What actually happens in the minutes a payment settles
The money is gone from your account. You can see that much. But where is it?
That question — deceptively simple, almost childish — turns out to be the most honest way to expose everything that is actually wrong with the payment infrastructure that governs how professionals close deals. When a wire goes out, the funds vanish from one ledger before they appear on another, and between those two events lies a gap that the banking system has trained everyone to treat as normal. It is not normal. It is a design artifact — an elaborate machinery built to manage the movement of obligations rather than the movement of money — and understanding exactly what fills that gap is the first step to understanding why onchain settlement feels like a different category of thing entirely.
This is not a story about technology. It is a story about time, and about what time costs.
The wire that isn’t a wire
Start with the vocabulary. A “wire transfer” conjures something immediate — electricity, instantaneous signal. The word dates to the age of the telegraph, when financial instructions genuinely did travel over wires at the speed of light. Wire transfers were named in the days of the telegraph wire, when direct transfers between financial institutions meant an instruction sent across copper. The wire survived the name but lost the metaphor. What actually moves when you initiate an international transfer is not money. SWIFT only passes messages between banks. It doesn’t actually move your money. When people talk about a “SWIFT transfer,” they mean a payment that uses SWIFT messages for coordination, not funds moved by SWIFT itself.
That distinction — message versus money — is the structural seam through which days get lost.
Here is the anatomy of a cross-border payment for a commission on a closed deal. A yacht broker in Monaco closes a sale to a buyer in Singapore. The gross proceeds flow from a Singapore account to the broker’s bank in France. Within the global banking system, not every bank holds accounts with every other bank worldwide. When the sending bank and the receiving bank lack a direct relationship, the payment must pass through one or more third-party banks called intermediary banks to reach its destination. Each of those intermediaries has its own processing window, its own compliance queue, its own cut-off time. Delays occur during compliance checks, currency conversions, and across time zones.
SWIFT transfers take one to five business days, with USD payments averaging 17 hours and EUR payments up to 23 hours. But that figure is an average for clean, direct-relationship transfers. For the broker in Monaco receiving from Singapore, the path likely runs through a correspondent in London or New York, then onward. When the sending and receiving banks have a direct relationship, funds move on a single path. When they don’t, the payment is routed through one or more intermediary banks. Each of these banks adds its own processing time; a transfer that involves several intermediaries may take four or five days, while a direct route will often only take one or two.
And what is happening to the money during those four or five days? Banks settle these transfers through Nostro and Vostro accounts — accounts that banks hold with one another to facilitate international payments. In practice, settlement occurs by adjusting the balances recorded in these accounts. The broker’s commission exists, during that window, as a ledger entry in transit. It is simultaneously not in the sender’s account and not in the receiver’s account. It is a claim in motion, passing through institutions that each take a processing fee for the privilege of holding it briefly before passing it along.
SWIFT transfer fees typically range from $25–$50 for sending, $10–$20 for receiving, plus $10–$30 per intermediary bank involved. Currency conversion markups of 1–4% over the mid-market rate add to the total cost. On a $200,000 commission, a 1.5% FX spread alone costs $3,000 — not as a fee anyone quotes upfront, but as a quiet trim applied somewhere in the chain. The total fee amount is unlikely to be broken down by all the intermediaries involved in a final statement. Instead, it is more likely to appear as a single fee charged by the recipient’s bank.
The broker has closed a deal. The money is moving. Nobody can tell him exactly where it is, or exactly when it will arrive, or exactly how much will remain when it does.
What “settled” actually means in traditional finance
There is a technical distinction in payments that almost nobody outside the back office articulates clearly, and it is worth holding in your mind precisely, because it is the lens through which everything else becomes visible.
“Cleared” means that the transaction details have been passed on. “Settled” means the two banks have exchanged money to honor the underlying transaction. “Disbursed” means the funds have been released to the end account holder.
Three separate events. Three separate moments in time. And between each event, uncertainty — not dramatic uncertainty, not the kind that makes headlines, but the low-grade ambient uncertainty that makes professionals refresh their banking portals compulsively on closing day.
When your business initiates an ACH payment, the funds do not move directly to the recipient’s bank. Instead, your bank groups the payment with others into a batch and submits it to an ACH operator, which sorts and routes each payment to the appropriate receiving institution. That batch processing structure is not a flaw; it is the system working as designed. The design prioritizes throughput — millions of transactions processed cheaply, in aggregate — over immediacy. The trade-off is settlement speed: standard ACH takes one to three business days to complete.
