment is not fundamentally a filtering problem.

Model guardrails remain valuable for reducing accidental misuse and raising the cost of casual abuse. But they are probabilistic by nature, and they assume an AI system can be prevented or persuaded from taking an undesirable action. A sufficiently capable agent optimizing toward a specific objective may instead look for a path around those controls. The durable security boundary, Katz argues, must exist below the model itself.

“The durable control is structural,” Katz writes. “Authority has to be constrained below the point of decision, at the key itself.”

His conclusion is simple: “An action outside the mandate is not blocked. It cannot be produced.”

That philosophy forms the foundation of AERPOLICE.

Rather than attempting to determine whether an AI model is behaving safely, AERPOLICE is designed to assess whether an organization’s infrastructure can contain autonomous AI agents through structural controls. The framework focuses on whether authority is cryptographically enforced, whether permissions are bounded, and whether autonomous agents are prevented from executing actions outside the mandates they have been given.

For Katz, the implications extend beyond an organization’s own AI deployments.

The question is no longer only whether personal AI agents can be trusted. Enterprises should also assume that increasingly capable external AI agents will eventually interact with their systems. Containment therefore becomes part of an organization’s overall security posture, defining how well its infrastructure can withstand autonomous, goal-directed agents regardless of where they originate.

This also changes how enterprises should think about responsibility. Security can no longer depend solely on the behavior of the model or on the policies of whichever AI provider an organization happens to use. Organizations need controls that enforce their own authorization boundaries independently of the model itself.

None of this, Katz argues, diminishes the importance of AI safety. Guardrails continue to play an important role in reducing accidental harm and improving the overall AI ecosystem. But they should not be mistaken for the security boundary that protects enterprise systems.

The broader lesson from the OpenAI and Hugging Face incident, according to Katz, is that enterprise AI security is entering a new phase. As autonomous AI agents become more capable, organizations will increasingly need infrastructure that can enforce what those agents are authorized to do, rather than relying solely on what they are expected to do.

The future of enterprise AI security, he argues, will depend less on whether an AI model behaves correctly, and more on whether it is structurally prevented from exceeding its authority.

The same brand runs a wallet-based blockchain venue with no identity checks and a federally licensed exchange requiring a government ID and a live selfie. They list different markets, settle differently, and answer to different law.

Knowing which one you are using is the first thing a participant should settle, and the interface will not tell you.Most explanations of Polymarket describe a single platform, and that description has been wrong since December. There are two Polymarkets. One is the venue crypto has known for years: a blockchain application where anyone with a wallet and some stablecoins can take a position on almost anything, with no account, no identity verification, and no intermediary. The other is a federally licensed American derivatives exchange that asks for a government identification document, a social security number, proof of residency, and a live selfie before it will accept a dollar. They share a brand, an interface language, and increasingly a rulebook. They do not share a legal status, a settlement asset, a custody model, a product range, or a regulator. A trader who does not know which one they are on does not know what protections apply, what happens if a market resolves against expectation, or whether their position is a blockchain token or a claim against a clearing organization. This guide draws the line clearly, explains why it exists, and flags the reasons it might disappear.

Two entities, one brand

Start with the corporate structure, because the split is real at the entity level and not merely a regional interface variation.

The international venue is the original Polymarket: an application whose markets are settled on the Polygon blockchain, collateralized in stablecoins, accessed through a self-custodial wallet, and open to anyone whose jurisdiction permits it. There is no account in the traditional sense. Positions are tokens held at an address, trades execute against a public order book with settlement on chain, and outcomes are determined by a decentralized oracle process this publication has examined separately. Access restrictions operate by internet address and not by identity, which is why the platform can be geoblocked from a country without knowing who any individual user is.

Polymarket US is a different animal, operated through the CFTC-licensed exchange and clearing organization the company acquired in 2025 for a reported $112 million. It received an amended order of designation in late November and opened to users on December 2. It is a designated contract market in the full regulatory sense, the license the US venue holds, which means it lists contracts under federal derivatives law, clears through a registered clearing organization, and carries the obligations that come with both. Users complete full identity verification, fund in dollars through approved intermediaries instead of by connecting a wallet, and hold positions as claims within a regulated system instead of as tokens they custody themselves.

