{"id":15393,"date":"2026-10-22T15:15:13","date_gmt":"2026-10-22T15:15:13","guid":{"rendered":"https:\/\/bjftradinggroup.com\/?p=15393"},"modified":"2026-10-22T15:15:13","modified_gmt":"2026-10-22T15:15:13","slug":"crypto-arbitrage-withdrawal-fees-transfer-times","status":"publish","type":"post","link":"https:\/\/bjftradinggroup.com\/ar\/crypto-arbitrage-withdrawal-fees-transfer-times\/","title":{"rendered":"Withdrawal Fees, Networks and Transfer Times: the Hidden Cost of Cross-Exchange Crypto Arbitrage"},"content":{"rendered":"

<\/p>\n

\n
BJF TRADING GROUP \u00b7 CRYPTO ARBITRAGE<\/span><\/p>\n

Withdrawal Fees, Networks and Transfer Times: the Hidden Cost of Cross-Exchange Crypto Arbitrage<\/span><\/h2>\n

A scanner shows a 0.6% gap between two exchanges. Whether that gap is money or a trap depends on things the scanner does not show: the withdrawal fee, the network you can use, how many confirmations the receiving exchange wants, and how far the price moves while your coins are travelling. This guide puts numbers on each of them and shows how professional arbitrage avoids most of them.<\/p>\n<\/div>\n

In this guide<\/strong><\/p>\n
    \n
  1. The short answer<\/a><\/li>\n
  2. The full cost stack of a cross-exchange trade<\/a><\/li>\n
  3. What happens during a transfer<\/a><\/li>\n
  4. Networks compared: fees, confirmations and credit times<\/a><\/li>\n
  5. Worked example: how a 0.60% gap shrinks<\/a><\/li>\n
  6. Break-even spread and trade size<\/a><\/li>\n
  7. Price risk while coins are in transit<\/a><\/li>\n
  8. How to calculate the true cost, step by step<\/a><\/li>\n
  9. Pre-funded arbitrage and rebalancing<\/a><\/li>\n
  10. Three ways to run cross-exchange arbitrage<\/a><\/li>\n
  11. Spreads that exist because nobody can close them<\/a><\/li>\n
  12. Practical rules to cut transfer costs<\/a><\/li>\n
  13. FAQ<\/a><\/li>\n<\/ol>\n<\/div>\n

    The short answer<\/h2>\n

    Short answer:<\/strong> a crypto withdrawal fee is a fixed amount of the coin that an exchange deducts when you move funds out. Together with the network you choose and the confirmations the receiving exchange requires, it decides how much a transfer costs and how long it takes. For cross-exchange arbitrage these costs are usually larger than the visible spread on small trades, and the waiting time adds price risk. That is why serious arbitrage keeps balances on both exchanges in advance and moves coins only in planned, batched rebalances.<\/p>\n

    Most explanations of how crypto arbitrage works<\/a> start with “buy where it is cheap, sell where it is expensive”. That sentence hides a transfer in the middle. On paper the transfer is a detail; in a real trade log it is where most of the edge disappears. The rest of this article is about that middle step.<\/p>\n

    The full cost stack of a cross-exchange trade<\/h2>\n

    A simple “buy on A, move, sell on B” trade has more cost lines than most people write down. Some are percentages of the trade size, some are fixed amounts, and one is not a fee at all but a risk.<\/p>\n

    \n\n\n\n\n\n\n\n\n\n\n\n\n
    Cost line<\/th>\nType<\/th>\nTypical size<\/th>\nWho sets it<\/th>\n<\/tr>\n<\/thead>\n
    Trading fee on exchange A (buy)<\/td>\n% of trade<\/td>\n0.02%\u20130.10% taker on major venues, higher on small ones<\/td>\nExchange A, your fee tier<\/td>\n<\/tr>\n
    Trading fee on exchange B (sell)<\/td>\n% of trade<\/td>\n0.02%\u20130.10% taker<\/td>\nExchange B, your fee tier<\/td>\n<\/tr>\n
    Slippage on both order books<\/td>\n% of trade<\/td>\nNear zero on deep pairs, 0.1%+ on thin ones<\/td>\nOrder book depth vs your size<\/td>\n<\/tr>\n
    Withdrawal fee on A<\/td>\nFixed, in coin units<\/td>\nFrom cents to several dollars per transfer<\/td>\nExchange A, per coin and network<\/td>\n<\/tr>\n
    Network fee (gas)<\/td>\nFixed, variable with congestion<\/td>\nUsually included in the exchange withdrawal fee<\/td>\nThe blockchain<\/td>\n<\/tr>\n
    Deposit fee on B<\/td>\nFixed<\/td>\nUsually zero; exists on some venues and coins<\/td>\nExchange B<\/td>\n<\/tr>\n
    Return leg (moving the quote currency back)<\/td>\nFixed<\/td>\nA second withdrawal fee<\/td>\nExchange B<\/td>\n<\/tr>\n
    Price change during transit<\/td>\nRisk, not a fee<\/td>\nGrows with time and volatility<\/td>\nThe market<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    Written as a formula, the net result of one transfer-based cycle is:<\/p>\n

