A7: the banks at the end of Russia’s payment detour

The most revealing feature of A7’s new offering is at the receiving end. In a commercial release distributed by Chainwire on 13 October 2026, the group presented a service in which customers fund an invoice with USDT, a dollar-pegged token, while the supplier receives a conventional bank transfer using Swift. A7 says the presentation took place that day, following its advance announcement on 12 October. A bank account therefore remains central to its promised independent international infrastructure. We have neither tested a transaction nor independently verified the service’s execution. ^[1][17]

Swift carries financial instructions. Institutions execute transfers and maintain the accounts in which money becomes available. As Swift itself explains, these are different functions. Replacing the first intermediary or the messaging method does not establish that every subsequent stage has become autonomous. For a business awaiting payment, the useful outcome is still spendable money in the currency specified by its contract. ^[2]

On 15 October, A7’s political backing became more visible. According to Denis Manturov’s secretariat, as reported by Interfax, Russia’s first deputy prime minister chaired the first meeting of a council attended by Finance Minister Anton Siluanov and Energy Minister Sergei Tsivilev, among others. The same announcement describes a consultative expert body bringing together government and business. That qualification matters: the documents obtained establish neither a sovereign payment guarantee nor the powers of a conventional corporate board. ^[3]

A7 may help sustain trade under financial restrictions while remaining dependent on institutions willing to complete its transactions. Following that dependence leads to overseas accounts, claims between intermediaries and the time an importer spends waiting for a supplier to be paid.

Political backing leaves contractual questions open

A7 was established in autumn 2024. Interfax reports ownership of 49% by state-owned PSB and 51% by Ilan Shor. Public backing and ministerial participation may help coordinate infrastructure considered strategically important. They do not establish which party must absorb a loss when a transfer is refused or an intermediary cannot return the money. Contracts, guarantees and refund provisions would be needed to answer those questions. ^[3]

That distinction matters as the United States steps up its action. On 1 October 2026, the US Treasury designated the A7 Network as a significant transnational criminal organisation under its sanctions framework and accused it of facilitating evasion, including for Iranian actors. This is an administrative action and the assessment of the authority taking it. A7 disputes the allegations. On 2 October, the company told Interfax that it did not work for Iran or terrorist organisations and said it applied compliance controls. We have not audited those controls. ^[7][8]

The US dossier combines alleged financial activity, public documentation and non-public information. Some explanations refer to the joint investigation published by Open Source Centre and TRM Labs in June. The relationship between those sources matters: repeated citations are not independent confirmations. Their accessible summaries describe substantial reliance on conventional banks; we did not obtain the leaked-document corpus or the complete Open Source Centre report. ^[5][9][10]

Still, opposing descriptions can be compared on one practical point. A7’s commercial offer promises a bank transfer to the beneficiary. The US assessment describes agents obtaining access to accounts and foreign-currency liquidity. Both assign an essential role to the banks at the receiving end, even though they give very different accounts of the uses and legality of the network. ^[1][5]

A payment unfolds in two places

In its proposal published in the Federal Register on 5 October, the Financial Crimes Enforcement Network, or FinCEN, describes a “mirror” system. Internal movements, including transactions in A7A5, help record relationships within the network. Separate operations performed by intermediaries abroad settle amounts in conventional currencies. A token-ledger entry and a bank credit to the supplier are different events. ^[5]

Consider a supplier waiting for an invoice to be settled. An intermediary may confirm that it has received the necessary value and recorded the instruction in its own system. The supplier still needs to see money credited to its account. Between those moments, the appropriate currency must be available, the payment accepted and the records reconciled. This simplified explanation does not reconstruct any actual A7 payment.

Method and scope

Simplified conceptual architecture from FinCEN’s description, compared with A7’s commercial announcement. No operational route, partner bank or individual transaction is reconstructed. Swift is a messaging system; institutions determine fund availability.

FinCEN estimates that, by June 2026, network subagents held accounts at approximately 435 financial institutions in at least 83 countries. It describes the messaging, banking relationships and foreign-currency liquidity available through those accounts. The agency also says institutions may be used unwittingly. Inclusion in that count is therefore not proof of knowing participation in sanctions evasion. The complete account list and underlying transactions were not public in the material obtained. ^[5]

A dispersed structure can be adaptable. It also creates many relationships that must continue to function. A business may switch payment providers without gaining access to genuinely different final-settlement banks. Apparent commercial diversification should therefore be compared with the concentration of actual settlement capacity. The material available does not support an A7-specific concentration measure.