The practical marker for most readers: if you connected a wallet, you are on the international platform. If you uploaded an identification document and took a selfie, you are on the US exchange. Those are not two doors into one building. They are two buildings.

What changes for the user

Four differences matter enough to change behavior, and they compound.

Custody. On the international platform, positions are tokens in a wallet you control, which means you bear the risks and hold the powers of self-custody: nobody can freeze your position, and nobody can restore your access if you lose your keys. On the US exchange, funds sit in a regulated system with customer protections attached, and the corresponding trade is that the venue can restrict, suspend, or close an account under its rulebook.

Settlement asset. The international venue runs on stablecoins on Polygon. The US venue settles in dollars through approved intermediaries. That difference determines how you fund, how you withdraw, how long each takes, and what your tax records look like at the end of the year.

Identity. No verification internationally, where access is gated only by network address. Full verification domestically, including government identification, a social security number, proof of residence, and a liveness check. The identity requirement is what makes the US exchange’s surveillance apparatus function, because screening lists only work against names.

Access and availability. The US exchange removed its invite-only waitlist in May and currently reaches users through an iOS application, with other platforms not yet launched. The international platform remains blocked from US addresses under the 2022 settlement and blocked entirely in more than twenty other countries. Using a virtual private network to reach the international platform from a restricted jurisdiction violates the platform’s terms, risks account closure, and forfeits any recourse the regulated venue would have provided.

What changes for the market

The user-facing differences are the visible half. The structural differences shape what you can actually trade and what happens after you do.

Product scope is the sharpest divergence. A designated contract market lists contracts under federal derivatives law, subject to the review provisions this publication has covered in its guide to event contract listing, which constrains what it may offer. That is why product scopes differ. The international venue, outside that perimeter, can list markets the regulated exchange cannot, including contracts tied to armed conflict, leadership changes, and other sensitive developments. Two users on what looks like the same platform therefore see materially different universes of tradable questions, and the difference is not a product decision but a legal one.

Resolution differs in kind. International markets resolve through a decentralized optimistic oracle process, with proposals, a challenge window, and token-holder voting on disputes, which this publication has examined in detail. That is how the international book settles. The regulated exchange resolves under its rulebook, with the accountability and the recourse that a licensed venue’s procedures carry. The resolution risk that attaches to every event contract is therefore differently shaped on each side, and it is the risk most often underestimated on both.

Surveillance is the third structural split, and here the architectures are almost opposites. The US exchange runs layered monitoring including a real-time control desk and a regulatory services agreement with the National Futures Association for trade practice surveillance and sanctions. The international platform leans on the transparency of public settlement, where every holder in a contract is visible on chain, supplemented by third-party monitoring. One model watches identified people through institutional machinery; the other watches pseudonymous addresses in public. Both catch things the other misses.

Why the split exists

The structure is a direct product of enforcement history, not a design preference.

In January 2022 the CFTC settled charges that Polymarket had operated an unregistered facility for event-based binary options, imposing a $1.4 million civil penalty and requiring the company to wind down non-compliant markets and stop serving American users. The company kept its New York headquarters and served everyone else, which is how a business headquartered in the United States came to be geoblocked from it. Returning legally required a license, and instead of applying for one, the company bought one, acquiring an existing CFTC-registered exchange and clearing organization, a route this publication has examined as a pattern in this sector, where regulatory status functions as a purchasable asset. Federal investigations closed in 2025, the amended designation order followed in November, and the US venue opened in December.

The two-track outcome was therefore not a strategy chosen at a whiteboard. It is what remains when a global business rebuilds a compliant version of itself for one jurisdiction while the original keeps operating everywhere else, and it is the same shape this publication has documented in stablecoins, where an offshore issuer built a separate American vehicle instead of restructuring the parent.

Whether the split survives

Two developments suggest the architecture may be transitional, and both are worth watching.

The company published harmonized market integrity rules in March, applying substantially the same prohibitions on insider trading, spoofing, wash trading, front-running, and self-dealing across the international platform’s terms of use and the US exchange’s rulebook, along with public integrity pages for both. Running one standard across two legal regimes is what a company does when it expects the regimes to converge, or when it wants regulators to see no daylight between its venues.