    Net = Q \u00d7 (Spread \u2212 FeeA<\/sub> \u2212 FeeB<\/sub> \u2212 Slippage) \u2212 Withdrawalcoin<\/sub> \u2212 Withdrawalreturn<\/sub> \u00b1 Q \u00d7 Price move in transit<\/p>\n

    Q is the trade size in dollars. The important detail is that the percentage lines scale with Q while the withdrawal lines do not. That single fact explains why the same spread can be a loss at $500 and a profit at $20,000, and why small accounts are hit hardest. If you are still sizing your account, our guide on how much capital arbitrage needs<\/a> covers the other side of that question.<\/p>\n

    \n
    0.20%<\/span>Round-trip trading fees<\/span>0.10% taker on each side; the floor before any transfer cost<\/span><\/div>\n
    $0.10\u2013$5+<\/span>USDT withdrawal fee<\/span>Same coin, different network: up to 50x difference<\/span><\/div>\n
    1\u201360+ min<\/span>Transfer time<\/span>Exchange processing plus required confirmations<\/span><\/div>\n
    0.55%<\/span>Typical move in 25 min<\/span>For a coin with 80% annual volatility (model, 1 sigma)<\/span><\/div>\n
    ~5%<\/span>Of edge spent on rebalancing<\/span>Pre-funded example below, cheap networks<\/span><\/div>\n
    ~100%<\/span>Of edge spent on transfers<\/span>Same trades, moving coins on every trade<\/span><\/div>\n<\/div>\n

    Card values come from the worked examples in this article. Fees differ by exchange, coin, network and date.<\/p>\n

    What happens during a transfer<\/h2>\n

    A withdrawal is not one event. It is a chain of steps run by two different companies and a blockchain, and each step can add time.<\/p>\n

    \"Diagram
    Every stage between the buy on exchange A and the sale on exchange B is time for both prices to move. Ranges are indicative for large centralised exchanges.<\/figcaption><\/figure>\n

    1. Withdrawal request and exchange processing<\/h3>\n

    When you request a withdrawal, exchange A deducts the withdrawal fee at once, then runs its own checks: risk scoring, address whitelist rules, daily limits for your verification level, and sometimes manual review for larger amounts. Many exchanges also batch withdrawals from a hot wallet. On a quiet day this takes seconds; on a busy day or for a flagged transaction it can take 30 minutes or more. This part is invisible on the blockchain and impossible to speed up from your side.<\/p>\n

    2. Broadcast and network fee<\/h3>\n

    The exchange signs and broadcasts the transaction and pays the network fee (gas). That is why the withdrawal fee you see is usually higher than the raw network fee: the exchange sets a fixed price that covers gas in busy periods, wallet operations and a margin. Exchanges review these fixed fees from time to time, so a number you wrote down last month may be out of date.<\/p>\n

    3. Confirmations required by exchange B<\/h3>\n

    The receiving exchange will not credit your deposit until the transaction has a certain number of confirmations, which it sets per network. The count reflects how the exchange sees the risk of a block reorganisation on that chain. Multiplied by the block time, it gives the minimum waiting time even when everything else is instant.<\/p>\n

    4. Deposit credit, then the trade<\/h3>\n

    Only after the credit can you sell on exchange B. If the gap has closed in the meantime, you now hold a coin you did not want, on an exchange where you did not want it.<\/p>\n