The distinction is relevant to a European manufacturer. An intermediary’s confirmation alone does not establish that a supplier’s bank account has been credited or that the bank will ultimately accept the transaction. The contract determines when the commercial obligation is discharged; settlement records establish execution. Looking only at the status displayed by a payment interface would leave the second half of the process unexamined.

A ruble token still meets a foreign-currency invoice

A7A5 and the USDT mentioned in the October announcement have different reference currencies and roles. A7A5 is presented as a ruble-pegged stablecoin issued by Old Vector in Kyrgyzstan, with advertised backing in ruble deposits at PSB. USDT references the dollar. FinCEN’s alert describes A7A5’s backing arrangement; it is neither a reserve audit nor validation of every holder’s contractual rights. ^[1][6]

Backing raises a straightforward question: which claim can be redeemed for which currency? Suppose the token’s ruble-value promise is honoured. A supplier requesting yuan or euros still needs that currency. A maintained token peg establishes neither immediate access to foreign exchange nor a usable overseas account through which the beneficiary can receive it.

The issuer’s bank deposit must also be distinguished from a token holder’s rights. A banking asset held as a reserve does not automatically give every user direct access to the reserve account. Redemption terms, eligible claimants and suspension rights belong in contractual documents. We did not obtain a complete set allowing those rights to be reconstructed across all network transactions. The distinction between reserve assets and money available to spend is developed in our analysis of stablecoin redemption delays.

Method and scope

Claims-and-conditions diagram with no amounts or reconstructed consolidated balance sheet. FinCEN describes the advertised backing by PSB deposits; l0g has not certified those reserves. Redemption rights, liquidity and FX access depend on terms not audited here.

Technology can accelerate instructions and reconciliation. It can also reduce some coordination requirements. Those improvements remain compatible with dependence on foreign exchange and banks. In its general research on FX settlement, the Bank for International Settlements distinguishes the risk of paying one currency without receiving the other. Payment versus payment makes final settlement of one currency conditional on final settlement of the other, eliminating that principal risk for covered transactions. We have no evidence of equivalent end-to-end protection for the transactions discussed here. ^[16]

The comparison provides a useful test without asserting an observed A7 failure. A token transfer can become final on its ledger while a claim remains outstanding elsewhere. Irreversibility tells the user something about that ledger entry. The associated commercial obligation still needs to be traced to completion.

Large numbers measure different things

A7’s advertised scale needs careful interpretation. Reuters’ 15 October report attributes to A7 an estimate that alternative channels account for up to 80% of Russian foreign-trade payments. That category is broader than A7 itself. The figure is an interested-party estimate without a complete methodology obtained for this investigation; it cannot be used as A7’s own market share. ^[4]

The US authorities count different objects. FinCEN identifies at least $179.1 billion in A7A5 transactions between February 2025 and June 2026. Treasury separately describes more than $17 billion in subagent transactions between January 2025 and June 2026. The periods, populations and types of transaction differ. Neither total directly establishes the value of goods delivered, intermediaries’ revenue or net available liquidity. ^[5][7]

Repeated movement of the same value is one reason. One hundred units transferred four times produce four hundred units of gross transfers. A single invoice for one hundred can remain the only commercial obligation in the example. Depending on the system, other transfers may support financing or internal rebalancing. Linking ledger entries to invoices and completed settlements would be necessary to establish what those entries represent.

Method and scope

Four fictional accounts and a loop of four 100-unit transfers. Gross transfers equal 4 × 100 = 400; the illustrated invoice is 100. This ratio is neither an estimate nor a correction factor for A7 volumes. Linking the two measures requires evidence.

This simulation supplies no correction factor for A7. It explains why cumulative transfers cannot substitute for a trade statistic. Adding the $179.1 billion token total to the $17 billion subagent figure could likewise count multiple representations of the same activity while mixing periods. The material obtained does not allow their overlap to be calculated.

A large counter can indicate substantial activity and still fail to measure the service delivered. A more informative series would track completed commercial payments on a consistent basis: the amount accepted by the supplier, elapsed time, incidents and refunds after failure. We found no public audited dataset of that kind covering all of A7. The scale of trade actually settled is less firmly established than the existence of the network.

