Guardian Analysis Warns Australia’s Data-Centre Boom May Compete With Housing for Builders

Australia’s data-centre boom is taking a much larger share of new building work even as housing construction accelerates. In a newly published Guardian analysis, economist Greg Jericho finds that the category covering data centres reached 20% of non-residential building starts by value. He argues that the boom could compete with housing for available builders.

The construction category behind the surge

The Australian Bureau of Statistics includes data centres in “commercial building not elsewhere classified.” That is a broader survey category, rather than a separately named data-centre measure. Jericho uses its sharp increase to describe the scale of the buildout in the latest building-activity figures.

The category usually represents around 1% of the value of non-residential building commencements—projects beginning construction. In the year to June 2026, it reached one-fifth. The distinction matters: this is a share of the value of work starting, not a count of facilities or a measure of completed computing capacity.

Factories offer a second measure of scale

The spending comparison with factories shows how unusual the balance has become. Usually, factory construction attracts almost twice as much spending as the commercial category covering data centres. Over the past year, that relationship reversed: Jericho calculates that spending on data-centre construction was more than 17 times factory-construction spending.

Jericho questions what that allocation will deliver after construction ends. His argument distinguishes the building work itself from its lasting benefits: factories provide ongoing employment, while homes provide somewhere to live. He contends that data centres employ relatively few people once built, making their eventual benefits central to his assessment.

Housing is growing, not standing still

The housing figures complicate a simple crowding-out story. Residential building volume rose 8.7% in the year to June, its strongest increase since 2016. Total building volume grew faster still, rising 10% over the year. Both residential construction and the data-centre category were expanding, rather than one rising alongside an overall housing decline.

New private-sector house starts also rose 11.6% to 31,707. Jericho describes that level as the highest since 1994, excluding the Covid homebuilder surge and the unusually high March 2000 starts ahead of the goods and services tax. Public and social housing, however, represented just 2.3% of residential building in the year to June.

Jericho’s resource warning is about the capacity to build more, not evidence that housing construction has stopped growing. He argues that housing ambitions depend on the number of builders available. His analysis does not quantify homes delayed or forgone because of data centres, leaving the size of any housing impact unresolved.