Industry InsightsSeptember 8, 2026
Before anything else, a correction that changes the arithmetic. Almost every guide to TRON fees currently in circulation, including several fee calculators, states that a USDT TRC-20 transfer consumes 31,895 energy, or 64,895 to an address that has never held USDT.
Measured on 1 September 2026 from confirmed on-chain receipts, the real figures are 64,285 and 130,285. Roughly 2.2 times what the internet says.
The cause is TRON's dynamic energy model, which applies a surcharge to heavily used contracts. The USDT contract currently carries an energy factor of 34,000, which is the network maximum, a penalty of 3.4 times its base cost. It cannot go higher. It can fall.
If you are budgeting withdrawal costs off the widely-quoted numbers, your model is wrong by more than half.
What a transfer costs today
TRON prices energy at 100 sun per unit, set by on-chain committee proposal 104 and effective 29 August 2025, readable live from the chain parameters. That was a cut from 210 sun, which itself replaced 420 sun in September 2024. Fees have fallen by roughly 76% in two years. One aside worth getting right: the 2025 cut is widely called "Proposal 789", which is the GitHub issue number for the improvement proposal. The on-chain governance proposal is number 104. Most coverage conflates the two numbering systems.
At 100 sun per energy unit, with TRX at $0.3316 as of 1 September 2026
There is also a bandwidth cost of around 345 bytes. Every account gets 600 free bandwidth per day, which covers exactly one transfer, because two would need 690. Past that, bandwidth burns TRX too. There is no free energy allowance of any kind. Energy comes from staking, from delegation, or from burning TRX. If the whole resource model is new to you, start with what TRON is and what gas limits do.
Staking is not an expense, it's a balance sheet decision
This is the part that gets modelled wrong, and it is why finance teams and engineering teams reach different conclusions from the same numbers.
Burning TRX spends it. The TRX is destroyed and it is gone from your treasury permanently. Staking TRX locks it. You still own it. It is illiquid for 14 days after you unstake, and it carries price exposure while it sits there, but it has not left your balance sheet.
Under Stake 2.0, energy is allocated proportionally: your share of the network's total staked TRX determines your share of the daily energy supply. As of today the network's total energy limit is 180 billion per day against 18.77 billion TRX staked for energy, which works out to roughly 9.59 energy per TRX staked per day.
Run that against a transfer. Covering one 64,285-energy transfer per day needs about 6,705 TRX staked, roughly $2,223 at today's price. That stake saves 6.43 TRX of burn per day, which is 2,346 TRX per year.
So a locked position of 6,705 TRX returns 2,346 TRX in avoided burn annually. Call it a 35% annual yield, denominated in TRX, on capital you still hold.
Scale it linearly. A payout service processing 100 USDT withdrawals a day needs around 670,500 TRX staked, roughly $222,000 locked, to eliminate a daily burn of 643 TRX. Over a year that is about 234,600 TRX not destroyed.
Two real caveats. The 9.59 figure moves as network staking shifts, so it needs re-checking rather than hard-coding. And 14 days of illiquidity plus TRX price exposure is a genuine cost that a treasury policy has to price.
The rental market, and what "energy ratio" means
Stakers with surplus energy delegate it to renters for less than the burn price, through marketplaces including JustLend DAO, TronSave, Feee.io, and TokenGoodies. Indicative pricing runs somewhere around 26 to 74 sun per energy against the 100 sun burn price.
Treat those as ranges, not quotes. I could not verify a live rate on any primary source. Comparison sites disagree materially on the same platform, and most of them are resellers with an interest in the answer. Prices move hourly.
The term "energy ratio" is worth pinning down, because it is used loosely for at least three different things: energy produced per TRX staked, rental price divided by burn price, and the share of TRX supply staked for energy. Say which one you mean.
Where the line sits
Below roughly ten transfers a day, burn. The stake required to cover it is capital locked against a cost that barely registers, and the operational overhead of managing delegations is not worth it.
Between ten and a few hundred, rent. You get most of the saving with none of the lock-up, at the cost of a dependency on a marketplace with no SLA and no guarantee it clears when you need it.
Above a few hundred a day, stake for your baseline load and rent the peaks. At that volume the 35% figure stops being a curiosity and starts being the difference between a withdrawal service that costs money and one that doesn't, which is the same calculation behind managing gas across the withdrawal queue.
One thing to watch: GasFree, which lets users pay USDT transfer fees in USDT with no TRX at all, is live and growing, and it belongs in the same conversation as gasless transactions generally.
TRON cut its fees by 76% in two years to defend its position in stablecoin transfers. The exchanges that noticed have already rebuilt their gas models. The rest are still quoting 31,895.
Verify before publishing
Every figure above moves. Re-run these at publication time:
curl -s https://api.trongrid.io/wallet/getchainparameters
curl -s -X POST https://api.trongrid.io/wallet/getaccountresource \
-H 'Content-Type: application/json' \
-d '{"address":"<any address>","visible":true}'
