A cross-border B2B settlement rail that charges a flat 0.3% is not offering a service. It is either subsidizing every transaction it touches or lying about the fee. There is no third option. When I read the A7 International announcement โ USDT in, SWIFT out, 0.3% flat โ I did not see a payment innovation. I saw a number that cannot survive contact with a real cost stack.
The announcement carried a timestamp: October 13, 13:13. Deliberate. Ceremonial. Marketing-driven, not engineering-driven. The entity behind it is A7 Group, operating in partnership with Promsvyazbank, a Russian state defense bank on the US SDN list since 2022. The chairman of that bank is Petr Fradkov, son of a former Russian prime minister, personally sanctioned. The CEO of A7 is reported to be Ilan Shor, a Moldovan businessman convicted in absentia in connection with the 2014 Moldovan bank fraud, also sanctioned. Three names. Three sanctions flags. One settlement rail.
That is the hook. Not the technology. The counterparties.
Ownership is an illusion without immutable proof. The corollary, for a settlement rail, is that trust is an illusion without verifiable custody. A7 International offers neither. What it offers is a flow โ USDT enters, rubles or dirhams or yuan exit through the SWIFT messaging layer โ and a story about bridging sanctioned economies. The story is real. The bridge is the question.
Context: The Market Vacuum That Built This
To understand A7, you have to understand what happened to Russian cross-border settlement after February 2022. The mechanism is not complicated. Major Russian banks were disconnected from SWIFT. Correspondent banking relationships collapsed. Western financial institutions that had spent decades building compliance infrastructure around Russian counterparties simply walked away, because the expected cost of a secondary sanctions finding exceeded any revenue the relationship could generate.
The result was a vacuum. Not a shortage of demand โ Russian exporters and importers still needed to move value across borders โ but a shortage of compliant plumbing. Into that vacuum poured everything that could not be sanctioned at the messaging layer: Chinese yuan through CIPS, Emirati dirhams through local clearing, Turkish lira, Central Asian intermediaries, and โ most importantly for this case โ USDT.
USDT has become the de facto dollar of sanctioned economies. This is not a controversial claim anymore; it is an observable fact. Tether's stablecoin offers what correspondent banking no longer does: deep liquidity, near-instant settlement, and โ critically โ a transfer mechanism that does not depend on any single bank's willingness to process a wire. For a Russian exporter who cannot receive a dollar wire, a USDT transfer is functionally a dollar.
A7 International's proposition sits precisely on this seam. The architecture, as described, is two-stage. The front end accepts USDT. The back end delivers fiat through what is described as standard SWIFT transfers to accounts in more than 100 countries. A7 claims 15,000 clients, roughly 2,000 transactions per day, and a 19% share of Russian corporate cross-border settlement as of 2025. It also cites an AA(RU) rating from ACRA, the Russian credit rating agency.
The company was founded in 2024.
Hold that timeline. We will return to it. First, the architecture, because the architecture is where the first structural lie lives.
Core: The Two-Stage Fracture
A payment rail is only as strong as its weakest interface. A7 International has two interfaces, and they are heterogeneous โ meaning they do not speak the same language, they do not share the same trust assumptions, and they do not settle on the same finality model. On one side is USDT on a public blockchain. On the other side is a SWIFT MT message that instructs a bank to debit one account and credit another.
These are not two halves of one system. They are two separate systems bolted together.
Between them, by necessity, sits a conversion step. USDT cannot be transmitted over SWIFT. SWIFT carries instructions, not tokens. So at some point in the flow, A7 โ or an entity acting for A7 โ must convert USDT into a fiat currency that a receiving bank will accept, and that conversion must happen somewhere, with someone, using some liquidity source. The original materials describe this in zero detail. That omission is not incidental. The conversion step is the entire business, and it is the one part of the rail that is completely undisclosed.
I have seen this pattern before. In 2017, I spent three weeks reverse-engineering the 0x Protocol whitepaper, cross-referencing its atomic swap proofs against the academic literature on slippage tolerance. The flaw I found was not in the math they published. It was in the liquidity assumption they left unpublished โ the silent premise that order books would always be deep enough for the tolerance to hold. The gap between the documented system and the operating system is where projects hide their fragility.
A7's gap is the OTC conversion. For USDT to become SWIFT-settleable fiat, it must land in a jurisdiction whose banks are not bound by the sanctions that severed the Russian banking system. The plausible candidates are the United Arab Emirates, Turkey, and Central Asian intermediaries โ the same corridors that have absorbed Russian trade flows since 2022. Whoever operates those conversion desks is the actual settlement counterparty. A7 is, at most, the orchestrator. At worst, the brand on top of an opaque liquidity pool.
