The data centre boom: $150 billion of construction that never appears in the infrastructure pipeline

Australia has around six gigawatts of potential data centre capacity in its development pipeline — roughly four times the operational capacity recorded at the end of 2025. The Commonwealth Bank estimates the build-out could total around $150 billion by 2030.

For comparison, the entire Major Public Infrastructure Pipeline — every road, rail line, hospital and transmission project across all states over five years — is $242 billion.

Almost none of the data centre work appears in it, because it is private capital. Which means any contractor reading only the public pipeline figures is materially underestimating the market they are competing in.

The numbers

• ~6 GW of pipeline capacity nationally, about 4x the operational capacity at end-2025
• ~$150 billion build-out to 2030 (Commonwealth Bank estimate)
• A separate estimate puts the investment pipeline above A$155 billion, equivalent to 5.6 per cent of annual GDP

• Data centres are estimated to add around 6 percentage points to real business investment growth in 2026, and around 5 percentage points in 2027
• Across Asia Pacific, the development pipeline reached a record 26.5 GW in the first half of 2026, adding 7.1 GW in six months

Where it is concentrated

New South Wales carries the largest pipeline: 44 facilities in development totalling 11.4 gigawatts — equivalent to the output of nearly four Eraring coal stations. During 2025 the state approved or received state significant development applications for a further 22 facilities totalling 3.67 gigawatts.

Victoria is smaller but growing fast. AirTrunk's MEL2 will deliver more than 354 MW at an investment above US$3.57 billion; combined with MEL1, Melbourne capacity exceeds 630 MW and total Victorian investment passes US$5.02 billion.

Microsoft's Australian data centre program, valued at around $25 billion, was reported as progressing in June 2026 and is among the largest single digital infrastructure commitments in the country.

The real constraint is transmission, not construction

This is the part contractors most often get wrong when assessing the sector.

Securing a grid connection for a new data centre in Sydney or Melbourne can take two to three years, with transmission projects queuing behind long-lead transformer deliveries and substation upgrades. The binding constraint has shifted from generation capacity to transmission capacity, and to how quickly new connections can be approved and built.

The practical consequence: announced capacity and buildable capacity are different numbers, and the gap between them is measured in years. A pipeline of 11.4 gigawatts in New South Wales does not mean 11.4 gigawatts of construction work is imminent. It means a queue.

For anyone pricing or resourcing against this sector, connection status is a better indicator of timing than announcement value.

The rules changed in July

In July 2026 the federal government announced that large-scale data centres would carry a legal obligation to meet their own energy needs by underwriting new renewable generation. Future facilities would be required to pay their full share of grid connection costs and to put at least as much energy into the grid as they take out.

That is a material change to project economics, and it has construction consequences:

Generation becomes part of the project. A data centre that must underwrite new renewable generation is no longer a single building on a single site. It is a building plus a generation asset plus a connection — potentially three separate construction programs with different contractors, different approvals and different timelines.

Full connection cost changes feasibility. Where connection costs were previously shared or socialised, some sites will not stack up. Expect pipeline attrition, concentrated in locations with weaker network capacity.

Earlier projects may be advantaged. Facilities already through approval under the previous arrangements sit in a different position to those behind them.

Separately, on 27 March 2026, fifteen data centre projects worth A$51.9 billion were endorsed for prioritised government support through the approvals process.

What this means for contractors

The trades overlap almost completely with infrastructure work. Bulk earthworks, concrete, structural steel, high-voltage electrical, mechanical services, fire, and controls. These are the same crews the metro projects, hospitals and transmission lines are bidding for. The data centre pipeline is not a separate labour market; it is additional demand on the existing one.

Electrical and mechanical capability is the differentiator. A data centre is, in construction terms, a large shed wrapped around a very large electrical and cooling installation. The building envelope is the straightforward part. The specialist M&E scope is where the value and the risk sit.

The clients behave differently to government. Hyperscale developers procure on speed and certainty rather than lowest price, and they repeat. A contractor who delivers one facility well is positioned for the next, which is a materially different commercial dynamic to competitive public tendering.

It is concentrated in fewer hands than the public pipeline. A small number of developers and a small number of head contractors account for most of the capacity, which makes relationships with those parties more decisive than tender monitoring.

What to watch

1. Connection queue movement, which is the actual determinant of when work starts.
2. Pipeline attrition following the July energy obligations — how many announced projects quietly disappear.
3. Whether the renewable underwriting requirement generates its own construction pipeline in generation and transmission.
4. Transformer and switchgear lead times, currently a constraint across both this sector and public transmission work.

Published 1 September 2026 · Updated 7 September 2026