Ask five teams how to buy TRX energy and you get five answers, each internally consistent and each computed against a different workload. That is why procurement comparisons in this category rarely settle anything. The numbers are not wrong; the assumptions behind them are not shared.
So this comparison of the ways to buy TRX energy fixes the workload first and holds it constant across five procurement models. Every model is then priced on the same energy requirement, the same bandwidth requirement, and the same failure exposure. Where a model wins, it wins on the same arithmetic everyone else was measured with.
The workload, stated once
Five thousand USDT payouts a day. Recipient mix of 85% existing holders and 15% fresh wallets — deliberately lower than a consumer wallet’s mix, because payout desks usually pay repeat recipients.
Average energy per transfer is 0.85 × 64,285 + 0.15 × 130,285 = 74,185. Across 5,000 transfers that is 370,925,000 energy a day. Bandwidth is 345 bytes per transfer, so 1,725,000 bytes a day, against a free allowance of 600 — the free allowance is a rounding error at this scale.
Chain parameters as read on 8 November 2026: energy burns at 100 sun per unit, bandwidth at 1,000 sun per byte, staking yields 9.70 energy and 1.62 bandwidth per TRX per day. TRX at $0.3197. Every figure below is derived from those, and every figure is a daily cost unless labelled otherwise.
Model one: burn TRX, which is what happens by default
When you buy TRX energy this way there is no supplier, no contract and no capital. Every transfer burns at the network price.
Energy: 370,925,000 × 100 sun = 37,092.5 TRX. Bandwidth: 1,725,000 × 1,000 sun = 1,725 TRX. Total 38,817.5 TRX a day, or $12,410. Per delivered transfer that is 7.76 TRX, about $2.48.
This model has exactly one merit and it is not a small one: it cannot fail. There is no delivery, no delegation lag, no partial fill, and no counterparty. For a desk sending fifty transfers a day, that reliability is worth more than the discount.
Model two: spot rental
Buy TRX energy on demand at the prevailing market price. At 55 sun the energy line becomes 20,400.9 TRX a day. Bandwidth is unchanged at 1,725 TRX, because nobody rents bandwidth at retail.
Total 22,125.9 TRX, or $7,074. Per delivered transfer: 4.43 TRX, about $1.42. That is a 43% reduction against burning, achieved by a single decision.
What the model buys with that 43% is operational exposure. Spot rental introduces delivery latency, price variance, and the requirement to size orders against a forecast that is wrong whenever the recipient mix shifts. It also introduces the failure tax: at a 1.5% failure rate, 75 transfers a day fail and are retried, consuming 5.6 million energy that delivers nothing.
Model three: term contract with a volume commitment
A supplier offers to sell you TRX energy at 40 sun against a commitment to 350 million energy a day, take-or-pay. The headline is 27% below the spot price and looks like the obvious next step.
Energy at commitment: 350,000,000 × 40 sun = 14,000 TRX a day, 20,725 with the shortfall priced in if actual volume runs at 300 million and you pay for 350 million — because 350 million at 40 sun is 14,000 TRX, but the energy you actually used was 300 million, so the effective rate on used energy is 46.7 sun.
At full utilisation the model delivers 14,000 + 1,725 = 15,725 TRX a day, $5,027, or 3.15 TRX per transfer. At 86% utilisation it delivers 20,725 + 1,725 = 22,450 TRX, worse than spot. The contract does not have a price; it has a range, and the range depends on a forecast.
The commitment model is correct for desks with contractual volume — a payroll processor with a fixed client book — and wrong for desks whose volume tracks merchant activity.
Model four: stake and self-supply
Own the capacity: buy TRX energy by buying the stake that produces it. At 9.70 energy per TRX per day, 370,925,000 energy requires 38,240,000 TRX staked — at $0.3197, about $12.23 million of capital. Bandwidth staking needs another 1,064,815 TRX, about $340,000.
The operating cost is not a fee; it is the cost of capital. At 8% a year on $12.57 million, that is $1.0 million, or $2,746 a day. Per delivered transfer: $0.55 before operations, compared with $1.42 for spot rental.
Two adjustments keep the comparison honest. First, staked TRX still earns voting rewards, which offset part of the capital cost but are not guaranteed and vary with network conditions. Second, staked energy is fixed per day: if volume doubles, the desk does not buy more energy, it burns TRX at 100 sun for the overflow.
Model four is the cheapest on paper by a wide margin and the least accessible in practice, because it converts a monthly expense into a $12 million balance-sheet position that most operators cannot justify to a treasurer.
Model five: aggregator or API resale
Buy TRX energy through a layer that routes orders across many suppliers and exposes one API. The spread over the underlying spot price is the product: typically 6 to 12 sun in this market, more during demand spikes.
At 62 sun the energy line is 23,000.4 TRX a day, 24,725 with bandwidth, $7,905, or 4.95 TRX per transfer — 12% more than direct spot and still 36% below burning.
What the spread buys is reduced operational burden: one integration instead of six, automatic failover across suppliers, and often better partial-fill behaviour because routing is the vendor’s core competence. For a desk without engineers who want to own energy procurement, this is frequently the correct answer even though it is never the cheapest quote.