Wire transfers are faster, but “faster” is doing a lot of work in that sentence. The Federal Reserve’s processing schedule shows that banks can submit same-day files through three transmission windows: 10:30 a.m., 2:45 p.m., and 4:45 p.m. Eastern Time, with corresponding settlement at 1:00 p.m., 5:00 p.m., and 6:00 p.m. Miss the 2:45 window by six minutes and your transaction waits for the 4:45 submission. Initiate anything after 4:45, or on a Friday afternoon before a long weekend, and the clock does not restart until the next business day. A transfer sent on a Friday afternoon before a Monday holiday can sit unprocessed for three calendar days before it officially begins traveling.
Missing a cutoff by even a few minutes pushes settlement to the next business day, so understanding these mechanics can save you from overdrafts, late payments, and unexpected delays.
This is not a fringe scenario. This is closing on a Thursday afternoon when the deal takes longer than expected. This is an international counterparty in a time zone that is already closed. This is routine.
The deeper issue is what those days cost. Settlement timing is a working capital variable: for a merchant capturing $2M per day, the difference between T+1 and T+3 settlement is $4M of capital tied up in transit. This is a real cost — funding in transit is capital not available for operations, inventory, or investment. For individual professionals managing commission flow across multiple ongoing deals, the same arithmetic applies, just scaled down. And the cost is not invoiced. The cost from settlement delays can exceed what businesses pay in processing fees altogether. Yet it never appears in payment cost analyses because it’s not invoiced separately — it’s simply cash you don’t have access to.
There is one more layer beneath the timing: the question of finality. In traditional banking, even a wire that has “settled” is not unconditionally final. In traditional finance, credit card authorizations take seconds but settlement takes days, and chargebacks can reverse transactions months later. The wire is harder to reverse than a card payment, but not impossible — not in cases of fraud, error, or disputed ownership. Finality, in the banking sense, is provisional. It is a strong presumption underwritten by institutional rules, not a mathematical certainty.
The moment of broadcast
Now go to the other architecture entirely.
When a payment is initiated onchain, the first thing that happens is a transaction is constructed and cryptographically signed. The process begins when a user or application cryptographically signs a transaction request. This signature proves authorization and ensures the request can’t be altered. The signature is not a credential submitted to a gatekeeper for approval. It is proof, embedded in the transaction itself, that the authorized party commanded this specific transfer of this specific amount to this specific destination — and that nothing about it has been changed since the moment of signing.
Once signed, the transaction is broadcast to a single node connected to the blockchain network. Upon receiving the transaction, the node performs a series of standard checks. It verifies the cryptographic signature, confirms the sender holds the necessary balance, and ensures the transaction format adheres to network rules.
What happens next is the part that most people have never had described to them at a human level, because it sounds technical and abstract, and it isn’t either. The transaction is broadcast across the blockchain network using a peer-to-peer gossip protocol. The metaphor is almost perfectly accurate: nodes whisper the transaction to neighboring nodes, which whisper it to theirs. When a node receives a new, valid transaction, it doesn’t keep the information to itself. It forwards — “gossips” — the transaction data to the other nodes it’s connected to. Within seconds, this transaction has propagated across thousands of machines distributed around the world. Not copies of an instruction — copies of the transaction itself, held in the pending queues of nodes waiting for a block producer to include it.
The term “mempool” — short for memory pool — refers to the decentralized staging area within a blockchain network where pending transactions wait to be processed. Whenever a user initiates an onchain action, the transaction does not immediately enter the blockchain. Instead, it is broadcast to the network and held in the mempool of individual nodes. Think of this environment as a digital waiting room for unconfirmed transactions.
That waiting room is public. Anyone with a block explorer can see the transaction sitting there, waiting. The sender can see it. The receiver can see it. The amount is visible, the destination is visible, the timestamp of broadcast is visible. There is no opacity, no transit mystery, no “your funds are being processed.” The transaction exists, verifiably, in a shared record that no single party controls.
A block producer — miner or validator — scans their mempool and selects transactions for the next block, typically prioritizing by fee. The economics of that selection process are straightforward: miners and validators act rationally. To maximize their profit, they scan the mempool and prioritize transactions with the highest fees. It’s a simple auction: the higher you bid, the more likely your transaction will be selected for the next block. There are no business hours, no submission windows, no manual reviews. The mempool is open around the clock, every day of the year.
