Australian policy analysis of managing unprecedented data center power demand growth amid constrained grid capacity.
New commentary from HSBC highlights lessons history could teach Australia's policymakers in managing the current data centre investment boom, drawing parallels to the liquefied natural gas boom of the early 2000s.
According to HSBC Chief Economist Paul Bloxham, while the comparison is not "perfect," the similarities are striking. Both booms "draw heavily on local energy supplies once completed, are highly import-intensive in the investment phase, have high foreign ownership, employ few workers particularly once operational, and make products—gas and compute—that are growth drivers whether used locally or exported."
Australia is currently experiencing one of the world's largest data centre investment booms. The Australian Energy Market Operator estimates the investment pipeline at AUD 97 billion, representing approximately 3.4% of GDP. Existing capacity stands at 1.4 gigawatts, with applications pending for another 5.4 gigawatts.
The critical question, Bloxham notes, is how much of Australia's economy will actually benefit. Near-term gains appear limited. While business investment has surged, 85% of equipment used is imported—semiconductors, hard drives, and cooling systems predominantly sourced from Asia—providing little immediate boost to local GDP growth. The longer-term outcome depends partly on what these data centres ultimately serve: local purposes or global compute markets remain unclear.
Meaningful economic benefits will hinge on successful AI adoption by local businesses that demonstrably lifts productivity. Progress here has been slow. Most large businesses and government departments have AI strategies in implementation, but adoption among small and medium enterprises—which employ around 70% of the workforce—lags significantly. More than half report no AI use in surveys, while only 6% claim "broad" AI deployment.
The data centre boom has intensified concerns about resource constraints. Industry estimates suggest insufficient available energy and water to support the full build-out. CBRE projects a 40% energy shortfall by 2028. If data centres draw from existing supplies without sufficient additional investment, they will likely drive up prices for energy and water, potentially displacing other industries and raising costs for households and firms. Additionally, data centres will create minimal employment, particularly once operational, and if AI adoption streamlines workforces, job growth could face headwinds.
The historical parallel is instructive. The LNG investment boom amounted to approximately USD 300 billion for Australia over five years. When LNG exports ramped up, GDP growth accelerated. However, with high foreign ownership, much profit flowed offshore. While tax revenues increased, depreciation of capital offset tax liabilities significantly. More striking: LNG sales shifted local gas prices from lower domestic levels to higher global rates, rendering gas-reliant industries less competitive—a dynamic that mirrors data centres' relationship to energy pricing.
During Australia's early 2000s resources boom, vigorous debate surrounded so-called "Dutch disease"—when high investment in one sector crowds out others partly through currency appreciation—and proposals for higher mining taxes to preserve boom-era returns for future generations. Australia ultimately benefited overall through increased exports and tax revenue, but the boom created both winners and losers. In retrospect, key questions persisted about whether tax settings were appropriate and whether more gas should have been reserved for local use. These remain live policy issues today, reflected in current discussions about higher gas taxes and domestic gas carve-outs.
Policymakers should "carefully" consider several issues when forecasting the data centre boom's local economic impact. Beyond the regulatory challenges AI raises, they ought to assess implications for local energy and water, effects on other industries and communities, and whether the benefits of large-scale data centre investment justify the costs. Tax and regulatory policy both matter, and establishing ground rules early carries real benefits.
Energy and water policies emerge as key priorities. There are potential upsides: if Australia could more rapidly become a renewable energy superpower, this could substantially support its emergence as a large and successful global AI player. Much depends on getting energy policy right. Data centres are energy-hungry—they measure capacity in gigawatts for good reason.