A sharp contraction, measured through June

A measurement independent of A7 provides a counterpoint to cumulative totals. In research published on 29 July 2026, Elliptic reported that average daily A7A5 transfer volume had fallen to $24.3 million in June, 96% below the July 2025 peak. These are the figures of a commercial blockchain analytics provider; we did not reproduce its raw series. They describe the token and that observation period. ^[11]

Method and scope

Source: Elliptic text published 29 July 2026. June daily average: USD 24.3 million. l0g index, July 2025 peak = 100; 100 × (1 − 0.96) ≈ 4 because the source rate is rounded. Zero-based bars. No monthly interpolation, reconstructed peak amount or October update. A7A5 is not the whole A7 network.

The fall is substantial, but its scope must remain attached to it. A historical cumulative total can remain impressive after new activity drops sharply. Conversely, less use of one token need not mean an end to activity through other instruments. A7’s October announcement explicitly mentions USDT and conventional bank transfers. Juxtaposing those facts raises a question about the network’s evolution; it does not quantify migration between its products. ^[1][11]

Elliptic discusses several possible drivers, including sanctions, Russian regulation and difficulties accessing exchange venues. Its analysis does not isolate a 96% causal effect of Western sanctions. Nor does it support the opposite claim that sanctions had no effect. As Elliptic sells compliance tools, its measurements should be assessed separately from its interpretation of their policy significance. ^[11]

As of 15 October, we lack a comparable series updating June’s level across all A7 services. Two observations remain: the company continues to promote its infrastructure, while an outside measurement documents an earlier sharp contraction in one of its instruments.

In a separate case, our analysis of Tether’s powers in an Iranian-oil seizure proceeding also distinguishes control of a wallet from the rights retained by a token issuer.

Enforcement moves towards the final bank account

Chronology clarifies the escalation. Core entities, including A7 and issuer Old Vector, were already targeted by the US Office of Foreign Assets Control on 14 August 2025. Designation of the network on 1 October 2026 expands that action. Treating October as the first sanction would erase more than a year in which the organisation could adapt. ^[12][7]

In the European Union, the regulation adopted on 23 October 2025 added A7A5 to the annex of crypto-assets covered by the transaction prohibition in Article 5ba. The date attached to the entry is 25 November 2025. Adoption and application are distinct events. This targeted historical reference is not a substitute for reviewing the EU law applicable to a particular transaction, including subsequent amendments and other restrictions. ^[13][14]

The October US initiative itself contains two different legal developments. OFAC’s designation is an adopted measure. FinCEN’s proposal concerns prohibiting certain transfers involving network subagents by covered US financial institutions. It was announced on 1 October, published on 5 October and opened for comments until 4 November 2026. At our cutoff, FinCEN’s status page still listed it as a proposal, with no final rule recorded. Existing sanctions continue to apply during the consultation; the proposal would add a separate instrument. ^[5][7][15]

Method and scope

Chronological markers; distances are not proportional to time. OFAC: entities in August 2025, network in October 2026. EU: A7A5 entry dated 25 November 2025 after 23 October adoption. FinCEN proposal announced 1 October, published 5 October, comments due 4 November. Jurisdictions retain their respective scope; this is not an exhaustive sanctions review.

An important feature is the proposed method for identifying the intermediaries concerned. FinCEN would provide institutions with a list through FI-Portal, its secure channel, and update it as the network changes. Some companies are named publicly; the contemplated complete list would not simply be an open directory. The text also provides a process through which an entity could seek removal from that list. ^[5]

The approach seeks to follow an organisation beyond its visible brands. More broadly, it shows where the confrontation takes place: banks’ ability to identify control relationships and refuse the transactions covered. A differently named company is not necessarily economically independent. Equally, allegations or name matches still require examination of evidence and the exact scope of the applicable rule.

Pressure may affect commercial decisions before a new rulemaking is complete: a bank can reassess a relationship under its existing obligations and risk policies. We did not, however, obtain a measurement of account closures caused by the October announcement. This architecture leaves execution exposed to settlement institutions’ decisions; the evidence obtained does not measure the resulting disruption.

When waiting reaches the purchase order

Macroeconomic risk emerges when a payment route still works but requires more time and liquidity to complete the same flow. A business may have a willing supplier, a profitable order and the necessary funds, yet remain constrained by uncertain settlement. The friction lies between deciding to buy and triggering shipment.

Take an entirely fictional example in accounting units. Payments of ten units are initiated every day and funded fully in advance. With a two-day average completion time, twenty units remain tied up in the process once it reaches a steady state. If the delay rises to eight days, maintaining the same pace requires eighty units. An additional sixty remain committed while payments are pending. The calculation assumes constant flow and excludes default, exchange-rate changes and fees.