    Networks compared: fees, confirmations and credit times<\/h2>\n

    The same token often exists on several networks. USDT, the most common way to move value between exchanges, is a good example: the coin is the same, but the cost and speed of moving it are very different depending on the chain you pick. Both exchanges must support the same network, and you must select it on both sides.<\/p>\n

    \n\n\n\n\n\n\n\n\n\n\n\n
    Network<\/th>\nTypical exchange withdrawal fee (USDT)<\/th>\nBlock time<\/th>\nConfirmations commonly required<\/th>\nTypical credit time*<\/th>\n<\/tr>\n<\/thead>\n
    Ethereum (ERC-20)<\/td>\n~1\u20135 USDT, moves with gas<\/td>\n12 s<\/td>\n12\u201364<\/td>\n3\u201315 min<\/td>\n<\/tr>\n
    Tron (TRC-20)<\/td>\n~1\u20131.5 USDT<\/td>\n3 s<\/td>\n19\u201330<\/td>\n1\u20133 min<\/td>\n<\/tr>\n
    BNB Smart Chain (BEP-20)<\/td>\n~0.1\u20130.3 USDT<\/td>\na few seconds or less<\/td>\n~15<\/td>\nabout 1 min<\/td>\n<\/tr>\n
    Solana (SPL)<\/td>\n~0.1\u20131 USDT<\/td>\n~0.4 s<\/td>\nfinalized slot<\/td>\nunder 1 min<\/td>\n<\/tr>\n
    Arbitrum, Optimism, Base<\/td>\n~0.1\u20130.5 USDT<\/td>\n0.25\u20132 s<\/td>\nexchange-specific<\/td>\n1\u20135 min<\/td>\n<\/tr>\n
    Polygon<\/td>\n~0.1\u20131 USDT<\/td>\n~2 s<\/td>\noften 100\u2013200 blocks<\/td>\n2\u201310 min<\/td>\n<\/tr>\n
    Bitcoin on-chain (for BTC)<\/td>\n~0.00001\u20130.0005 BTC<\/td>\n~10 min<\/td>\n1\u20133 (some venues more)<\/td>\n10\u201360 min<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    *After the exchange has broadcast the transaction; exchange-side processing comes on top. Ranges summarise publicly listed fees and confirmation rules of large exchanges in 2026; they change often. Always read the live withdrawal and deposit pages before a transfer.<\/p>\n

    Three practical conclusions follow from the table. First, the cheapest and fastest rails for moving dollars are usually stablecoins on low-fee chains, provided both of your exchanges support the same one. Second, Ethereum mainnet and Bitcoin on-chain are slow and expensive for frequent moves, even if they are the most widely supported. Third, the coin you are arbitraging may only exist on one chain, so the coin leg of a transfer can be much slower and more expensive than the stablecoin leg.<\/p>\n

    Worked example: how a 0.60% gap shrinks<\/h2>\n

    A mid-cap coin trades 0.60% higher on exchange B than on exchange A. You have $2,000 in USDT on exchange A and decide to buy, transfer and sell. All numbers are illustrative but realistic for 2026 fee levels.<\/p>\n

    \n\n\n\n\n\n\n\n\n\n\n\n\n
    Step<\/th>\nCost or result<\/th>\nIn dollars<\/th>\n<\/tr>\n<\/thead>\n
    Gross spread on $2,000<\/td>\n0.60%<\/td>\n+$12.00<\/td>\n<\/tr>\n
    Buy on A, taker fee 0.10%<\/td>\n\u22120.10%<\/td>\n\u2212$2.00<\/td>\n<\/tr>\n
    Withdraw the coin from A (fixed fee in coin units)<\/td>\nfixed<\/td>\n\u2212$1.60<\/td>\n<\/tr>\n
    Sell on B, taker fee 0.10%<\/td>\n\u22120.10%<\/td>\n\u2212$2.00<\/td>\n<\/tr>\n
    Send USDT back to A on TRC-20 for the next trade<\/td>\nfixed<\/td>\n\u2212$1.00<\/td>\n<\/tr>\n
    Net if the price does not move<\/strong><\/td>\n0.27%<\/td>\n+$5.40<\/strong><\/td>\n<\/tr>\n
    Transit: 10 min processing + 15 min confirmations<\/td>\n1-sigma move at 80% volatility \u2248 0.55%<\/td>\n\u00b1$11.00<\/td>\n<\/tr>\n
    Realistic outcome range<\/strong><\/td>\n<\/td>\nabout \u2212$5.60 to +$16.40<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    On paper the trade earns $5.40. In practice the result is dominated by the price move during the 25 minutes the coins spend travelling, and the expected value of that move is not zero: price gaps between exchanges tend to close, so by the time you can sell on B a good part of the 0.60% is often gone. If the full gap closes during transit, the trade loses the $6.60 of costs, plus or minus whatever the market did.<\/p>\n