Inside the confirmation
The transaction is included in a block. This is the moment most people point to and call “confirmed,” but the architecture of what happens next is where the genuine power of onchain settlement becomes legible.
The transaction is included in a block and confirmed on-chain. The block propagates to all nodes. Every node removes the confirmed transaction from its mempool. The lifecycle is complete.
On different networks, that lifecycle completes in vastly different timeframes. Traditional ACH takes two to seven days. Credit cards take twenty-four to seventy-two hours. PayPal takes one to three days. Bitcoin requires roughly sixty minutes for six confirmations. Ethereum takes roughly fifteen minutes across two epochs. Avalanche achieves it in approximately one second. Solana provides optimistic confirmation at around 400 milliseconds, with full finality in roughly 13 seconds.
But what does “finality” actually mean in this context? It is not the same as “confirmed” in the banking sense — a strong presumption backed by institutional procedure. It is something more absolute. Settlement finality is the point at which a blockchain transaction becomes irreversible. After finality, the payment cannot be reorganized out of history, double-spent, or unwound by a chain reorg.
On Ethereum’s proof-of-stake architecture, validators vote on checkpoint blocks, and once two successive epochs receive sufficient attestations — a two-thirds majority — the earlier epoch is finalized. The math matters here. The economic cost of reversing even a handful of confirmations makes rollback practically impossible. On Ethereum, reversing a finalized block would require an attacker to control more than one-third of all staked ETH and sacrifice it in the attempt — billions of dollars of capital destroyed in the service of an attack that would immediately be visible to the entire network. The economics of reversal are not just unfavorable; they are structurally catastrophic for the attacker.
Deterministic finality works differently. In these systems, validators agree on the state of the network using sophisticated voting protocols. Once a supermajority agrees, the transaction is finalized instantly. There is no waiting for extra blocks. It is either in or out.
The language of “either in or out” sounds binary and simple. In practice, it represents something that traditional payment rails cannot offer at any price: a state change that is simultaneous, global, and irreversible. Not provisionally settled, not pending final posting, not subject to reversal under enumerated conditions. Final.
The anatomy of a split
Here is where the architecture starts to bear directly on the work of deal professionals.
A closed deal rarely pays one person. In almost every high-value transaction that moves through a broker, an advisor, an attorney, or an agent, the economics of the payment are multi-party from the start. There is a lead broker, perhaps a co-broker on the other side, a closing attorney taking a fee, a referral to be honored. In a traditional wire workflow, that multiplicity creates cascading complexity. Each payee needs their own wire instruction. Each wire needs to be initiated separately, confirmed separately, reconciled separately. The entire exercise happens after the deal closes, when everyone involved is already exhausted and eager to move on, and when any clerical error in an account number or routing code means the process restarts from a position of anxiety.
The fees compound. Wire transfers typically cost $15–$30 for domestic sends and $35–$50 for international transfers. Four parties to pay means four wires. Four fees. Four confirmation windows. Four opportunities for something to be wrong.
The more serious problem is not the fees. It is the sequence. Traditional payment infrastructure requires that funds aggregate before they distribute. Money lands in one account. Someone then divides it and re-sends it outward. That intermediate concentration step — the moment when all the deal’s economics sit in one place, controlled by one party — is a structural friction point that everyone involved simply accepts because there has never been a practical alternative.
This is the structural problem that onchain payment logic directly addresses. A transaction on a blockchain is not limited to one recipient. A smart contract can encode the deal’s payment structure before a single unit of value moves — splitting percentages, destinations, priorities — and execute all of those distributions in a single atomic operation. Blockchains remove intermediaries and rely on automated consensus instead of manual bank approvals. This dramatically shortens settlement times from days to minutes or seconds. But the more significant compression is not temporal. It is structural: the split and the settlement are the same event.
That is what a tool like Shaka makes possible in practice. A broker creates a payment link in advance, encoding the deal’s economic structure — who receives what, in what proportion — and when the transaction executes, funds move simultaneously to each destination wallet in one transaction. The professional who structured the payment closes the deal. Shaka handles how the money lands. There is no subsequent distribution step, no intermediate aggregation, no sequential wire chain. The payment architecture reflects the deal architecture, and both resolve in the same moment.
What finality feels like at human scale
Numbers compress experience. Let’s decompress one.