Method and scope

Steady-state averages: pending balance = daily flow × average delay. Fully prefunded, unchanged flow of 10 units/day; no default, fees or exchange-rate movements. The two- and eight-day delays are entirely fictional. Insufficient resources may prevent the same throughput from continuing; no A7 delay or cash shortfall is estimated.

This is an additional financing requirement, not a loss of sixty units. The money may eventually reach every beneficiary. While it is pending, however, it cannot finance another purchase. A firm with ample cash or credit may absorb the delay; another must slow its ordering. Physical availability and the ability to pay for goods can therefore move at different speeds.

The economic impact depends on what is delayed. A hard-to-replace component can interrupt production, whereas inventories and alternative suppliers provide room to adapt. Slower receipt of export earnings can also reduce the liquidity available for the next operating cycle. The documents obtained support no A7-specific estimate of GDP lost, national average delay or missing imports.

Government participation could help mobilise liquidity or coordinate reorganisation. Its effectiveness still depends on the currency and institution needed for final settlement. Making rubles available may assist a Russian customer; crediting an overseas account requires the rest of the chain to work. That connection is what would need to be documented to assess the economic reach of the political support announced.

For the domestic financing side, our analysis of Russia’s proposed 2027 budget follows the links between public borrowing, banks and credit to businesses.

Continuity is the test of resilience

The strongest argument for A7’s resilience is practical: an intermediary can enable payments that no longer went through directly. It need not own every infrastructure it uses to provide a service. Ordinary payment systems also rely on partners. Dependence on foreign banks alone does not establish that A7 is ineffective.

It does identify the evidence that matters. Robustness would appear in completed settlements with stable delays, reliable refunds for refused transactions and continuity when a partner becomes unavailable. It would weaken if delays, pending balances and concentration at a few settlement points rose together. Those tests concern an observable economic service rather than the number of addresses, companies or transfers displayed.

The public record establishes the importance of the banking detour. A7 markets a service that reaches conventional accounts; US authorities seek to target intermediaries providing access to those accounts. Between them, the supplier is still waiting for the same outcome: a claim that has actually been settled. This is where the network’s capacity to sustain trade will be tested, and where the public data remain most incomplete.

Sources and limits

Documentary research as of 15 October 2026. A7 communications and Russian announcements are compared with US decisions and proposals, European documents and public financial-intelligence research. The council announced on 15 October is described as consultative by Manturov’s secretariat, as quoted by Interfax; its founding instrument was not obtained. A7’s claims, authorities’ allegations and third-party measurements retain their respective attribution.

The FinCEN alert and relevant passages of the Federal Register proposal were read; pages 1 and 3 of the alert were visually checked. The complete Open Source Centre PDF and leaked corpus could not be obtained: only its summary and TRM’s public analysis are used. Full direct EUR-Lex access was blocked, so official indexed passages establishing the A7A5 entry and its date were checked. This targeted reading is not an exhaustive EU sanctions review or legal advice.

No interviews, reserve audit, bank-account access, exhaustive independent blockchain tracing or test payment were performed. FinCEN and Elliptic figures retain their dates and were not reproduced from raw data. Joint investigations and their subsequent citations are identified to avoid overstating the number of independent confirmations.

Figures distinguish documented mechanisms, a published measurement and fictional simulations. Accounts A to D and accounting units are invented to explain arithmetic identities. No operational routing, wallet identifiers or details enabling the circumvention of controls are reproduced. Conclusions concern dependencies visible in the documents; no net trade volume, macroeconomic loss or general failure of sanctions is estimated.