More directly, the company filed with the CFTC in April seeking permission for US users to access the main global exchange. If granted in any form, that would begin dissolving the very split this guide describes, folding the deep-liquidity international book into the American perimeter. The company also applied for a margin trading license in July, and separately faces a reported regulatory review of its influencer marketing practices, which concerns advertising and not the legality of trading on the regulated venue.

Volume explains the motive. The international book cleared a record $10.8 billion in June on World Cup markets while the US exchange did more than $3.5 billion. The liquidity is offshore; the legal future is onshore; and no operator wants those two facts to stay separated indefinitely.

What the volume says

Numbers settle arguments that architecture descriptions leave open, and the volume split between the two venues is the clearest statement available about where this business actually lives.

In June the international platform cleared a record figure above ten billion dollars, driven by World Cup markets, while the regulated US exchange did more than three and a half billion. Both numbers are large, and their ratio is the point: the deepest liquidity, the widest market selection, and the largest share of activity sit on the venue that American users cannot legally reach, operated by a company headquartered in New York. That is the central awkwardness of the two-track structure, and it explains the company’s regulatory filings better than any strategy statement.

For a participant, the split has a practical consequence beyond the legal one. Liquidity is not a nicety; it determines the spread you pay, the size you can take without moving the price, and how reliably a market price reflects genuine information instead of the opinion of the last few traders. A market that exists on both venues will generally price better on the deeper one, and a market that exists only on the international platform has no domestic equivalent at all. Users restricted to the regulated venue are trading a smaller, newer book by construction, which is the cost of the protections that come with it.

The direction of travel is worth watching for exactly this reason. The company’s April filing asking the CFTC to let American users reach the global exchange is, read commercially, an attempt to resolve the split in favor of the liquidity. If regulators allow it in some form, the two-track structure this guide describes becomes a transitional phase in the sector’s history. If they do not, the structure hardens, and the American market develops its own liquidity separately over years. Both outcomes are plausible, and the filings are public.

The volatile layer

One category of information in this guide changes faster than the rest, and it should be treated as a snapshot, not a rule.

State-level access is contested and moving. Federal registration has not settled the question, because state gaming regulators across many jurisdictions maintain that sports event contracts are wagers requiring state licensing, producing cease-and-desist letters, litigation, and at least one enacted state ban with an effective date this year and a court challenge pending. The CFTC has sued multiple states asserting exclusive jurisdiction, its chairman has publicly described the conflict as a likely Supreme Court question, and a parallel line of cases brought under tribal gaming law, which this publication has covered separately, adds a third sovereign to the dispute. That is the state fights over access.

The practical instruction: verify current availability in your own jurisdiction at the moment you intend to trade, from the venue’s own disclosures, and treat any published state list, including any implied by this guide, as potentially out of date. The architecture described above is stable. The map of where each half may legally operate is not.

A final orientation point, because the two-track structure is not unique to this company and recognizing the pattern is more useful than memorizing one platform’s arrangements. The same shape appears across crypto wherever a business built globally meets a jurisdiction that regulates it: an offshore original continues serving most of the world while a smaller, licensed, identity-verified version operates domestically, with the parent carrying the liquidity and the twin carrying the legal future. This publication has documented the identical structure in stablecoins, where the largest issuer built a separately chartered American token instead of restructuring its global one, and it recurs in exchanges, custodians, and derivatives venues.

The pattern has a predictable life cycle worth knowing. It begins as compliance necessity, matures into deliberate strategy once the operator realizes the domestic vehicle is an option on regulatory outcomes, and resolves in one of three ways: the regulated version scales until the offshore one is redundant, the perimeter tightens until the offshore one is cut off, or the two converge because the regulator permits it. Polymarket’s April filing seeking access for American users to the global exchange is an attempt at the third path, which is the fastest and least costly of the three for any operator who can obtain it. Watching which path each of these dual-track businesses takes is one of the more informative things a reader can do with the next two years, because the answer will describe how much of crypto ends up inside the perimeter and how much stays outside it.