A commercial real estate advisor closes a transaction after four months of work. The buyer is based in the UAE. The seller is a fund domiciled in Luxembourg. The advisor is in London. There are two co-advisors: one in Dubai, one in Amsterdam. Total fees due across all parties on closing day: approximately €340,000.
Under the traditional wire workflow, the closing attorney — who receives the gross proceeds — initiates outgoing wires to each party after the final documents are signed. Each wire goes through the SWIFT network. SWIFT payments typically take two to five business days to settle, although they can take longer if weekends or public holidays get in the way. The Dubai advisor’s bank may require an additional compliance review. The Amsterdam co-advisor’s institution is not in a direct correspondent relationship with the English sending bank; there will be an intermediary stop in Frankfurt. Currency conversion introduces FX risk for transfers that take multiple days to settle, which can be significant when dealing with large sums. Exchange rates can fluctuate during the several-day gap between initiation and settlement, meaning the recipient can end up with a slightly different total.
The closing attorney sends four wires. Total outgoing fees: roughly £200. That sounds manageable until you account for the spread on EUR/GBP conversion and the correspondent bank deductions that come out of the amounts before they arrive. The correspondent fee can be deducted at any stage of the process. It could be an outgoing transfer fee charged by your own bank, an intermediary fee from a correspondent bank, or even an incoming fee charged by the recipient’s bank. Nobody told the advisor in Amsterdam that €1,200 was going to be quietly trimmed before the funds cleared. The advisor calls the attorney. The attorney calls the bank. Nobody has a clear answer. Swift’s own data shows that 90% of payments reach the beneficiary’s bank within an hour, but only 43% are credited to the recipient’s account that quickly. The biggest delay often isn’t the journey; it’s the receiving bank’s internal processing.
Meanwhile, the closing attorney waits three business days for confirmation that all four wires have landed correctly before closing the matter file. Three days of follow-up emails, portal checks, reconciliation notes. The deal closed on a Tuesday. Everyone gets paid by Friday — if nothing goes wrong, and if no one is in a time zone where Thursday’s wire missed the cut-off window and pushed to Monday.
Now run the same deal onchain.
The payment structure is encoded once, before closing: four wallet addresses, four proportions, one transaction. When the deal closes and the transaction broadcasts, peers receive, validate independently, and re-broadcast to their peers. The transaction ripples outward across the network within seconds. The block is produced. For treasury teams moving stablecoins between participants, finality is the moment risk leaves the books and cash is truly cash. All four parties — London, Dubai, Amsterdam, and wherever the third party is located — see their wallets update within the same confirmation window. Not four transactions. One. Not four confirmation timelines. One. Not four opportunities for something to arrive short of the agreed amount. One transaction that either executes as specified, or does not execute at all.
The attorney’s reconciliation task is not a multi-day process of chasing confirmations. It is the transaction hash: a permanent, immutable record that the payment happened, when it happened, in precisely the amounts specified. Once a transaction is finalized, it cannot be reversed on the blockchain. This immutability is a core security feature. The record does not depend on the memory or the cooperation of any single institution. It exists in the same decentralized ledger that every party can inspect independently.
The clock that doesn’t stop
There is one further dimension of the banking settlement problem that rarely gets surfaced explicitly, and it is perhaps the most corrosive: the banking system does not run continuously.
Many routine transactions clear in one to two days. Transfers initiated later in the day, near weekends, or before holidays may take the full three days or longer. That is not a quirk. That is the inherent constraint of a settlement infrastructure built around institutional business hours and centralized clearing windows. The Federal Reserve’s Fedwire system processes transactions between 9:00 a.m. and 7:00 p.m. Eastern Time on business days. Outside those hours, the system is not slower — it is closed. Most settlement processes operate on business days. Transactions captured on Friday typically settle the following Monday or Tuesday, depending on the settlement schedule.
A deal that closes at 4:30 p.m. on a Friday in New York is, from a payment standpoint, a Monday deal. The hours between Friday close and Monday open are not transit time. They are dead time — a structural silence built into the architecture of correspondent banking, inherited from an era when physical clearinghouses needed time to balance their books overnight.
The blockchain has no overnight. Polygon’s decentralized payment network operates 24/7 across borders. The network processes transactions at 2:00 a.m. on Christmas Day with exactly the same efficiency it applies on a Tuesday morning in September. There are no banking holidays in cryptographic consensus. There are no cut-off windows. Technological progress is moving finality toward near-instant certainty — and that near-instant certainty is available continuously, to any party in any time zone, without exception.
