Data centres are rewriting Australia's electricity demand curve

Advisers who want portfolio exposure to the energy transition usually reach for the utilities and the generators. Claudia Kwan, portfolio manager at NorthStar, thinks the earnings will land further down the chain, with the companies that connect, service and supply the build rather than the ones that own it.

The demand case keeps firming. Electricity makes up around 20 per cent of Australia’s final energy consumption, according to the latest Australian Energy Statistics for 2023-24, and meeting the federal government’s 2035 target of a 62 to 70 per cent cut on 2005 emissions requires electrifying almost everything else.

Deloitte puts the additional investment needed at more than $420 billion on top of business as usual spending.

The Australian electricity demand shock

Data centres have graduated from a rounding error to a genuine load. AEMO’s 2026 Electricity Statement of Opportunities, released in August, forecasts data centre consumption on the National Electricity Market rising from roughly 5 TWh in 2025-26 to 34 TWh in 2035-36 under its central scenario, lifting the sector from about 3 per cent to 13 per cent of grid-supplied electricity. Its high-growth case reaches 52 TWh.

“Data centres are emerging as one of the most important new sources of Australian electricity demand, and the effect is likely to be considerably larger than their current contribution suggests,” Kwan says. “In addition to the initial capex, they consume electricity almost continuously, have very high load factors and are increasingly power dense.”

Weather compounds it. Kwan estimates the current El Niño event alone will add roughly 3 TWh of incremental annual demand to the NEM, and she does not expect that to unwind when the cycle turns.

“Following heatwaves across Europe and a warm winter in Australia, the increased demand for electricity from hotter weather will continue beyond this year’s El Niño,” she says.

Down the chain

A decade of connection work, substation upgrades and maintenance flows through contractors long before it shows up in the returns of the asset owners. Southern Cross Electrical Engineering makes the point cleanly.

“It is the electrician, working as a contractor to the utilities and generators rather than competing with them,” Kwan says. The company takes the work whoever wins the asset, which removes the need to pick the eventual owner correctly.

Tasmea offers a different kind of exposure again, weighted to the servicing end of the same activity. Kwan describes it as “a portfolio of businesses focused on maintenance and repairs, much of it in remote locations, which makes it less a single bet and more a diversified way to own the servicing side of the build-out”.

Heat, copper and water

The thesis reaches past the grid. Heat drives both the energy and the water consumption of a data centre, which turns thermal management into a cost and growth constraint rather than an engineering footnote, and Kwan points to it as the problem every operator and semiconductor company now works on.

Advanced Energy Materials, a chemicals business making high-purity aluminium derivatives used in thermal management, supplies a direct input into solving it.

Copper carries the same logic. “The market has been structurally short of supply for years, and existing mines are approaching the end of their productive life,” Kwan says.

“Substitute technologies already exist, but most are only economical at higher copper prices and as copper rallies, expect substitution technology to move from theoretical to commercial.”

Aluminium and fibre substitution has moved in and out of favour with the copper price for decades, so the trade depends on the price holding rather than on the technology arriving.

Water rounds out the list, and Kwan treats it as the least appreciated of the four. Decades of underinvestment in water infrastructure are now visible at the same moment that extreme weather forces industry and communities to rethink flood risk and water security.

Vysarn, an integrated water company working across use, reuse, treatment and advisory, gives investors a wider exposure to that repricing than a single infrastructure asset would.

“Water, in our view, is the constraint the market is still under pricing entirely,” observes Kwan.

Testing the thesis

The contractor argument depends on the pipeline converting, and AEMO’s own numbers show how often it does not. More than 40 per cent of data centre projects tracked since 2025 have dropped out or gone backwards, and around 36 per cent of the projects on AEMO’s 2025 list have been cancelled outright.

Connection requests total 66.9 GW as at June 2026, a number that measures ambition rather than committed capacity. Contractors trade on order books, and an order book built on announced projects can shrink faster than an operating asset’s earnings.

That history explains part of the market’s hesitation. “Many investors are still not positioned for the structural theme,” Kwan says. “That may be because contractors have historically been viewed as cyclical plays rather than beneficiaries of a multi year, possibly multi decade, structural theme.”

For advisers building portfolios around Australian electricity demand and electrification, the practical question is which part of the chain holds pricing power once the capital starts moving in volume. On Kwan’s reading, the answer keeps landing with the businesses holding the tools.