Costs and pricing

How the money moves

Four streams pay for a payout business: the FX spread (the gap between the real rate and the customer's), fees, float and, in workforce products, subscriptions. Against them sit partner and correspondent fees, the cost of money parked abroad, compliance and losses.

$2,000 to a contractor in Mexico

Back to the contractor, at an illustrative 18.50 pesos per dollar (easy arithmetic): MXN 37,000 at mid.

The wire costs about $102, or 5.1%: $45 to the US bank, $20 to the intermediary, $4 to the Mexican bank and about $33 in the rate. At this size two-thirds of the cost is visible fees, which is the best evidence I have against "the fee is the rate". Had the customer sent pesos, the US bank would have waived the $45 and taken a markup instead. On the local route the provider keeps about $16 plus any fee, minus small bulk-FX, partner and SPEI costs and the cost of capital on pesos parked in Mexico.

The 0.8% markup is my assumption, inside vendors' claimed range of roughly 0.5% to 2%; card push pricing isn't public.

$10,000 to a software contractor in India

Mid-market rate 96.3163 rupees per dollar on October 1, 2026, from a public provider's price feed.

The wire stacks a sending fee, an FX markup of 1% to 3.5% by whichever bank converts (my assumptions and a vendor survey), intermediary fees, Indian processing and FIRA fees (the FIRA is the bank's proof of inward remittance), and GST, India's sales tax, on the conversion. The provider route delivered about INR 958,000. That $250 to $400 gap is the market "the fee is the rate" describes. Either way the contractor needs a purpose code (P0802 for software consultancy), a FIRA as proof for zero-rated GST, and the money home within the 9-month realisation period.

$1,500 to a freelancer in the Philippines

Mid-market rate 62.5972 pesos per dollar on October 1, 2026.

The wire figure rests on a lot of assumptions. PHP 93,000 is over InstaPay's standard ₱50,000 cap; before its business version (₱500,000) arrived on July 29, 2026, that meant PESONet, a split or a wire. A catch I haven't confirmed: the Philippines' 0% VAT on services to non-residents relies on foreign currency remitted through banks, which a peso credit from a local partner may not prove. For scale, two consumer benchmarks: the World Bank found sending $200 cost about 6.5% on global average in Q1 2025, and the Dallas Fed puts US to Mexico at "slightly below 5 percent".

An engineer in Mexico City: EOR or contractor (estimates)

The overview's engineer, on my assumptions: gross MXN 90,000 a month (inside one staffing firm's 2026 senior range), 18.07 pesos per dollar (Banxico's FIX, September 30, 2026), one provider's EOR list price (about $700 a month), a 1% FX markup, no profit sharing.

The client saves about $2,226 a month, 31%, which is why misclassifying is tempting. Reclassified after two years, I estimate back social security, bonuses, three months' pay, 20 days a year and seniority premium at about MXN 970,000 (about $54,000) before surcharges, tax and fines; whether that's collectable from a foreign client with no Mexican assets is open. The contractor's net assumes Mexico's simplified tax regime (1% to 2.5%) and 0% VAT on exported services apply. In Brazil an employee costs about 1.7 to 2.0 times gross.

How companies of different sizes use it

Headcount says less here than two other measures, how many people a company pays abroad and how much money it sends abroad a year, so I give both for each size in the table below. Nobody publishes spend by company size, so the costs are my rough estimates from list prices and one buyer-data platform's contracts. The market prices by size anyway: in one market-data firm's 2026 sizing, small and mid-size businesses are the biggest revenue pool in business payments without being the biggest flow, so small companies pay in percentage points and the largest in basis points (hundredths of a percent).

A startup (under 50 people) usually pays a handful of contractors abroad, and maybe one or two employees through an EOR when a key hire wants benefits or local law requires employment. It signs up for a contractor platform such as Deel or Remote, both of which offer startups first-year discounts through investors and banks, and pays the odd supplier from a free multi-currency account at Wise or Airwallex. A founder or the first finance or operations hire runs it. My rough estimate is $5,000-40,000 a year in platform fees, plus the FX markup on everything sent. The push to the next step is the first contractor who has to become an employee, for benefits, a visa or local law, or one who has worked full-time inside the team for a year and now looks like an employee under most countries' tests.