For professionals whose deals do not politely conform to the Federal Reserve’s processing calendar — which is to say, for nearly all of them — this is not a minor convenience. It is a structural advantage that changes what closing day means. The transaction broadcasts when the deal closes, not when the next processing window opens. Finality arrives in minutes, not the next business day. The parties receive confirmation — not a pending status, not a clearing notice, but cryptographic proof of receipt — before anyone has left the room.
Why days became acceptable
It is worth pausing to ask an honest question: why did the industry ever accept this?
The answer is that for most of financial history, there was no alternative. Blockchains remove intermediaries and rely on automated consensus instead of manual bank approvals — but until blockchains existed, manual bank approvals were the only mechanism available for global value transfer with any reasonable security guarantee. The correspondent banking network, for all its friction, represents an extraordinary feat of institutional coordination: SWIFT consists of a network of 11,000 financial institutions located in over 200 countries. Building trust across that many institutions, across that many regulatory jurisdictions, required the kind of slow, procedural, heavily documented infrastructure that SWIFT represents. The days of settlement were not indifference to efficiency; they were the cost of operating at global scale without a shared cryptographic ledger.
The second reason days became acceptable is that the cost of settlement delay is largely invisible. Payment processing fees are visible, negotiated, and intensely managed. Settlement timing costs are invisible, assumed to be non-negotiable, and completely ignored in most payment cost analyses. Nobody invoices you for the two days of interest you did not earn on funds sitting in transit. Nobody itemizes the FX spread as a line item before you authorize the transfer — it simply compresses the amount that arrives. Nobody bills you for the time your operations staff spent chasing confirmations. The costs are real; they are just structurally designed to be illegible.
And the third reason is simpler: the professionals who most need fast, accurate, multi-party settlement — brokers, agents, advisors, closing attorneys — were not, until recently, in a position to use an alternative. The payment infrastructure was upstream of them, institutional and immutable. Whatever the banks offered was what settlement meant.
What the minutes actually contain
So let us return to the question at the beginning. The money is gone from the account. Where is it?
In traditional finance, during the settlement window, the money is in motion across a chain of institutional ledger entries. It exists as a debit in one place and a pending credit in another, traveling through correspondent relationships and clearing windows at the pace of institutional procedure. This is systematic float capture by banks through payment settlement delays — not entrepreneurial financial engineering, but infrastructural value extraction built into the architecture of payment systems themselves. The delay is not incidental. In many cases, the delay is profitable — for the institutions, not for the parties to the deal.
In onchain settlement, during the confirmation window, the money is not “in motion” in the correspondent banking sense. The transaction has been broadcast. It exists in the mempool — visible, verifiable, waiting only for inclusion in the next block. The moment that block is produced, the payment cannot be reorganized out of history, double-spent, or unwound. The receiving wallet reflects the new balance. The state of the network has changed, globally and simultaneously, in a way that every node independently verifies and records.
This shift is crucial because it aligns blockchain behavior with how humans expect digital interactions to work. When you send an email, it arrives. When you submit a document, it is received. The expectation of immediacy — or at least rapid confirmation — is baked into every other form of digital interaction. Payments have been the exception, governed by legacy infrastructure that predates the internet and has never been fully replaced, only incrementally patched.
The minutes a crypto payment takes to settle are not empty minutes. They are active — a global distributed computation, thousands of independent nodes reaching cryptographic consensus, the network permanently updating its shared record of who owns what. What those minutes replace is not hours or days of faster processing. What they replace is the institutional machinery of correspondent banking: the nostro accounts, the clearing windows, the compliance queues, the FX spread, the correspondent fees, the cut-off times, the dead weekends.
All of that infrastructure exists to solve one problem: how do you move value between parties who do not share a ledger? The answer that the banking system developed over two centuries is complex, expensive, and slow by design. The answer that public blockchains developed is: share the ledger. Make it a public good. Make the rules cryptographic rather than procedural. Let anyone verify the state of the network without asking permission. And let the settlement happen not when the institutions agree it has happened, but when the mathematics says it has.
The professional who structures the deal — the broker, the advisor, the attorney — does not need to understand the consensus mechanism to benefit from it. They need only to understand what it changes: that when the deal closes and the payment broadcasts, the question of where the money is has a simple, verifiable, immediate answer.
It is exactly where it was sent. And it arrived the moment the network said it did.