Now consider the custody model. There is no trust minimization here. There is no bridge contract with verifiable state. There is no proof of reserves, no on-chain settlement finality, no independent attestation. The trust is fully concentrated in A7, in Promsvyazbank, and in the unnamed intermediary banks that clear the fiat leg. That is a custodial architecture wearing the vocabulary of a payment network.
I audited the Bored Ape Yacht Club contract line by line in 2021 and found twelve minor but structurally significant vulnerabilities in the metadata logic. The lesson that audit taught me was not about NFTs. It was that centralization does not announce itself. It hides in the parts of the system the documentation does not describe. A7's documentation describes the on-ramp and the off-ramp. It does not describe the custodian holding the funds between them.
There is no whitepaper. There is no audit report. There is no open-source repository. Every claim about the system's behavior comes from A7's own marketing material โ the "About A7" page and a promotional piece on a crypto media outlet that included a livestream invitation. This is not a technology milestone. It is a press release with a timestamp chosen for numerological effect.
Core: Stress-Testing the 0.3%
Let me do the arithmetic the announcement avoided.
Cross-border B2B settlement carries four cost components. First, the FX spread โ the difference between the mid-market rate and the rate the customer actually receives. For exotic or restricted currency pairs, that spread routinely runs 1% to 3%, and for sanctioned-corridor currencies it can exceed that. Second, the intermediary bank fees โ every correspondent bank in the chain takes a cut, and in a multi-hop route through third countries, there are more hops. Third, the compliance cost โ KYC, AML screening, sanctions screening, and the legal overhead of maintaining relationships in jurisdictions that tolerate this business. Fourth, the float cost โ the time value of funds held during settlement.
A flat 0.3% does not cover the first component alone in most sanctioned corridors. It certainly does not cover all four. So the question is not whether 0.3% is profitable. It is where the real revenue comes from. There are only three honest answers, and all three are red flags.
The first is that the advertised 0.3% is a teaser rate that applies to a narrow band of transaction types, and the effective all-in cost is materially higher once the FX spread is applied. This is the most benign explanation, and it still means the headline number is misleading.
The second is that the business monetizes the float โ holding client funds during the settlement window and earning on them. That is a duration mismatch, and it introduces liquidity risk into a rail that already has no disclosed risk controls.
The third, and the one I consider most likely, is that the revenue comes from the spread embedded in the OTC conversion step โ the same step A7 does not disclose. If A7 is buying USDT cheap and selling the fiat leg at a markup, the 0.3% is theater. The real fee is hidden inside the conversion.
A fee is not a business model. It is a hypothesis about where the money actually comes from. A7 has published the hypothesis and hidden the mechanism.
Core: Tether Is the Hidden Single Point of Failure
Here is the structural irony of the entire enterprise. A7's model does not control its own settlement medium. It accepts USDT, which means its front-end liquidity depends entirely on Tether's continued willingness to honor transfers involving A7-related addresses.
Tether has a documented history of freezing addresses in cooperation with law enforcement. The company maintains a freeze capability and has exercised it repeatedly. For A7, that capability is an existential threat. If Tether โ under OFAC pressure or independently โ flags and freezes the addresses A7 uses to receive client funds, the front end of the rail stops functioning within a single block. The SWIFT leg becomes irrelevant. The clients lose access to their inbound liquidity. The service dies.
This is a single point of failure that A7 cannot engineer away, because it does not sit inside A7's architecture. It sits inside a third party's compliance department. The stablecoin that made the business possible is also the switch that can turn it off.
There is a detail in the original materials worth flagging: a reference to "digital financial assets" and to potential promissory notes. In the Russian legal framework, a digital financial asset is a defined category that can operate outside the USDT rails entirely. If A7 is developing a parallel instrument โ an internal promissory note or a DFA-based settlement token โ that is not diversification. It is the creation of a quasi-currency, an unregulated credit instrument issued by a private entity in a sanctions environment. That is a shadow money market waiting to be born, and it deserves far more scrutiny than it has received.
Core: The SWIFT Off-Ramp Is the Technical Evidence
Return to the claim that A7 settles through "standard SWIFT transfers." This deserves forensic attention because it is technically impossible as stated, and the workaround is the crime.
Russian banks were disconnected from SWIFT. Promsvyazbank is a Russian bank. Therefore Promsvyazbank cannot originate or terminate a SWIFT message. If A7 is clearing through SWIFT, the messages must be sent and received by banks that are not disconnected โ non-sanctioned intermediary banks, likely in the UAE, Turkey, or Central Asia. Those banks are the technical enablers of the off-ramp.
This is where the case stops being about payments and starts being about sanctions evasion. Providing a settlement path that reconnects a disconnected bank to the international messaging layer is not a neutral commercial act. It is the functional equivalent of building a bypass around a blockade. The intermediary banks that process those messages are exposing themselves to secondary sanctions risk โ the same risk that caused compliant banks to exit Russia in the first place.