One practical addendum on record-keeping, since the two-track structure creates a bookkeeping problem most users discover in April. Positions on the international platform are blockchain transactions in stablecoins, with cost basis and proceeds derived from on-chain records you are responsible for reconstructing. Positions on the regulated exchange run through a supervised system that produces the reporting a domestic financial account produces. Those are entirely different tax documentation situations arising from what looks like the same activity on the same brand, and a participant who used both in one year has two separate reconstruction problems, one of which nobody will do for them. Capture transaction records at the time of trading on the on-chain side, because interfaces change and explorers do not organize themselves around your filing needs. Crypto.news has also explained how the DeFi side’s positions work.

Frequently asked questions

Disclaimer: This article is for information and educational purposes only and does not constitute financial, investment, or legal advice. Platform availability, regulatory status, and product scope change frequently and vary by jurisdiction, and pending litigation may alter the arrangements described. Always verify current terms with the venue directly. Always do your own research. Information is accurate as of July 28, 2026.

Ten European financial institutions have launched Regulated Layer One (RL1), a permissioned blockchain cooperative aimed at supporting regulated capital markets and tokenized assets. The network was established as a European Cooperative Society in Luxembourg and has started operations with an initial group of founding members.

RL1 announced that its governance model gives each founding institution equal decision-making rights over the network’s development. The cooperative is positioned as shared infrastructure for institutional workflows that typically rely on separate distributed ledgers, with the stated goal of reducing fragmentation across financial networks.

Key takeaways

• RL1 has been incorporated in Luxembourg as a European Cooperative Society and is already operating with founding institutions.
• ABN AMRO, Cecabank, Chartered Investment, Crédit Mutuel Alliance Fédérale, DekaBank, DZ BANK, LBBW, Natixis CIB, SC Ventures, and Seturion are the initial members.
• The permissioned network is built on infrastructure developed by SWIAT, which transferred ownership to the cooperative.
• SWIAT said the underlying platform processed 50+ transactions worth more than €700 million during three years of production use.
• RL1 is designed to support institutional applications such as tokenized bonds, collateral, and blockchain-based settlement.

A cooperative built for regulated tokenization

According to RL1’s announcement, the initiative is intended to serve regulated financial markets and tokenized asset use cases. The cooperative’s structure is designed to align governance with participating institutions, with each member holding equal decision-making rights for RL1’s network direction.

Among the founding organizations named by RL1 are major banks and capital market entities across Europe, including ABN AMRO, DekaBank, DZ BANK, LBBW, Natixis CIB, and NatWest is mentioned as an institution RL1 is currently in discussions with regarding joining. The list also includes Cecabank and Chartered Investment, as well as SC Ventures and Seturion.

From SWIAT infrastructure to RL1 ownership

RL1’s technical foundation traces back to infrastructure developed by German fintech Secure Worldwide Interbank Asset Transfer (SWIAT). RL1 said SWIAT has transferred ownership of the network to the cooperative, moving the platform from a fintech-led build to an institution-led shared asset layer.

SWIAT also provided performance context for the underlying platform. The company said it has processed more than 50 transactions totaling over €700 million (about $808 million) during three years of production use. For investors and market participants, this kind of prior operational record matters because permissioned blockchain deployments in finance often face scrutiny around scalability, reliability, and throughput under real-world conditions—areas that are difficult to assess without production history.

Why RL1 is positioning itself as shared settlement infrastructure

RL1 described the network as permissioned and geared toward institutional workflows. The stated target applications include digital money, tokenized bonds, collateral management, and blockchain-based settlement.

A central theme in RL1’s framing is interoperability within regulated environments. RL1 said that having a shared network could help reduce fragmentation that can occur when different financial institutions operate separate distributed ledger systems. In practice, this addresses a common friction point in tokenization efforts: without shared standards or compatible infrastructures, value transfer and settlement can become siloed across networks, complicating liquidity and operational integration.

RL1’s coop governance structure is also designed to reinforce this “shared infrastructure” approach. Rather than relying on a single operator, the cooperative model gives participating institutions equal say in governance and ongoing development, which RL1 suggests is intended to support long-term adoption across a wider group of market players.