A scale-up (about 50 to 500 people) pays dozens to a couple of hundred people in several countries, a mix of EOR employees and contractors, and starts paying suppliers abroad: agencies, data vendors, outsourced support. One platform covers most countries, usually Deel, Remote, Rippling or Papaya Global, and suppliers get paid through Wise, Airwallex, Payoneer or the bank's wire desk; Payoneer says it's moving upmarket toward customers like these. The head of people operations owns payroll and contracts, and the finance lead owns funding and FX. My rough estimate is $25,000-300,000 a year in fees, and the provider's FX markup on funding can cost as much as the subscription, which is "the fee is the rate" at company scale. What pushes it on is a country growing big enough for a local entity (vendors that sell both EORs and entity setup put that anywhere from a handful of people to a couple of dozen), and payroll FX losses that the CFO starts to notice.

A mid-market company (about 500 to 2,000 people) has its own entities in its biggest few countries, an EOR for the rest, and real supplier flows abroad. It runs global payroll through an aggregator or a payroll bureau in each country, and buys FX and supplier payments from a specialist aimed at this size, such as Corpay, Convera or Ebury, or from its house bank. Hedging starts here, and only partly: in a hedging vendor's survey of US and UK firms this size, four in five lost money on unhedged FX over 2025. The controller or a first treasurer owns FX and supplier payments, a global payroll lead owns payroll, and procurement picks the vendors. My rough estimate is $100,000-600,000 a year in platform and payroll fees, plus FX, whose cost depends on whether the company takes the specialist's quoted spread or negotiates tiers. Acquisitions that bring foreign payrolls and banks, preparing for an IPO with audited controls over payroll and treasury, and an unhedged quarter that shows up in earnings push it on.

An enterprise (roughly 2,000 to 10,000 people) pays most of its people abroad through its own entities and keeps an EOR to test a new market or bridge an acquisition. Its money moves through several banks (JPMorgan, Citi, HSBC, BNY, Deutsche Bank), with FX bought from their dealing desks against a credit line and specialists such as Corpay or Convera as second providers; Citi's WorldLink, for one, pays out in many currencies without the client holding local accounts. Supplier payments run out of the ERP, and payroll through an aggregator such as Deel or Papaya Global or a bureau in each country. Treasury owns FX, bank relationships and payment formats, shared services owns payables, and a global payroll team owns payroll. Platform contracts stay in six figures even at this size; the money is in FX and bank fees, which I'd put at roughly 5-30 basis points of what it sends. Bank tenders, a payment factory project and format deadlines such as Swift's structured addresses drive what changes next, and sanctions duties sit with the company itself, since OFAC's strict liability applies to it and not only to its provider.

A Fortune 500 company (revenue above about $7.5 billion) pays its people abroad from its own entities and routes the money through a global treasury: an in-house bank, a payment factory (one central unit that makes payments for every subsidiary) and a panel of global banks for clearing, cash pooling and FX. In PwC's 2025 treasury survey, about two-thirds of companies above $10 billion of revenue run an in-house bank. Hedging covers months of exposure, specialists handle niche currencies, and an EOR is a stopgap for a country with a handful of people. The group treasurer and a treasury center own it, with shared services for payables and payroll and HR deciding how people are employed. My rough estimate is millions to tens of millions a year in bank and FX costs, at single-digit basis points on major currencies, though nothing public confirms it. There's no next size up; the pressure comes from the board's FX policy and from rules that bite hardest at this scale, such as FATF's richer payment data.

What changes as a company grows (people, volumes and costs are my rough estimates):

On the way up, who carries the licence flips. A startup's provider holds the money transmission licences and the partner banks; a Fortune 500 company pays from its own accounts and carries the sanctions and reporting duties itself. FX buying changes the same way, from taking the provider's quote to negotiating spreads and hedging.

Interest on held balances (float)

Interest on customer money is a real business of its own at Wise and Payoneer: about a third of revenue at Wise (year to March 2026), and more than a fifth at Payoneer (2025, down 10%). Workforce products hold balances by design: in my Mexico example, two weeks of pre-funded payroll earns about $11 a month per employee at 4%.

Float moves with rates in both directions. Falling rates cost Corpay about 2 points of growth in 2025. Then on September 16, 2026 the Fed raised its target range by a quarter point, to 3.75-4.00%, its first rise since July 2023, so for now float has stopped shrinking. It could turn again, which is why Payoneer says it has hedged much of its interest income through 2028. It's limited elsewhere: send-and-done remittance holds little, the GENIUS Act bars stablecoin issuers from paying yield, and safeguarding rules limit where balances sit.

So I'd compare routes on local currency received per dollar paid, and on when it becomes usable. The visible fee is most of the cost on a small wire, a small part on a large one, and the whole price on a transparent route.