What JustLend energy rental really costs
JustLend DAO holds the deepest pool of rentable energy we have measured — when we probed its ceiling on 2 September 2026, one call could have delegated stake worth roughly twenty-six billion units of energy. A JustLend energy rental is quoted as a rate per second of occupancy, and that quote is not the bill. On the same day we rented a position with our own money, spent the energy on a real USDT transfer, returned it, and read what the contract charged out of the chain. Here is the arithmetic.
The bill has two terms, and the rent is the small one
There is no fixed term in the contract. An order is a triple — payer, receiver, resource type — carrying an amount of delegated stake and a deposit, and rent accrues per second against a global index for as long as the position stays open. The rental duration a renter picks in the interface is never written on-chain at all: it is an input to the prepay arithmetic, and beyond that only headroom before the position can be liquidated. What crosses to the receiver is an ordinary delegation, the mechanism in the guide to TRON energy.
The second term is the one the rate says nothing about. Energy that has been spent does not return to the pool when the delegation ends — the stake behind it produces nothing until it regenerates, which on TRON takes a day — so the contract bills for that separately, at settlement. For a position returned at once — which is what a per-transfer product does — the charge is a ratio constant of 0.75 times one full day of rent on the position, scaled by the share of the delegated energy actually spent. How that charge behaves if you wait instead is the one thing our order could not settle. Return a position with everything spent and you owe three quarters of a day's rent no matter how briefly you held it. Hold energy and never touch it and you owe only the seconds.
The prepay, and what a guaranteed refund guarantees
Renting is paid in advance. The prepay covers the chosen duration plus one whole day plus a liquidation threshold, with a fee on top, and the contract's own minimum corresponds to a duration of zero — what it insists on is a day of rent and the fee, not a term. Overshooting is safe but not free: the excess is not handed back at rent time, it joins the deposit and comes back when the position closes.
Beside the prepay the interface shows a "refund at least" figure. It promises the fee back plus the quarter of a day that the usage charge cannot reach — a floor rather than an estimate, and on our own order the chain paid back slightly more than the screen had said it would.
The deposit is also what stands between a forgotten position and liquidation. Liquidation is permissionless: anyone may close a position whose deposit has run down and keep the keeper's fee, which is the larger of a flat floor and a small fraction of the delegated stake. It is rare — three liquidations against more than seventeen hundred rentals in the day we sampled — but it is somebody's running business, and on that day at least two independent keepers were doing it. The usage charge applies to a liquidation exactly as it does to a voluntary return.
What one real order cost
On 2 September 2026 the operator rented 131,000 units of energy by hand, delegated them to a second wallet, sent one USDT transfer with them to an address that had never held USDT, and returned the position 144 seconds after the rent. The transfer spent 130,285 units and burned nothing of its own: the delegation did the job it was bought for. Every figure below is from the two transactions' own event records.
| The position, 2 September 2026 | As the contract recorded it |
|---|---|
| Energy rented | 131,000 units |
| Held | 144 seconds |
| Energy actually spent | 130,285 units |
| Prepay sent | 29.062504 TRX |
| Occupancy rent accrued | 0.014032 TRX |
| Usage charge on return | 6.269262 TRX |
| Refunded on return | 22.77921 TRX |
| What the position cost | 6.283294 TRX |
The prepay minus the refund is exactly those two charges added together, with nothing else in between. And they are nothing like each other: the rent for those 144 seconds came to about one part in four hundred and fifty of the bill, and the usage charge was all the rest. Spread over the energy rented, the position came to a little under half of what the transfer would have burned — the network prices energy at 100 sun a unit, and a transfer to an address with no USDT history burns 13.1 TRX.
Twenty-six more positions, from the public record
One order of ours is one order. Every rental on that contract is public, though, so we paired the twenty-four hours of its own event feed ending 2 September 2026 — 2,621 calls, of which 1,758 rents, 860 returns and 3 liquidations — and pulled the full bill for 26 complete positions, from 199,964 units up to just under 200 million. On all 26, as on ours, the two charges summed exactly to prepay minus refund. There is no third fee.
What those 26 show is that one product carries two prices, and which one you get depends on what you do with the energy. In twenty-one of them the energy went almost untouched, and the bill was between four tenths of a sun and a little over five sun for each unit rented — occupancy rent, essentially the rate. The five that were spent paid between roughly a sixth and roughly a half of the burn rate per unit. Duration had almost nothing to do with it: one of them held a position for three seconds, spent it, and was billed just under a quarter of a day's rent.
The rate moved more than we expected. Pool utilisation swung from around nine parts in a hundred to forty-two and back inside that same day, and the price per second followed it: the day's high was about half again its low. A rate cached for a minute is fine. A quote extrapolated across a hold of several hours is not.
When it is cheap, and when it is not
Cheap for energy you hold and do not spend. Not cheap for energy you spend: roughly half the burn rate per unit, near enough whatever the hold. That is the finding, and it is arithmetic rather than opinion — the usage charge does not decay with a shorter rental, so the one lever a per-transfer product has is not connected to anything.
Against the sellers we already buy from, a single spent transfer on it lands between not quite twice and almost three times what we pay them — and that is with a wallet staked well enough to cover the two calls' own energy, which at our daily flow would mean freezing capital of the same order as the whole treasury. On our market table for 7 September 2026, JustLend DAO's computed one-hour rate sat in the middle of the field the day this was written — above the cheapest sellers, below the burn rate. Our own price for a delivered transfer moves with the time of day.
Where it is not in the middle is capacity. The bulk window we priced — a few hundred million units for an hour — is about one part in eighty of what a single call can delegate, and the sample contains a third party holding nearly 200 million units for close to three hours. Nothing else we route through is within two orders of magnitude of that. The catch there is capital rather than price: an hour-long window on that scale wants a prepay in the tens of thousands of TRX, of which only about a third comes back at settlement, because the usage charge on a fully spent window takes the rest. It is a position to size, fund and watch — a different job from buying a transfer's worth of energy by API.
What we did not measure, plainly: a liquidation of our own, an unstaked wallet of our own, and anything past that single order of our own money — and because the position was returned seconds after the spend, we could not tell apart the two readings of how the usage charge decays while spent energy regenerates.
What the same volume comes to here, quoted before anything is reserved: POST /v1/estimate in the API reference →