To accept USDT and stablecoins, a merchant connects a gateway, exposes a checkout that lets the payer choose a token and network, and settles the received balance on demand. The two decisions that shape cost are which networks you accept and how settlement closes the rate. This guide covers both, with transfer costs measured on-chain on 11 September 2026 — including the part most guides still get wrong: on that date, an ERC-20 USDT transfer is cheaper than TRC-20.
If you need the definitions first, start with what is USDT and what a stablecoin is; this article is only about accepting one.
Why accept a stablecoin instead of a volatile coin
A stablecoin fixes the settlement value at the asset, not at the checkout. When a customer pays in USDT, the amount that reaches your balance equals the invoice minus the gateway fee, because the token holds a 1:1 peg to the dollar. With Bitcoin or Ether, the price can move between the moment the invoice is issued and the moment the payment confirms, and the merchant carries that gap. On a small order the gap is noise; on a large one, or across a day of orders, it becomes a real line in the accounts. A 1% move on a $10,000 invoice is $100 the merchant never priced for, and it lands on whichever side the coin moved against — with a stablecoin that variable is removed before it starts.
Reach is the second reason, and it is measurable. USDT and USDC together are about 83% of all stablecoin supply — roughly $311 billion in circulation as of 11 September 2026 (DeFiLlama). The businesses and companies that accept USDT are settling in the asset their counterparties already hold, which removes a conversion step on both sides: the payer does not buy a coin to pay you, and you do not sell one to bank the money.
For a merchant this narrows to two concrete gains. The settlement value matches the invoice, so pricing and reconciliation stay simple. And the payer base already holds the token, so checkout completion does not depend on the customer first acquiring something. The trade-off is that a stablecoin is still a crypto asset with a network and a network fee attached, and the network choice is where the actual cost sits. Everything below is about that choice.
You can see the full supported coins and how USDT works on Speend on the product pages; the operational reasoning is here.
USDT is not one network
USDT is a single token issued across several blockchains, and each blockchain is a separate address, fee, and confirmation time. A USDT balance on TRON cannot be sent to an Ethereum address, and a payment sent on the wrong network does not arrive — it is the most common way funds go missing at a crypto checkout. The payer selects a network when paying; the merchant decides which networks to expose in the first place.
On Speend, USDT is available on five networks, each with its own confirmation speed:
- TRON (TRC-20) — confirmation under 30 seconds
- Ethereum (ERC-20) — 1 to 3 minutes
- BNB Smart Chain (BEP-20) — 5 to 15 seconds
- Polygon — 5 to 15 seconds
- Solana — under 1 second
The five behave identically as dollars and differently as rails. Confirmation speed spans two orders of magnitude, from sub-second on Solana to a few minutes on Ethereum, and the fee spans a wider range still. Supply is concentrated on two of them: Ethereum carries about 48% of all stablecoin circulation and TRON about 31%, so together the two networks hold roughly 79% (DeFiLlama, 11 September 2026). That concentration is why most payer wallets default to Ethereum or TRON, and why those two are the networks a merchant almost always exposes first.
The choice of which additional networks to add is a cost decision (USDT is one of 300-plus coins across 18 networks a gateway like Speend can carry), and cost is where the received wisdom is out of date. For the underlying token mechanics, the ERC-20 vs TRC-20 network breakdown and the TRON network page go deeper than this guide needs to.
What a USDT transfer actually costs by network
On 11 September 2026, an ERC-20 USDT transfer on Ethereum costs about $0.13, while a TRC-20 transfer on TRON costs about $1.07 to $2.18 for a sender without staked energy. The common guidance — TRC-20 is the cheap default, ERC-20 is expensive — is inverted on this date, because Ethereum gas fell after its recent fee-reducing upgrades while TRON’s energy price stayed where it was.
