Why size it top-down at all
Most target lists for remote energy work are built from the bottom up: find the sites, estimate their load, rank them. That is the right way to find a client and the wrong way to find out whether a market exists. A bottom-up list cannot tell you what it is missing, and it has no denominator — there is no point at which it says this is all of it.
National energy statistics can. Australia publishes consumption by industry division and by fuel, and generation by fuel, from the same statistical spine. Reconcile a bottom-up site list against those totals and you get two things a site list alone never gives you: a ceiling, and a named residual. Everything below comes from that reconciliation, for the 2023–24 financial year.
Four steps and the market is gone
The third step is corroborated from the other side of the accounts: national oil-fired generation is reported at 4,823 GWh, or 551 MW average. It is a ceiling — mine sites, remote town stations and factory standby sets are all inside it. The final bar is drawn wider than scale so it remains visible.
Source: Australian energy statistics (consumption by industry and fuel; generation by fuel), 2023–24; national greenhouse and energy reporting register of designated generation facilities. Analysis by Heliovulcan. Conversion from fuel to electrical output assumes 33 per cent thermal efficiency.
Where the other 96 per cent is
| Use | PJ | Share | Form |
|---|---|---|---|
| Road transport | 679.4 | 53.5% | Mobile |
| Mining | 299.6 | 23.6% | Predominantly haul fleet |
| — Coal mining | 151.4 | — | Qld 95.3 · NSW 53.2 |
| — Metal ore and quarrying | 145.4 | — | WA alone 120.3 |
| — Oil and gas extraction | 2.8 | — | LNG plants burn gas, not diesel |
| Agriculture, forestry, fishing | 88.8 | 7.0% | Tractors, harvesters, vessels |
| Rail transport | 52.0 | 4.1% | Mobile |
| Electricity supply | 42.1 | 3.3% | The only generation line in the table |
| Commercial and services | 32.1 | 2.5% | Dispersed small sets |
| Construction | 26.5 | 2.1% | Excavators |
| Manufacturing | 13.9 | 1.1% | Mostly standby |
| Total | 1,269.9 | 100% |
Source: Australian energy statistics, consumption by industry and fuel, 2023–24. Analysis by Heliovulcan.
One line in that table is a power station. The rest is engines that move.
What is left after mining, line by line
Four categories, four closures. This is not a market that competitors have already taken. It is a market that was never there — and knowing the difference matters, because the two call for completely different responses.
The 390 megawatts that is not booked to mining
The one generation line in the diesel table — 42.1 PJ under electricity supply — is larger than the registered town stations can account for. Subtract them and roughly 37.8 PJ is unexplained, equivalent to about 390 MW of average electrical output.
Approach it from the generation side instead and the same figure appears. Oil-fired generation in the three states where off-grid operation is material — Western Australia, Queensland and the Northern Territory — totals 3,779 GWh, or 431 MW. Remove those states' town stations and 387 MW remains.
The explanation is structural. At the majority of Australian mine sites the power station is owned and operated by an independent power producer under a long-term power purchase agreement. Its fuel is reported under the operator's industry classification — electricity supply — not the mine's. The generation is real, it serves mining, and it is invisible in any analysis that reads the mining line and stops there. This attribution is an inference drawn from the accounts, not a field in any register.
It has a corollary worth stating plainly: if most mine power station diesel is booked elsewhere, then mining's own 299.6 PJ is almost entirely mobile plant. Haul trucks, dozers, drills, light vehicles. Which is exactly what the cascade above implies, arrived at independently.
The one door that is not closed
Electrifying the haul fleet does not trim this market. It rewrites it.
| Segment | Diesel today | Equivalent electrical load | Fuel tax credit forgone |
|---|---|---|---|
| Australian mining, total | 299.6 PJ | 3,900–4,650 MW | ~A$3.9bn/yr |
| — Metal ore and quarrying | 145.4 PJ | 1,900–2,250 MW | ~A$1.9bn/yr |
| — Coal mining | 151.4 PJ | 1,970–2,340 MW | ~A$2.0bn/yr |
Source: Australian energy statistics, diesel consumption by industry, 2023–24. Electrical equivalents assume diesel end-use efficiency of 35–45 per cent against an electric drivetrain at 85–92 per cent; fuel tax credit at the full off-road rate. Analysis by Heliovulcan.
Set that against the sector's present position. Australian mining consumes 5,339 MW of electricity in total, and the off-grid mine generation fleet we screen sits at roughly 340–525 MW. Electrifying the fleet would raise mining electricity demand by something like seventy to ninety per cent, and in Western Australian metal mining it would multiply off-grid station capacity several times over.
Two things hold it back, and only one of them is technology.
The first is the fuel tax credit. Off-road diesel in Australian mining attracts a full rebate, so an electric fleet is not competing against the pump price — it is competing against a subsidised price, and every litre displaced gives the rebate back. We have written elsewhere about why any mine energy business case has to be net of that credit; at fleet scale the same arithmetic becomes a national-order number.
The second is timing. Haul truck service life and remaining mine life are often the same order, and a mine whose reserve life will not outlast one more fleet cycle will never buy the fleet. The window, where there is one, is an alignment: fleet replacement due, mine life long enough to amortise, and the power purchase agreement approaching its term. Fleet age is not publicly available site by site. Mine life and contract expiry are. That asymmetry should decide the order in which a screen asks its questions.
That last point is not hypothetical. On one remote gold operation we screened, the site registers zero generation and zero purchased electricity — it runs on diesel plant. Convert its energy use to an electrical equivalent and it becomes an 8.7–10.4 MW average load, implying an installed station of 25–35 MW: several times the size of a real power station at a neighbouring operation in the same group. A screening layer built on who has a power station today classifies that site as no target at all.
A bottom-up site list, tested against a national account built a different way
Heliovulcan maintains a site-level screen of Australian mine on-site generation — 370 sites, with fuel, scale band and grid position established from regulatory filings rather than assumed from industry codes. That screen answers “which sites” but cannot answer “is that all of them”, because a list has no denominator.
The denominator comes from national energy statistics, which publish consumption by industry division and by fuel, and generation by fuel, from the same statistical spine. Building the market from the sites and then testing it against those totals gives two things a site list alone never gives: a ceiling, and a residual that has to be named rather than shrugged off.
One check governs whether any of this counts. Two figures agreeing is only evidence if their inputs do not overlap — a facility’s reported emissions and its regulator-published intensity factor will always agree, because one is built from the other. The 390 MW finding in this note is reported because two calculations reached it from inputs with no shared source: an industry-classified energy balance on one side, contractor project registers and state generation licences on the other.