The ISP's sentence
Start with the passage the headline share comes from. The 2026 Integrated System Plan has data centres at "around 2% of today's grid-supplied electricity use", growing at around 25 per cent a year to "almost 10% of the NEM's underlying demand by 2050", an endpoint the document also describes as five times today's share and the equivalent of 20 per cent of today's total demand.
That sentence changes base midway. The 2 per cent is measured against grid-supplied electricity use, the almost-10 against underlying demand, which in Australia is a substantially larger quantity. The document is also unsettled on the multiplier: the 2050 endpoint is "four times the share it has today" in one passage and "five times the share it has today" in three. Both versions compare a share on one base with a share on the other. Held to a single base, the factor comes out differently again: from roughly 2 per cent of grid-supplied consumption in FY25 to 12 per cent by FY50 is about six-fold. That six is my arithmetic on the published shares; neither AEMO document prints it.
Grid-supplied and underlying demand
AEMO's operational, or grid-supplied, demand is the electricity that centrally dispatched generation has to meet. Underlying demand adds the output of distributed rooftop solar, which is consumed where it is generated and never reaches the wholesale market. Underlying is always the larger figure.
How much larger depends on the day. The 2026 ISP records that rooftop solar has, in aggregate, met just over 60 per cent of underlying electricity demand across the NEM on a sunny weekend when industrial demand was low. That is a peak for a single period rather than an annual average.
The same forecast, published two ways
The consumption numbers are stable across the documents. Oxford Economics Australia, in the report AEMO commissioned and publishes, puts Australian data centre consumption at 3.9 TWh in FY25, with the NEM accounting for 98 per cent of it, or approximately 2 per cent of NEM grid-supplied consumption. Under Step Change, growth averages 25.1 per cent a year, reaching 12.0 TWh by FY30, 6 per cent of the NEM's grid-supplied electricity, and 34.5 TWh by FY50.
On the FY50 share, that report says: "Of the 34.5 TWh, 33.8 TWh of this is forecast to be on the NEM, representing 12% of NEM grid-supplied consumption by FY50." Its figure note defines the share as data centre consumption over NEM operational sent-out consumption from the 2025 Electricity Statement of Opportunities, so on that side there is no ambiguity about the base. The ISP states the same trajectory as almost 10 per cent of underlying demand by 2050. Both figures are accurate readings of one forecast; the 10 is measured against the larger base.
| Published statement | Numerator | Denominator | Share | Source |
|---|---|---|---|---|
| Today | ~3.9 TWh (FY25) | Grid-supplied | ~2% | 2026 ISP; Oxford Economics for AEMO |
| FY30, Step Change | 12.0 TWh | Grid-supplied | 6% | Oxford Economics for AEMO |
| FY50, Step Change | 33.8 TWh on the NEM | Grid-supplied | 12% | Oxford Economics for AEMO |
| 2050, Step Change | Same trajectory | Underlying | almost 10% | 2026 ISP |
Each row appears as its document publishes it. The FY50 and 2050 rows are the same Step Change trajectory quoted on different bases.
Backing the two totals out
The two published shares put numbers on the bases themselves. A numerator near 34 TWh at almost 10 per cent implies underlying demand around 340 TWh in 2050. The same numerator at 12 per cent implies grid-supplied demand around 283 TWh. The difference comes to 55 to 60 TWh, about a sixth of the implied underlying total. All of this is arithmetic on the published percentages, with the ISP's "almost 10 per cent" rounded to 10.
| Implied quantity | 2050 value | How it is derived |
|---|---|---|
| Underlying demand | about 340 TWh | A numerator near 34 TWh at almost 10 per cent |
| Grid-supplied demand | about 283 TWh | The same numerator at 12 per cent |
| The difference | 55 to 60 TWh | Consumption that never clears the wholesale market |
Derived here from the published shares. AEMO publishes neither implied total.
The ISP does publish components that make a residual of that size plausible. Businesses are forecast to invest in 15 TWh more rooftop solar by 2050 and 27 TWh more large on-site generation, household grid-supplied energy needs are forecast to fall 44 per cent to 20 TWh even as households consume more, because more of what they consume comes off their own roofs, and business and industry, after drawing on their own rooftop solar, are forecast to need 280 TWh from the grid in 2050.
A proposal for flexible connections
On 23 July 2026 the Clean Energy Council published Powering the digital economy, proposing a Flexible Contracting Framework: new data centres would match their consumption with additional firmed renewables, backed by large-scale generation certificates (LGCs) until 2030 and guarantee of origin certificates after that, on a glide path aligned to each facility's ramp-up. The paper puts current data centre consumption at 4 TWh a year, about 2 per cent of NEM demand, across more than 250 operating facilities, and projects that share reaching almost 10 per cent by 2050.
The matching obligation is written against electricity a facility draws from the grid. The paper's almost-10-per-cent projection is, numerically, the ISP's underlying-demand headline; whether it is meant on that base is a question to settle before setting the share beside the obligation. The practical rule I use: for anything with an obligation or a connection application attached, quote the grid-supplied series (2 per cent now, 6 by FY30, 12 by FY50, all Oxford Economics for AEMO), and keep underlying demand for statements about total energy use across the economy. Mixing them is easy enough that the ISP's own drafting switches base inside a single sentence.
Two counts of the fleet
The facility counts do not line up either. The 2026 ISP states there are currently over 160 operational data centres in Australia, almost half in Sydney and most of the rest in Melbourne, Brisbane and Perth. The Clean Energy Council paper describes more than 250 facilities operating. Neither document defines "data centre" in terms that make the counts comparable, and a count that includes small edge and telco sites will not match one built from grid-connected load. Anyone carrying either count into a model has to pick a definition first.
The FY30 waypoint
The near-term part of the forecast can be checked against outcomes. At the end of the March 2026 quarter the ISP records 11 large data centres, representing over 5 GW of maximum demand, in the connections pipeline, with connections historically taking about two years from application to energisation and ramping to full demand over five to ten years. The nearest checkpoint is FY30: 12.0 TWh at 6 per cent of grid-supplied electricity.
| Year | Consumption | Share of grid-supplied | Status |
|---|---|---|---|
| FY25 | 3.9 TWh | about 2% | Base year |
| FY30 | 12.0 TWh | 6% | Forecast, checkable within this decade |
| FY50 | 34.5 TWh (33.8 on the NEM) | 12% | Forecast |
All three readings on the grid-supplied basis.
The 12.0 TWh at 6 per cent falls due in FY30. I will mark this piece against the actuals when they arrive.