Measured cost of a single USDT transfer:
| Network | Median cost (11 Sep 2026) | Note |
|---|---|---|
| Ethereum (ERC-20) | ≈ $0.13 | rises to ≈ $1.20 under peak load |
| TRON (TRC-20) | ≈ $1.07–$2.18 | burn model, no staked energy; higher end is a new recipient address |
| Polygon | ≈ $0.005 | fractions of a cent |
| Solana | < $0.001 | sub-cent, sub-second |
Method, so the numbers are reproducible: Ethereum and Polygon costs come from eth_feeHistory over a recent block window, applied to about 65,000 gas for an ERC-20 transfer. The TRON figure comes from the on-chain energy fee of 100 sun per unit of energy, at 31,895 energy for a recipient who already holds USDT and 64,895 for a new address that has to be initialized, priced at TRX ≈ $0.34 and ETH ≈ $2,600. The energy difference between an existing and a new recipient is why TRC-20 has a range rather than a single figure.
Two caveats keep this from being a gotcha, and both matter operationally. First, the ERC-20 number is a median. Under network load its p90 climbs to roughly $1.20, at which point Ethereum draws level with TRON’s lower bound but still stays under the cost of a new-address TRON transfer. So ERC-20 is cheaper on a normal block and comparable on a busy one, not cheaper always. Second, the TRON figure is the burn model — the cost when the sender pays with TRX rather than with pre-staked energy. A merchant who stakes TRX pays less, sometimes nothing, on outbound TRON transfers, so the $1–2 range is an upper bound on the merchant side, not a typical payout cost.
The takeaway is not that TRON is expensive. It is that “TRC-20 by default because it is cheapest” is no longer a fee argument, and the direction of the gap now depends on the date. The TRC-20 vs ERC-20 comparison holds the standing version of this; the figures above should be re-measured against your own date before you rely on them, because they move.
Which network to enable for which job
Enable the network that fits your payer’s geography and your ticket size, not a single default. With the fee near zero on three of the five rails and around a dollar on the fourth, the deciding factor is where your payers are and how large each payment is, not a headline about which chain is cheapest.
Want to accept crypto payments on your website?
Fast setup and KYC/KYB, fee starts from 0.5%
Contact Us| Network | Payer profile / geography | Ticket size | Enable when |
|---|---|---|---|
| TRON (TRC-20) | Retail, Asia, existing USDT holders | Any | Payers expect it; stake energy for your payouts |
| Ethereum (ERC-20) | B2B, desks and wallets already on Ethereum | Mid to high | Payer already holds ERC-20; the fee is minor on a large ticket |
| Solana | High-frequency, low-value checkouts | Micro to small | Many small payments where a sub-cent fee matters |
| Polygon | SaaS, subscriptions, in-app | Small to mid | Recurring low-value charges |
| BNB Smart Chain | Binance-ecosystem payers | Any | Your base sits inside that ecosystem |
The rule under the table is simple to apply. On a large B2B invoice a $1 network fee is a rounding error, so the right move is to route to whatever the payer already holds and let convenience win. On a stream of small payments the fee is the whole economics, so Solana or Polygon takes over, where the cost per transfer is a fraction of a cent. TRON stays relevant not on price but on habit: a large share of retail USDT holders keep balances on TRC-20 and will pay from there regardless of the fee, so exposing it is about not losing the payment. Exposing two or three networks rather than one raises checkout completion for the same reason, because the payer can pay from the balance they already hold instead of first moving USDT onto a specific chain.
One floor is worth knowing before you set this up. On TRC-20 the practical minimum sits near 1 USDT because of the network’s energy economics, which makes it a poor rail for micro-payments — those belong on Solana or Polygon. Speend lets a merchant set per-network minimums in the dashboard, so you can accept large TRON payments while pushing small ones onto a cheaper network.
Who pays the network fee, and what staking energy changes
On an incoming payment the payer covers the network fee; on a payout the merchant covers it. A customer sending USDT pays their own gas or energy to move the funds, so the cost of accepting a payment does not land on the merchant at all. The merchant meets a network fee only when moving funds out — a withdrawal to a wallet or exchange, or a payout to someone else.
