Policymakers and economic commentators are all abuzz about data centres. This week, the Federal Government released its National AI Plan, including plans to boost data infrastructure. A week earlier at the National Press Club, Shadow Treasurer Ted O’Brien pitched in some ideas to support data centre investment. And the latest National Accounts show data centre investment has jumped dramatically this year.
Is Australia about to switch gears and become a data centre leader? And what would that mean for the economy as a whole? Recent strong growth in investment suggests that the economic returns are there for at least some domestic Australian capacity. Geopolitical concerns are one consideration, but more practically important are legal concerns with storage of sensitive datasets in overseas jurisdictions, as well as cost considerations and processing speed lags. These lock up a certain amount of ‘non-tradable’ domestic demand.
A more uncertain question is whether the economic returns are there for massive Australian capacity and a role exporting compute to the ‘tradable’ global market. Do we have a comparative advantage in this industry due to our stable, rules-based institutions, preferred access to global trade in computer chips, and our potential for ample low-cost renewable energy? Or will bottlenecks in energy generation, high labour costs in construction and aggressive investment by other nations mean that Australia has little profitable role to play?
Prior to this year, Australia was well off the pace internationally when it came to data centre investment. Looking at imports of automated data processing (ADP) equipment, which serve as a decent proxy for the scale of data centre development in most countries, the United States was streets ahead. China and other East Asian nations had also seen rapid growth, though there will be some double counting of imports that circulate between these tightly linked nations and – on the flip side – some undercounting of domestic production for own use. Australia only made the top 20 on this measure in 2024 but is expected to rise up the rankings given the surge in investment in 2025 and predictions of further strong growth. Some proprietary data sources already suggest Australia is a leading location for data centre investment. Of course, Taiwan dominates the export of ADP equipment.
So, what should policymakers be doing – and not doing – to help unlock the opportunity of data centre investment in Australia?
The new National AI Plan grapples with this at a high level. Only so much specificity is possible when AI technologies and the global data industry are evolving rapidly. But the plan provides a little more regulatory certainty, which should help encourage AI-related investments.
The Government’s approach resembles the pragmatic, technology-neutral approach to regulation described in e61’s joint policy report with the UTS Human Technology Institute ‘AI, Productivity, and Australia’s Choice of Regulatory Framework’.
Beyond the positive marketing of the Government’s plan, there are a few interesting specifics. Federal, State and Territory Treasurers will meet soon to consider ‘approvals and barriers for data centre projects’ as well as the resource usage of those data projects. In NSW, the Government has already established a new body to fast-track approvals of major investment projects like data centres.
Government can also act to help coordinate big-data infrastructure. However, it should be careful to avoid slowing constructive private deals. A sensible path is to mirror the broader AI strategy. Government should aim to use existing technology-neutral rules, expedite approvals according to those rules, and prune only genuinely unhelpful regulation.
The Government’s $1 billion National Reconstruction Fund allocation for ‘critical technologies’ could also help. But with a growing VC market, which naturally skews towards tech firms, funding must be structured carefully to add value rather than displace private capital.
Another area of policy concern is competition. The ACCC suggests that the cloud computing industry will tend towards anti-competitive behaviour because it can be difficult to change provider and the industry has high barriers to entry, among other things. The Federal Government is rightly focused on improving competition across the economy, and should consider how its policy decisions on data centres and AI can contribute to better competition dynamics.
More broadly, policymakers must recognise that building data centres alone won’t drive Australia’s economic growth. Much of their value is imported – the recent surge in ICT CAPEX has been offset by a rise in ADP equipment imports, which mechanically detract from GDP growth.
Data centre operators are also capital-intensive and relatively small employers. Administrative BLADE data suggests that only around 11,500 workers are employed in the relevant industry class, with construction labour adding only temporary demand that competes with other projects. And despite technological improvements, data centres still pose significant energy challenges that can spill over into energy scarcity for other industries and households.
The limited direct benefits of data centres mean that the bigger prize potentially lies in how much domestic data centre investment can be leveraged to help unlock Australia’s digital infrastructure and encourage digitisation among individual firms. There are links between ICT use and productivity. Data centre investment is a first step on the road, but for substantial productivity gains Australia will need to follow up with investment in knowledge capital like R&D and software. This is all the more important because Australian businesses have fallen behind on technology adoption, with ICT investment running below the OECD average for the decade to 2024.
Later this month, the Productivity Commission will issue its final report into ‘Harnessing data and digital technology’. With this in hand, and in the fresh air of the new year, it will be a good moment for policymakers to work on a richer plan to support productivity growth and living standards in the era of big data and AI.
