Choosing your deposit network

The network is the one choice in the chain where a mistake can be irreversible. A wrong payout rail gives you a rejected transfer that comes back; a wrong network gives you a successful transaction in the wrong place, and recovery depends entirely on the technology.

This guide sets out the selection criteria, network by network, and the three traps that account for nearly every deposit incident.

The four criteria that matter

Sending cost

The dominant criterion on small amounts and negligible on large ones. Two costs stack: the withdrawal fee charged by the platform you send from, which it sets freely, and the real network cost, which depends on congestion.

On a 100 EUR equivalent send, choosing Ethereum over Tron can cost several percentage points. On 10,000 EUR the gap is noise compared with the spread. The full calculation is in USDT on TRC-20 or ERC-20.

Confirmation count

We require a confirmation count specific to each network, calibrated on the depth beyond which a chain reorganisation becomes improbable. That number sets your wait.

| Network | Confirmations | Average block | Estimated wait | | --- | --- | --- | --- | | Bitcoin | 2 | 600 s | ~20 min | | Ethereum | 12 | 12 s | ~2-3 min | | Tron | 20 | 3 s | ~1 min | | Solana | 32 | 0.4 s | < 1 min | | BNB Smart Chain | 15 | 3 s | ~1 min | | Polygon PoS | 128 | 2 s | ~4 min | | Arbitrum One | 20 | 0.25 s | < 1 min | | Litecoin | 6 | 150 s | ~15 min | | XRP Ledger | 1 | 4 s | seconds | | Dogecoin | 20 | 60 s | ~20 min |

These values come from product data and are repeated as-is on every asset page. A high count is not a sign of slowness: Polygon requires a hundred and twenty-eight confirmations but its block lasts two seconds, giving four minutes.

Memo requirement

Three networks we accept require an extra field alongside the address:

• XRP Ledger: Destination Tag, an integer.
• Stellar: Memo, text or integer depending on the type.
• TON: comment, a text message.

On those networks the deposit address is shared between customers and the memo is the only thing identifying you. Forgetting it turns an automatic deposit into a manual case, and sometimes into a permanent loss when the recipient is not us. Full detail in The missing memo.

Address format ambiguity

The least known criterion and the most dangerous.

An address starting with 0x followed by forty hexadecimal characters is syntactically valid on six networks: Ethereum, BNB Smart Chain, Polygon, Arbitrum, Base, Optimism and Avalanche C-Chain. No input validation can tell which one you meant. The only protection is the network selector, and it is often pre-filled with a value that is not yours.

Conversely, some formats are unambiguous: a Bitcoin bc1... address, a Tron T... address, a base58 Solana address, an XRP r... address. A network mistake on those formats produces a refusal, not a misdirected send. That is a real advantage.

Recovery, case by case

What "recoverable" means depends entirely on the technology.

Between EVM networks. The same private key controls the same address on every compatible chain. Sending ERC-20 USDT to a Polygon deposit address is recoverable: the operation needs manual work on the chain concerned, it has a cost, and it takes time.

To an unsupported network. If we run no node on the target chain we cannot sign a transaction to move funds out of it. The answer is no.

Without a memo, on an account we control. Recoverable after proof of control of the sending address.

Without a memo, to another service. Out of our reach. Only the recipient service's support can act, under its own policy.

An asset we do not accept. An unlisted token sent to a valid EVM address is technically recoverable, but we can neither convert it nor hold it. Only a return is conceivable.

The three traps that explain nearly everything

The network selector left on its default

On withdrawal screens the network is a dropdown under the address field, pre-filled. The user pastes a correct address and confirms without looking at that menu.

The countermeasure is mechanical: choose the network first, paste the address second. In that order, an address incompatible with the selected network is rejected by the form.

The maximum amount that deducts fees

"Send max" or "withdraw all" deducts fees from the displayed amount. You think you are sending 0.05, and 0.0498 leaves. That is not serious — underpayment is processed at the locked rate — but it surprises people. Detail in Underpaying an order.

Transaction fees set too low

On Bitcoin, a wallet's "economy" setting can leave a transaction in the mempool for hours. It is not lost, but it overshoots the rate lock window, and the order is re-quoted.

If your wallet supports RBF you can raise the fee afterwards. Otherwise waiting is the only option.

Choosing in practice

Here is the logic we recommend, in order.

1. Is the network forced by where my funds are? If your USDT sits on Ethereum, compare the cost of a bridge plus a withdrawal against sending directly over ERC-20. Direct often wins.
2. Is the amount small? Below a few hundred euros equivalent, systematically prefer a low-cost network: Tron, Solana, Polygon.
3. Do I need speed? Solana, Arbitrum and XRP Ledger confirm in under a minute. Bitcoin and Dogecoin take around twenty.
4. Can I avoid a memo network? If so, do. Not because memos are hard, but because forgetting one is the only deposit error that can be permanent.

Layer 2 networks and bridges

Arbitrum, Base and Optimism are layer 2 networks: they execute transactions on their own chain and periodically publish their state to Ethereum. Their appeal for a deposit is direct — very low cost, near-immediate confirmation — but two points are worth understanding.

Finality there has two levels. A transaction is confirmed on the layer 2 in a fraction of a second, and definitively settled on Ethereum much later. We work on the layer 2 confirmation, which is standard behaviour and sufficient for a deposit of this size, but it means the guarantee is not identical to that of a deeply confirmed Ethereum transaction.

A bridge is not a send. Moving funds from Ethereum to Arbitrum through a bridge is a separate operation with its own cost, its own timing and its own counterparty risk depending on the bridge used. Bridging then sending sometimes costs more than sending directly over ERC-20, especially on small amounts.

The practical rule: if your funds are already on a layer 2, send from that layer 2. If they are on Ethereum, compare the total cost of both paths before bridging. A bridge is only justified if you have other reasons to be on the destination chain.

Finally, bridge or no bridge, the address stays in 0x format: the network selector trap applies in full, and it is the only thing protecting you.

What we check before accepting a network

A network does not enter the catalogue because it is popular. We check:

• Finality. Does the network offer a clear notion of an irreversible transaction, and at what depth?
• Fee stability. A network whose inclusion cost varies by a factor of a hundred is hard to quote honestly.
• Availability of a reliable node. Without direct chain access you depend entirely on a third party to detect a deposit.
• Address format. An ambiguous format shared with other chains mechanically raises the incident rate.

Those criteria explain why some heavily used networks do not appear, and why the ones that do each carry their own confirmation count rather than a uniform value.

Articles attached to this guide

• USDT on TRC-20 or ERC-20 compares the real cost of networks for the same asset.
• The missing memo details the recovery procedure on XRP, Stellar and TON.
• Underpaying an order covers amount gaps and late sends.