Leadership and funding support

RL1 named Henning Vollbehr, former SWIAT Managing Director, as the network’s leader. KfW and L-Bank will continue supporting the initiative, RL1 said, indicating that development backing is expected to remain in place as the cooperative transitions from launch to broader deployment.

RL1 also said it is discussing additional institutional partners, including NatWest, about joining the network. The ability to onboard more institutions will likely be a key checkpoint for the project: the value of a permissioned network grows with participation, and RL1’s ambition to support settlement and tokenized financial instruments depends on whether additional banks, investors, or market infrastructure providers choose to integrate.

As RL1 begins operations, market observers will likely watch for concrete milestones beyond launch—such as which tokenized asset workflows the network will prioritize first, how quickly new institutions join, and whether the cooperative can convert its production track record from SWIAT into expanding real-world use across regulated markets.

In eighteen months, the prediction market industry has assembled the surveillance apparatus that took stock exchanges decades: named detection engines, encrypted prohibited-person lists, league partnerships, forensic academics, and employment disclosure for high-risk traders. Here is the architecture, who is building it, and the cost nobody in the sector wants to price.

Financial markets build their police forces after the scandal, not before it, and the construction usually takes a generation. The New York Stock Exchange operated for over a century before anything resembling modern trade surveillance existed; the systems that now watch equity markets for spoofing and insider activity accumulated in layers across decades of enforcement actions, statutes, and technology. The American prediction market industry has compressed that build into roughly eighteen months, and it has done so in public, under a congressional investigation, while growing volumes at a pace that makes each month’s controls obsolete by the next. The result is a surveillance stack with proper names: an in-house detection engine at Kalshi called Poirot, an encrypted prohibited-persons list called ProhiBet, an AI monitoring platform called HALO, a forensic academic partnership, a league relationship, a whistleblower button on market pages, and, as of last month, a requirement that traders in high-risk markets disclose their employer. 

This piece maps that architecture, identifies who is actually building it, and examines the trade it embodies, because a category whose entire value proposition is open participation has spent a year and a half constructing the machinery of exclusion.

Cointelegraph is committed to providing independent, high-quality journalism across the crypto, blockchain, AI, and fintech industries.

All news, reviews, and analyses are produced with full journalistic independence and integrity. For more details on our standards and processes, please read our Editorial Policy.

USAT went from $22 million to $140.8 million in a single month, the fastest growth of any regulated dollar token this year. In the same window, its offshore parent shed roughly six billion. One of those numbers is a rounding error on the other, and the gap between them is the most interesting structure in stablecoins.

There is a specific kind of corporate structure that appears when a very large business decides it may eventually need to be a different business, and Tether built one in January.

USAT is a dollar token issued through a federally chartered bank, designed from the ground up to satisfy the American stablecoin statute, run by a separate US entity with its own chief executive, and it is, by any measure of scale, almost nothing. It launched at $10 million. By the end of its first week, it held $17.6 million. Six months later, after the fastest month of growth any regulated dollar token has posted this year, it holds roughly $141 million, which is about eight hundredths of one percent of the $184 billion its parent has in circulation. Read one way, that is a failure to launch. Read another, it is a 540% month, faster growth than Circle, PayPal, or Ripple managed at any point this year, off a base small enough that the percentage means less than it appears. The interesting reading is the third one: USAT is not primarily a product. It is an option, written on a regulatory outcome, held by a company whose main business currently sits outside the perimeter the option would let it enter. This piece takes the numbers seriously, examines the structure that produced them, and asks what the twin is actually for.

The numbers, in order

Start with the sequence, because the growth story and the scale story are both true and point in opposite directions.

USAT launched on January 27 with a $10 million initial supply as an ERC-20 token, immediately available on several major exchanges. Anchorage Digital Bank’s first reserve attestation, dated January 31, reported 17,501,391 tokens outstanding against $17,604,716 in reserves, roughly 0.6% overcollateralized. By the end of March, circulation stood near $22 million. Then April: the Deloitte-signed reserve report published in late May showed circulating supply at $140.8 million as of April 30, an increase of about 540% in a single month, which the US entity’s chief executive attributed to institutional treasury operations, settlement flows, and regulated dollar liquidity management.

