
What troubles people most when a token swap fails in TronLink is often not the failure itself. It is not knowing where the money went. The wallet shows one error line. From it you can hardly tell whether the transaction never reached the chain, the contract refused the swap, or execution ran short of energy.
Raising the fee limit or tapping Swap again may not solve it. The first job is to find the step where it broke, and to see what that step consumed.
The causes differ widely in both the fix and the cost. On TRON, energy and bandwidth already spent are not returned, even when no token comes back.
Copy the transaction ID from the wallet history and open it in TronScan. Two fields carry the answer: Result, and Energy Usage.
| What the record shows | What actually happened | Go to |
| TronScan cannot find it, balance unchanged | It never reached the chain | Cause 1 |
| FAILED, a few thousand energy used | It executed but missed the swap conditions | Causes 2, 3, 5 |
| FAILED, OUT_OF_ENERGY shown | Energy ran out during execution | Cause 4 |
Why not just check the TRX balance? An address that stakes or rents energy can fail a transaction and spend no TRX at all. The energy it used is still gone. An unchanged balance does not mean the attempt was free.
On 11 November 2026 we sampled recent swaps on SunSwap. In one group, 78 of 199 calls failed, close to four in ten. We opened 40 of those failures. Twenty were rejected because the amount received would have fallen below the user’s minimum. Fourteen had an input above the allowed maximum. Six failed in the token transfer itself.
An older group failed differently: 19 of 195 calls ran short of energy, and 6 reverted after execution.
Both kinds carry a resource cost, and the two are nowhere near the same size. The reverted swaps in our sample used 4,070 to 11,983 energy, median 4,149, which is the cheap end. The out-of-energy calls are the expensive kind, and Cause 4 has their numbers. Neither produced a swap.
Before you retry, confirm the signing address holds enough energy and bandwidth. Energy runs the contract, and bandwidth carries the transaction data.
A plain USDT transfer moves a token from one address to another. A swap also prices the trade against a pool, checks the conditions, and sends the output token back to you. So it usually costs more energy, and an estimate built for a transfer does not hold here.
In the TronLink mobile app, the swap button opens SUN.io. You are inside TronLink, but the swap is executed by SUN.io’s contracts. The wallet signs. How much energy the swap needs depends on the route it took and on what the contract did.
On 11 November 2026 we measured successful trades on three SunSwap routes. The spread was wide:
| Route | Successful trades | Median energy | Cost if paid in TRX |
| SunSwap V2 Router | 121 | 111,437 | 11.14 TRX |
| SunSwap V3 SwapRouter | 196 | 164,482 | 16.45 TRX |
| Earlier V2 router address | 170 | 182,828 | 18.28 TRX |
Those costs are worked out at 100 sun per unit of energy. On the first row, 111,437 energy comes to 11.14 TRX. That is well above the usual energy cost of a USDT transfer. What a USDT transfer costs on TRON against other networks shows how far above.
These are medians from a sample, not a price list. The real figure moves with the route, the state of the pool contract, and the dynamic energy factor.
That is how a 65,000 energy rental fails halfway. Someone sizes the order against the energy a USDT transfer needs, a figure that already doubles with size. The swap runs two-thirds of the way and stops. The rent is paid and the swap is not.
One cost is easy to miss. The first swap of a TRC-20 token may also need an approval, which lets the swap contract spend the tokens in your wallet. It is a second transaction with its own energy bill.
We simulated a USDT approval to the V2 router. It returned 99,764 energy, of which 77,076 came from the dynamic surcharge the network applies to heavily used contracts. At 100 sun per unit, paying that in TRX comes close to 10 TRX.
So before a first swap, price the approval as well as the swap. Check whether the wallet already holds an allowance, since a route that uses Permit2 can change the steps you sign.
Nothing is charged here, because nothing was mined. The usual reasons:
The fix. Confirm the wallet is on TRON mainnet, and that the account holds enough TRX for bandwidth and for any resource shortfall. Then reopen the swap page, take a fresh quote, and sign once. Do not resubmit an expired transaction. Build a new one.