The contagion logic is straightforward. OFAC secondary sanctions target non-US entities that materially assist sanctioned parties. A third-country bank that knowingly clears A7's SWIFT traffic is a candidate. An OTC desk that converts A7's USDT into fiat is a candidate. The risk does not stay contained at A7. It propagates upstream and downstream through every node in the rail.
Trace the settlement leg and you find the liability. The liability is not in the code, because there is no code. It is in the banking relationships, and those relationships are the actual exposure.
Core: The Timeline Contradiction
A7 was founded in 2024. It claims a 19% share of Russian corporate cross-border settlement as of 2025. That is a twelve-to-eighteen-month window in which a new entity allegedly captured roughly one-fifth of a national settlement market.
I have run enough due diligence to be suspicious of any single-source market share number, and this one fails the plausibility filter on two grounds. First, the denominator is undefined. A 19% share of what? All Russian corporate cross-border settlement, including yuan and dirham flows that never touch crypto? A 19% share of the crypto-adjacent settlement niche? A 19% share of A7's own addressable customer base? Each of these is a radically different claim, and the materials do not specify. A share figure without a defined denominator is not data. It is a rhetorical device.
Second, the operational numbers do not obviously support a 19% claim. A7 reports 2,000 transactions per day and 15,000 clients. Run the arithmetic: 2,000 transactions per day across roughly 250 business days is 500,000 transactions per year. Divided across 15,000 clients, that is approximately 33 transactions per client per year โ a plausible frequency for a small-to-medium importer or exporter. The operational data is internally coherent. But operational coherence does not validate the share claim. It only tells us the volume numbers are not obviously fabricated.
The share number remains unverified. Every figure in A7's disclosure set originates from A7. There is no independent audit, no third-party verification, no regulatory filing that would let an outsider confirm the market position. When the Curve 3Pool depeg simulation I built in 2020 showed the pool's stability mechanisms would fail under simultaneous large-scale withdrawals, the team dismissed it as theoretical. The point was not that the failure was certain. The point was that the team had no independent verification of its own stability claims, and neither does A7.
Contrarian: What the Bulls Get Right
The reflexive critique of A7 is that it is a scam and the story is fake. That critique is lazy, and it misses the more important truth.
The demand is real. Russian cross-border settlement is a genuine, urgent, unmet need, and the market vacuum is not a narrative construct โ it is a measurable consequence of financial exclusion. When you sever a major economy from the international messaging layer, you do not eliminate its trade. You force it underground, into alternative rails, at higher cost and lower transparency. A7 is a symptom of that process, not a cause of it.
The broader thesis โ that stablecoins are becoming geopolitical instruments โ is also correct, and it is arguably the most important financial story of the decade. USDT is functioning as a parallel dollar system for economies the conventional dollar system has expelled. That is not a fringe observation. It is a structural shift, and the A7 case is a clean data point illustrating it.
Here is the contrarian move. The bulls are right about the trend and wrong about the counterparty. Conflating the macro reality of stablecoin geopolitics with the specific trustworthiness of A7 is a category error. The fact that the vacuum is real does not mean the entity filling it is sound. In fact, the opposite is often true: vacuum-filling entities in sanctions environments are selected for their willingness to absorb legal risk, not for their operational quality. The very conditions that make A7's market exist are the conditions that make A7 dangerous to transact with.
There is also a second-order point the bears miss. A7's existence is evidence that the sanctions architecture is porous. That is a finding about the architecture, not about A7. The interesting question is not whether A7 is trustworthy โ it is not โ but whether the sanctions regime can actually prevent the formation of these rails. The evidence to date suggests it can disrupt them, not eliminate them.
Takeaway: The Freeze Is the Only Certainty
A7 International is not a payment innovation. It is a sanctions arbitrage rail dressed in the vocabulary of fintech, and its commercial value is entirely exogenous to its own competence. It exists because the market is fragmented. If the fragmentation closes, the value closes with it. And the fragmentation is closing in the other direction โ toward tighter enforcement, not looser.
The signals to monitor are specific. Watch Tether's freeze actions against A7-related addresses. Watch OFAC's SDN updates for A7 and its intermediaries. Watch the sanctions status of the third-country banks carrying the SWIFT leg. Watch whether any independent auditor ever validates the 15,000-client and 19%-share claims. And watch for the alternative rail โ because in a market defined by a vacuum, the first credible competitor takes the share.
The macro story will continue. Stablecoins will keep functioning as geopolitical instruments, and the A7 case will be cited in parliamentary hearings and regulatory drafts for years. The question for anyone considering this rail is narrower and colder: do you want to be the counterparty on the day the freeze executes? Because the freeze is the only part of this system that is guaranteed to be immutable.