The five models side by side
| Model | Daily TRX | Per transfer | Capital locked | Main exposure |
|---|---|---|---|---|
| Burn TRX | 38,818 | 7.76 | none | none operational; highest cost |
| Spot rental | 22,126 | 4.43 | none | delivery, price variance |
| Term contract | 15,725–22,450 | 3.15–4.49 | none | volume forecast |
| Self-stake | ~8,600 equiv. | ~1.72 | $12.57M | volume growth, capital |
| Aggregator | 24,725 | 4.95 | none | vendor spread, routing |
The self-stake row needs a footnote, because comparing a capital cost against a fee is the most common error in this analysis. The 8,600 figure is the capital charge at 8% a year plus bandwidth, expressed as a daily equivalent. It is a real cost, and it is also not a cash outflow in the way the other rows are. A treasurer will not treat them as identical, and neither should a cost model.
The hidden line items that change the ranking
Three items sit outside all five models and are large enough to reorder the table.
Bandwidth. 1,725 TRX a day, $16,500 a month, identical in four of the five models. Staking for bandwidth costs about $340,000 of capital and removes it permanently. Nobody markets this because there is no margin in it, and it is the easiest 4% to recover.
Failures. Each percentage point of failed transfers at this volume costs 3.7 million energy plus support time. Between a well-run supplier and a badly-run one, the difference is often larger than the gap between 55 sun and 62 sun.
Treasury operations. Prepaid balances, reconciliations, and the working-capital float of paying a month ahead. A model that saves 12% on energy but requires $400,000 of idle prepayment may be worse than the one that saves 8%.
Which model wins at which volume
The ranking of ways to buy TRX energy is not stable across scale, and the crossover points are worth remembering.
- Under 50 transfers a day: burn. The absolute saving from renting is a few dollars, and the operational cost of any supplier relationship exceeds it.
- 50 to 500 a day: spot rental or an aggregator. Capital is not justified, and the saving is real — around 40%.
- 500 to 5,000 a day: spot rental with a second supplier, or a term contract if volume is contractually stable. This is where procurement discipline pays for itself.
- Above 5,000 a day, stable: self-staking enters the comparison seriously, provided the capital has a cost below roughly 19% a year and the volume is genuinely stable.
- Above 5,000 a day, volatile: stake for the base load and rent the peak. The hybrid dominates both pure models, because it uses capital for the part of demand that is predictable.
For how to compare quotes once you have chosen a model, the wholesale procurement guide covers the contract terms that matter. The integration side, if the chosen model has to run inside a payout pipeline, is in wiring TRX energy into a payment pipeline. A worked month at this exact volume is in the monthly teardown, and the assumptions behind the cost model itself are in our guide to buying TRX energy at scale.
A decision procedure in five steps
- Measure the recipient mix over thirty days before you buy TRX energy at a new volume. Everything downstream depends on it.
- Compute the energy requirement with an 8% buffer. Under-buffering is the most common cause of failures that get blamed on suppliers.
- Price the requirement in all five models, using the same chain parameters on the same day.
- Add bandwidth and failure cost to every model, not just the chosen one.
- Choose by cash cost and operational capacity together. The cheapest model that your team cannot operate is the most expensive one you will actually run.
Sensitivity: what would break this ranking
A comparison that only holds at one set of assumptions is a snapshot, not an analysis. Four variables move the ranking, and knowing which ones matter is more useful than the headline table.
Recipient mix. Moving from 15% fresh wallets to 30% raises average energy per transfer from 74,185 to 84,085, a 13% increase in every model. The ranking does not change, but the absolute gap between burning and renting widens, because the expensive transfers are where the saving is largest.
Rental price. At 45 sun spot rental pulls further ahead of the aggregator; at 75 sun the gap between spot and a well-utilised contract widens enough that commitment starts to look attractive again. The self-staking line is unaffected by rental price, which is precisely why it becomes more attractive as the market tightens.
Cost of capital. The self-staking row is the only one sensitive to it. At 5% the capital charge drops to about 1,720 TRX a day and staking dominates everything. At 20% it doubles and the model stops being interesting for most operators. This single input can reorder the table more than any market variable.
Utilisation of a commitment. A term contract at 100% utilisation is the best non-capital model. At 80% it is worse than spot. Since utilisation is a forecast, the contract’s position in the ranking is a statement about forecasting skill rather than about price.
The variable that does not move the ranking much is TRX price. It scales the dollar cost of every model roughly in proportion, so the ordering survives a 30% move in either direction. What it does change is the size of the capital commitment in dollars, which is the number a treasurer will react to.
Frequently asked
Why is it cheaper on paper to buy TRX energy through staking? Because the capital cost is spread over years while the fee is paid monthly. The comparison is legitimate only if the capital genuinely has an 8% alternative use.
Does the ranking change if TRX price moves? Less than expected. A lower TRX price reduces the staking capital requirement and the dollar cost of burning proportionally, so the relative ordering is fairly stable. What changes is the dollar sensitivity of the staking decision.
When you buy TRX energy through an aggregator, is it always more expensive? Per unit, usually. Per delivered transfer including operations, sometimes not. The spread is buying operational capability, and the question is whether you already have it.
What is the most common modelling error? Comparing a rental quote against the burn price without adding bandwidth. It understates the all-in cost by roughly 4% in every model and makes the rental discount look larger than it is.