This was the most common failure in our sample. The price you saw when you placed the order differed from the price at execution by more than you allowed, so the swap stopped.
Two messages appear on TronScan:
In the 40 failures we opened, those two accounted for 20 and 14.
Two things move the price. A large order of your own pushes along the pool curve, and other trades land first and change the ratio in the pool.
The fix. Take a fresh quote, cut the order size, or move to a pool with deeper liquidity. Do not reach for 5% or 10% slippage on the first failure. A wider tolerance only buys room for a worse fill. It does not add liquidity to a thin pool.
If the allowance is zero, or smaller than this swap needs, the swap can fail.
The message is sometimes TRANSFER_FROM_FAILED, which appeared in 6 of the 40 reverts we opened. The same message can come from a token balance that is too low. On its own it does not prove the approval is the problem.
Two details catch people out:
The fix. Read the contract address the swap page asks you to approve, and confirm it is the one this service uses. Send the approval, wait for it to confirm, then run the swap. If a route you have used before asks again, check the spender address before you sign rather than approving twice.
OUT_OF_ENERGY means the swap began executing and ran short of energy partway. The swap failed, and the energy and TRX already spent are not returned.
In our sample, the 19 out-of-energy calls reached 672,214 energy used. That is more than three times the median of a successful swap. The TRX burned for energy ran from 0 to 67.22, median 1.96.
Two limits produce this. One is the energy available on the wallet. The other is the fee limit, which caps how much TRX the call may burn. A fee limit set too low stops the call even when the wallet holds TRX.
The fix. Rent or stake energy to the address that sends the transaction, not to the address that receives the tokens. Size it for a swap rather than a transfer. Even 150,000 energy may not be enough on a heavy route, so simulate the swap first and leave a margin. The order is a two-number decision: volume and duration. How TRON energy rental works, and how its price compares with burning TRX covers both, and where the delegation lands.
If buying works and selling keeps failing, the token contract is the likelier cause than your settings. Thin liquidity or a slippage setting that is too tight can produce the same symptom.
The fix. Check the contract address against the project’s own channel, then look at whether sells are restricted and whether the pool is deep enough. If nothing looks wrong, test with a small amount you can afford to lose. No slippage setting and no amount of extra energy fixes a token that blocks sells.
This is the order we use ourselves.
The habit that costs the most is retrying without reading the cause. Every attempt that reaches the chain spends bandwidth and energy, and a revert gives neither back.
Does a failed swap still cost me?
Yes, if it reached the chain. A plain revert used about 4,100 energy and 539 bytes of bandwidth in our sample. Paid in TRX, both come to about 0.95 TRX. Rented energy was paid for in advance and is not charged again, but the energy it covered is gone. A swap that never broadcast costs nothing on-chain.
Why does the same swap work in the morning and fail at night?
Pool prices move, so the same order can miss its limit. Energy demand moves too. The network recalculates it every six hours, and the factor can raise the base requirement to as much as 3.4 times.
Can I swap with USDT only, and no TRX?
You can try, but not reliably. Rented energy covers the contract call. Bandwidth and any remaining shortfall still need TRX, so keep some in the wallet.
Will a higher slippage setting always get it through?
Sometimes. A wide tolerance is the right tool when the pool is deep. If the price genuinely moved a little between quote and confirmation, a slightly worse fill is still a fair trade. It is the wrong tool when the pool is thin or the token takes a transfer tax. There the fill gets worse without getting any more likely.
Does TronLink decide whether the swap succeeds?
No. The swap service quotes the price and builds the route, the contract executes on-chain, and TronLink connects, signs and submits. The reason for a failure sits on the TronScan record.
Disclaimer. Figures were checked on 11 November 2026 and will drift. Energy prices are set by committee vote, contract energy requirements are recalculated every six hours, and rental prices move hourly. Verify the numbers before you size a large order. Nothing here is financial advice.