Now the context that the percentage conceals. Circle’s USDC sits around $75 billion. PayPal’s PYUSD is roughly $5.5 billion. Ripple’s RLUSD, itself a young institutional token, is about $1.7 billion. USAT at $141 million is therefore under a fifth of one percent of USDC, roughly two and a half percent of PYUSD, and about eight percent of RLUSD, which makes it the smallest meaningful entrant among the regulated dollar tokens competing for American institutional use. Against its own parent, the ratio is starker still: USDT’s circulation of roughly $184 billion makes USAT about 0.08% of the group’s outstanding dollar liabilities.

One further number completes the picture and is the reason this is a story instead of a launch update. While the twin grew, the parent shrank. USDT peaked near $190 billion in May and stood at approximately $184.1 billion on July 21, a decline of roughly $5.4 to $6 billion over sixty days, alongside a broader stablecoin market contraction of about $10 billion from its May high. The compliant American token is growing quickly from nothing while the offshore token it exists alongside is contracting by amounts larger than the twin’s entire supply, several times over, every month.

The structure, and who is in it

The corporate architecture explains more about the strategy than any growth figure, and each participant is worth naming.

Anchorage Digital Bank, N.A. is the issuer. It holds a national trust bank charter granted conditionally by the Office of the Comptroller of the Currency in 2021, well before the current administration, and describes itself as the first federally regulated crypto bank. Its chief executive has framed USAT as evidence of what stablecoin issuance looks like inside the US banking system, under supervision, with accountability. That is the structural core of the arrangement: Tether does not issue USAT. A chartered American bank does, under federal supervision, which is precisely the arrangement the offshore parent cannot currently replicate.

Cantor Fitzgerald serves as designated reserve custodian and preferred primary dealer, the same firm that has handled Tether’s Treasury holdings, and its former chief executive is now the sitting Commerce Secretary. The US entity is led by Bo Hines, previously executive director of the White House Council of Advisers on Digital Assets, appointed in September to run the American vehicle. Neither fact implies impropriety, and both were reported at launch. Together they describe something worth stating plainly: the compliance vehicle for the world’s largest offshore stablecoin issuer is staffed and served at the precise intersection of the policy network that wrote the framework it is designed to satisfy. In an industry where our own reporting has documented the crypto sector supplying more than a third of all corporate election money this cycle, that adjacency is part of the strategic picture, not a curiosity.

Distribution has been assembled in parallel: availability across major exchanges from day one, a payments integration with a commerce platform announced in February, and, in the chief executive’s framing, a stated ambition that Tether could become one of the largest buyers of US Treasury bills as demand for its dollar tokens grows.

The disclosure inversion

The most revealing detail in the entire structure is one almost nobody has commented on, and it runs opposite to what anyone would predict.

USAT, at $141 million, publishes reserve reports signed by Deloitte. USDT, at $184 billion, has operated for its entire existence on attestations rather than a full audit, a gap this publication has documented repeatedly and which S&P cited when it downgraded the token to the weakest grade on its stablecoin scale in December, alongside the rising share of higher-risk assets in the reserves. The small token has the stronger disclosure regime. The enormous one does not.

That inversion is not an accident; it is the whole design. USAT exists inside the federal framework, which imposes reserve composition, custody, and reporting requirements, and satisfying them is the token’s entire purpose. USDT operates outside that framework by choice and by history, with reserves that, as reported at USAT’s launch, do not currently align with the statute’s standards, while the company describes itself as progressing toward compliance. The group therefore runs two dollar tokens with opposite regulatory postures: one built to the American rulebook and audited to it, one built for global liquidity and disclosed on its own terms.

For anyone assessing Tether, this is the most useful lens available. The twin is proof that the group can meet the standard when it chooses to, on a token small enough that meeting the standard costs almost nothing. Whether the $184 billion business ever moves onto that footing is a different question, involving reserve composition changes at a scale that would reshape the company’s economics, and nothing in USAT’s existence answers it.