    Key point:<\/strong> the visible spread has to pay for fixed transfer fees and for waiting time. On small sizes the fees eat it; on slow networks the waiting time does.<\/div>\n

    Break-even spread and trade size<\/h2>\n

    Because withdrawal fees are fixed amounts, the spread needed to break even falls quickly as the trade size grows. The chart below uses 0.20% round-trip trading fees and four levels of fixed transfer cost per cycle.<\/p>\n

    \"Chart
    Calculated: break-even spread = 0.20% + fixed transfer cost \/ trade size. Slippage and price risk in transit come on top.<\/figcaption><\/figure>\n

    At $1,000 a $5 transfer cost already needs a 0.70% spread just to break even, and a $15 transfer needs 1.70%. Gaps of that size between liquid exchanges are rare and short-lived. At $20,000 the same $15 transfer adds only 0.075%, but at that size two other limits appear: order book depth on the thinner exchange and withdrawal limits of your account level.<\/p>\n

    Price risk while coins are in transit<\/h2>\n

    A transfer turns an arbitrage into a short directional trade: from the moment you buy on A until you sell on B, you own the coin. A simple way to estimate how far it can move is to scale annual volatility by the square root of time:<\/p>\n

    Typical move \u2248 annual volatility \u00d7 \u221a(minutes \u00f7 525,600)<\/p>\n

    \"Chart
    Illustrative model. Real crypto returns have fat tails, so large moves happen more often than this curve suggests.<\/figcaption><\/figure>\n

    For a large-cap coin with 45% annual volatility, 15 minutes in transit already means a typical move of about 0.24%, half of a 0.50% spread. For a small-cap altcoin with 130% volatility, the typical move over the same 15 minutes is about 0.69%, larger than the whole spread. You can hedge the coin with a short position on a perpetual future during transit, but that adds its own fees and funding payments; our funding rate arbitrage guide<\/a> explains how those payments work.<\/p>\n

    How to calculate the true cost, step by step<\/h2>\n

    Before trusting any cross-exchange spread that needs a transfer, run these six checks. They take a few minutes by hand and can be automated.<\/p>\n

    \n
    \n

    1<\/span><\/p>\n

    Check wallet status on both sides.<\/strong> Withdrawals of the coin must be open on exchange A and deposits open on exchange B, on the same network.<\/div>\n<\/div>\n
    \n

    2<\/span><\/p>\n

    Read the withdrawal fee and minimum.<\/strong> Take them from the live withdrawal screen of exchange A, in coin units, and convert to dollars at the current price.<\/div>\n<\/div>\n
    \n

    3<\/span><\/p>\n

    Add trading fees and slippage.<\/strong> Use your real taker fees on both exchanges and the order book depth for your size, not the top-of-book price.<\/div>\n<\/div>\n
    \n

    4<\/span><\/p>\n

    Estimate total transit time.<\/strong> Exchange A processing plus the confirmations exchange B requires multiplied by the block time.<\/div>\n<\/div>\n
    \n

    5<\/span><\/p>\n

    Convert time into price risk.<\/strong> Typical move = annual volatility \u00d7 \u221a(minutes \u00f7 525,600). Require the spread to cover all costs plus at least this move.<\/div>\n<\/div>\n
    \n

    6<\/span><\/p>\n

    Price in the return leg.<\/strong> Decide how the quote currency gets back to exchange A: another withdrawal now, or a planned rebalance later.<\/div>\n<\/div>\n<\/div>\n

    If a spread survives all six checks, it is a candidate. If it fails step 1, it is not an opportunity at all, however large it looks.<\/p>\n