1. A7 / Chainwire · A7 to unveil a service that will re-invent the international payments market. 2026-10-12. Announcement for 13 October: USDT funding and a conventional bank transfer to the beneficiary, using Swift. Explicitly sponsored content. No operational test, execution audit or compliance verification.
2. Swift · Who we are, question « Does Swift move money? ». changing page, read 15 October 2026. Swift carries instructions; financial institutions perform the actual transfer of funds. Changing page read on 15 October 2026. Does not describe A7 specifically.
3. Interfax / secrétariat de Denis Manturov · Мантуров, Силуанов и Цивилев вошли в набсовет А7. 2026-10-15. First meeting chaired by Manturov, attended by Siluanov and Tsivilev; the body is described as consultative and expert. Formation in 2024; PSB and Shor ownership. Attributed to the secretariat by the Russian agency; original government release and council charter not obtained. No sovereign guarantee inferred.
4. Reuters / Euronext · Russian ministers join board of key trade payments firm A7. 2026-10-15. Ministerial participation; A7 estimate of the share of alternative routes, not its own market share. Figures attributed to A7 or PSB are interested-party claims; complete methodologies were not obtained.
5. FinCEN / Federal Register · Proposal of Special Measure Prohibiting the Transmittal of Funds Regarding Transactions Involving the A7 Network’s Sub-Agents. 2026-10-05. 91 FR 63208–63226: mirror system, overseas accounts, figures and periods, FI-Portal list, proposed process and 4 November comment deadline. Announced 1 October; a proposal, not a final rule. Public and non-public inputs; account-level data and invoice-to-transfer matching not reproduced by l0g.
6. FinCEN · FIN-2026-Alert007: A7 Network. 2026-10-01. Network description, currency distinctions, token’s advertised backing by ruble deposits; pages 1 and 3 visually checked. Financial-intelligence alert, not a judicial ruling on each transaction. Operational indicators and identifiers omitted from the article.
7. US Treasury · Operation Economic Outcast Takes Unprecedented Action Against Sanctions Evasion Network Used by Iran. 2026-10-01. Network designation, expansion of earlier measures, US allegations and an estimate exceeding USD 17 billion for January 2025–June 2026. Source is the sanctioning authority. Allegations are attributed and A7’s response is included; no equivalence with a criminal conviction.
8. Interfax / déclaration A7 · «А7» опровергла обвинения в работе с Ираном на фоне введения новых санкций США. 2026-10-02. A7 denies working for Iran or terrorist organisations and says it operates compliance controls. Statement obtained by Interfax; controls not independently verified and no l0g interview conducted.
9. Open Source Centre · The Big Shor: A7 and the Illusion of Russian Financial Innovation. 2026-06-12. Joint investigation with TRM Labs, leaked documents, dependence on conventional banking. Public summary read; full PDF and leaked-document corpus not obtained. Joint work is not two independent confirmations.
10. TRM Labs · The A7 Leaks: TRM’s On-Chain Analysis of Russia’s Cryptocurrency Connections. 2026-06-12. Internal-ledger role and links to the joint investigation; explicitly cited by FinCEN. Commercial compliance provider; raw data not reanalysed. Documentary dependence with [9] and [5].
11. Elliptic · The fall of A7A5: how sanctions strangled the ruble stablecoin. 2026-07-29. June 2026 daily average: USD 24.3 million, down 96% from the July 2025 peak; several possible drivers discussed. Published textual figures, rounded rate; no raw series, October measurement or causal estimate isolating sanctions obtained.
12. US Treasury · Treasury Sanctions Cryptocurrency Exchange and Network Enabling Sanctions Evasion and Cyber Criminals. 2025-08-14. 2025 measures covering A7, A71, A7 Agent and issuer Old Vector, among others; predating the October 2026 network designation. US jurisdiction and its conditions; no universal prohibition inferred.
13. Conseil de l’Union européenne · 19th package of sanctions against Russia: EU targets Russian energy, third-country banks and crypto providers. 2025-10-23. Adoption of EU measures including restrictions concerning A7A5. Adoption date differs from the application date in the regulation’s annex.
14. EUR-Lex / Conseil de l’Union européenne · Règlement (UE) 2025/2033, article 5ba ajouté et annexe XVIII ajoutant l’annexe LIII. 2025-10-23. A7A5 is listed in Annex LIII, dated 25 November 2025. Full EUR-Lex access was blocked; official indexed English passages on Article 5ba and Annex LIII were read. No complete legal PDF was obtained or visually inspected. Targeted historical reference, not a full review of the law applicable to a transaction.
15. FinCEN · Special Measures. changing page, read 15 October 2026. A7 Network Sub-Agents row: NPRM dated 1 October 2026; no final rule recorded as of 15 October. Changing page; dated finding to be checked again if publication occurs after the research cutoff.
16. Banque des règlements internationaux · Uncovering FX settlement risk: new measures from the 2025 BIS Triennial Survey. 2026-06. FX settlement-risk definition and the role of payment versus payment. General mechanism. No A7 audit, score or assessment is attributed to the BIS.
17. A7 / Chainwire · Meet A7 International, an Innovative Global Platform Designed to Reinvent the International Payments Market. 2026-10-13. Corporate release, also republished by CryptoSlate on 14 October. A7 reports that the presentation took place on 13 October and describes USDT funding with the recipient paid by bank transfer using Swift. The distributor’s original was read through its newsroom link. Interested-party claims; no independent execution test or compliance audit.