This is where TRON’s energy model changes the merchant math. A TRC-20 transfer can be paid two ways: by burning TRX, which is the ~$1–2 figure measured above, or by consuming energy the account has staked to obtain. Staking TRX gives an account a recurring energy allowance that regenerates daily, and transfers that fit inside it cost little or nothing. A merchant with meaningful TRON payout volume stakes once and pays near-zero on outbound TRC-20 afterward, which is exactly why the burn-model cost is an upper bound rather than the number a busy merchant actually sees.
The other four networks have no staking equivalent. On Ethereum, Polygon, BNB Smart Chain and Solana the payout fee is simply the measured gas, and on three of those it is already fractions of a cent. So the practical read splits cleanly: on the accept side the network fee is the payer’s concern, and your job is to expose networks your payers can pay from cheaply; on the payout side, if TRON is a large part of your flow, staking energy turns the cost from a dollar-scale line into a near-zero one.
Settlement and withdrawal: a stablecoin sits on your balance
By default the USDT you receive lands on your merchant balance in crypto, and you withdraw on demand. Speend’s base fee starts at 0.5%. There is no conversion cost if you keep and withdraw the same USDT you accepted; if you auto-convert that USDT into a different withdrawal currency, a 1% conversion fee applies. Settling in the stablecoin you already received is therefore the cheaper default, and the reason to convert is a treasury preference, not a requirement. Stacked, a merchant who accepts USDT and withdraws it unchanged pays the base fee from 0.5% and no conversion; one who converts into another withdrawal currency pays that base fee plus 1%.
Auto-conversion closes the rate at the moment the payment arrives. For a merchant who accepts a volatile coin but wants a stable ledger, converting to a stablecoin at receipt fixes the value at that point rather than at withdrawal; for a merchant accepting USDT directly, the value is already fixed and the step is skipped. USDC behaves the same way as a settlement target, a dollar-pegged stablecoin you can hold or convert into on the same terms, and on Solana both USDT and USDC circulate as SPL tokens. If you want the distinction between the two, the what USDC is entry covers it.
Withdrawal runs to three destinations: a self-custodial wallet you control, an exchange account, or a supported bank channel where a fiat off-ramp is available in your jurisdiction. Funds credit to your balance at the provider and stay yours to withdraw, and keeping the same stablecoin you accepted preserves the invoice value from checkout through to payout. The full network and settlement detail for USDT on Speend sits on the product page; the operating decisions are the ones above.
FAQ
Which network should I accept USDT on?
Match the network to your payers and your ticket size. Expose TRON for retail and Asian payers, Ethereum (ERC-20) for B2B counterparties already on that chain, and Solana or Polygon for high volumes of small payments, where the fee is a fraction of a cent. Enabling more than one lets the payer pick the rail they already hold.
Is TRC-20 really the cheapest network?
Not on 11 September 2026. Measured on-chain, an ERC-20 transfer costs about $0.13 against roughly $1.07 to $2.18 on TRC-20 without staked energy, because Ethereum gas fell while TRON energy did not. The figures move over time, so measure on your own date rather than repeating the older claim that TRC-20 is always cheapest.
Why is a stablecoin better than a regular coin at checkout?
It fixes the settlement value. A stablecoin holds a 1:1 dollar peg, so the amount you receive equals the invoice minus the fee; a volatile coin can move in price between invoice and confirmation, leaving the merchant to absorb the difference.
Should I accept USDT or USDC?
Accept whichever your payers hold, since you can settle in either. USDT carries the larger share of stablecoin supply and of payer balances; USDC is the second dollar-pegged stablecoin and converts on the same terms. Enabling both takes the choice off your side and leaves it with the customer.
Who pays the network fee?
The payer pays it on the incoming transaction; the merchant pays it only on a payout or withdrawal. On TRON, staking TRX for energy cuts the merchant’s payout cost toward zero, while on the other networks the payout fee is just the measured gas.
What happens to the exchange rate, and how do I withdraw?
Received USDT sits on your balance, and you withdraw on demand to a wallet, an exchange, or a supported bank channel. Keeping the same USDT carries no conversion fee; auto-converting into another withdrawal currency costs 1%, and it fixes the rate at the moment the payment lands.