    Pre-funded arbitrage and rebalancing<\/h2>\n

    The way around both fixed fees and transit risk is not to transfer on every trade. Instead you keep balances on both exchanges in advance: quote currency (for example USDT) on the exchange where you buy, and the coin on the exchange where you sell. When a gap appears, both legs are executed at the same time, within seconds, so there is no transit and no directional exposure. This is how hedge arbitrage across exchanges is normally run, and it is the model used by VIP Crypto Arbitrage<\/a> for its hedge strategy.<\/p>\n

    The cost does not disappear; it moves. After a series of trades in the same direction, exchange A runs low on USDT and exchange B runs low on the coin. At some point you must rebalance: move coin from A to B and USDT from B to A. The difference is that a rebalance happens a few times per many trades, in larger amounts, at a time you choose.<\/p>\n

    \"Simulated
    Simulation: $10,000 per exchange, $1,000 trades, 72% of opportunities in the same direction, rebalance when one side drops to 20% of its starting balance.<\/figcaption><\/figure>\n

    Rebalancing cost in numbers<\/h3>\n

    Take the same simulation: 80 trades of $1,000, each with a 0.45% gross spread and 0.20% trading fees, so $2.50 net per trade before transfers, $200 in total. The balances needed 4 rebalances.<\/p>\n

    \n\n\n\n\n\n\n\n
    Model<\/th>\nTransfers<\/th>\nCost per transfer pair<\/th>\nTotal transfer cost<\/th>\nShare of $200 net edge<\/th>\n<\/tr>\n<\/thead>\n
    Move coins on every trade, cheap networks<\/td>\n80 pairs<\/td>\n$2.50<\/td>\n$200<\/td>\n~100%<\/td>\n<\/tr>\n
    Pre-funded, rebalance on cheap networks<\/td>\n4 pairs<\/td>\n$2.50<\/td>\n$10<\/td>\n~5%<\/td>\n<\/tr>\n
    Pre-funded, rebalance on Ethereum mainnet<\/td>\n4 pairs<\/td>\n$10.00<\/td>\n$40<\/td>\n~20%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    Calculated from the simulation above; transit risk is not included in the first row, which makes that model look better than it is.<\/p>\n

    Pre-funding has its own costs. Capital is split across exchanges and partly idle, and you carry counterparty risk on every venue that holds your funds. The coin balance also changes in value with the market unless you hedge it. These are the reasons to keep the number of exchanges small and to choose venues carefully; the list of supported exchanges<\/a> is a good starting point for planning which pairs of venues to fund.<\/p>\n

    Three ways to run cross-exchange arbitrage<\/h2>\n
    \n\n\n\n\n\n\n\n\n\n\n
    <\/th>\nTransfer-based<\/th>\nPre-funded (inventory)<\/th>\nPre-funded with perp hedge<\/th>\n<\/tr>\n<\/thead>\n
    Transfers<\/td>\nEvery trade<\/td>\nOccasional, batched<\/td>\nOccasional, batched<\/td>\n<\/tr>\n
    Transit price risk<\/td>\nHigh<\/td>\nNone during trades<\/td>\nNone; coin inventory hedged<\/td>\n<\/tr>\n
    Speed of execution<\/td>\nMinutes to hours<\/td>\nSeconds<\/td>\nSeconds<\/td>\n<\/tr>\n
    Capital needed<\/td>\nOne balance<\/td>\nBalances on each exchange<\/td>\nBalances plus futures margin<\/td>\n<\/tr>\n
    Main extra cost<\/td>\nWithdrawal fees on each trade<\/td>\nRebalancing, idle capital<\/td>\nFunding payments, futures fees<\/td>\n<\/tr>\n
    Realistic for<\/td>\nRare, large and persistent gaps<\/td>\nRegular hedge arbitrage<\/td>\nLarger books, volatile coins<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n

    Transfers can be avoided entirely by staying on one exchange, for example with triangular arbitrage across three pairs on the same venue<\/a>, or by trading the gap between spot and perpetual futures. Each of those has its own cost structure, but none of them pays withdrawal fees per trade. For the order-routing and data side of multi-exchange setups, see our guide to exchange APIs for arbitrage<\/a>.<\/p>\n

    Spreads that exist because nobody can close them<\/h2>\n

    The largest gaps on any scanner are often not opportunities. A price difference between two exchanges stays open when arbitrageurs cannot move coins to close it. Typical reasons:<\/p>\n