Australia’s data centre bet rests on some big geopolitical assumptions
Australia, like other countries, is making a very large bet on data centres. Governments see potential for investment, economic growth, and productivity gains from artificial intelligence. Not everyone sees data centres in such a positive light – there exists considerable public debate about whether we have enough electricity, water, land, and community support to accommodate them. In several cases, the public response has been enough to slow or entirely halt a data centre’s development.
However, alongside these concerns sits a less visible issue which could be at least as disruptive – global geopolitical volatility. The potential positives from Australia’s data centres depend on things outside Australia’s control, like advanced chips produced through highly concentrated supply chains, energy technologies sourced largely from elsewhere, and AI models and cloud services governed by US companies and US law.
Currently, Australia’s published plans for data-centre growth address energy, water, land, communities, skills, security, and local capability. Separate critical-infrastructure guidance also recognises risks arising from foreign ownership, offshore services, and supply-chain dependencies. What is less clear in the public policy framework is whether and how these concerns come together, that is, how geopolitical disruption could affect access to the advanced chips, AI models, cloud services, and clean-energy equipment on which Australia’s data-centre ambitions depend. Geopolitical volatility means we need to make that connection explicit.
Start with the computing itself. Data centres need advanced chips, specialised equipment, cloud infrastructure, and frontier AI models. The supply chains behind these capabilities are highly concentrated. Taiwan’s TSMC manufactures the most advanced AI chips, US company Nvidia dominates their design, and Dutch company ASML supplies the specialised lithography machinery needed to produce them. Concentration at several points means that export controls, trade barriers, conflict, or logistical disruption can affect access to the entire system. While hosting data centres in Australia does give greater control over land, buildings, grid connections, physical security, and planning approvals, control of much of the computing inside them remains elsewhere.
The United States already uses export controls to determine who can access advanced computing chips, semiconductor equipment, software, and technical knowledge. Australia is a close US ally, but that does not make future access automatic. US policy will continue to be shaped by its own national-security priorities, assessments of technology leakage, and strategic competition with China. This matters directly to Australia’s productivity ambitions.
Energy creates a second set of dependencies. Data centres require large and reliable supplies of electricity. The government expects new developments to secure additional clean generation or storage, cover the infrastructure costs they impose, use water efficiently, and maintain community support. Proposed mandatory standards will extend across energy, water, and land use. Meeting these requirements will increase demand for solar panels, batteries, grid equipment, and the minerals and components used to manufacture them. China accounts for about 85 per cent of global solar manufacturing capacity and 80 per cent of lithium-ion battery production capacity, with even higher concentration at some stages of those supply chains.
Australia therefore faces an uncomfortable interaction between its objectives. Requiring data centres to support new renewable generation can ease pressure on the grid and help maintain public support. At the same time, it can also increase reliance on a single source at a time when the Australian government is trying to diversify critical supply chains. Domestic production may reduce some external exposure, but it brings its own requirements for capital, technology, infrastructure, environmental approvals, and community consent. Diversification takes time, and in the meantime, access to affordable energy technology will remain sensitive to trade policy, industrial policy, and Australia’s relationship with China.
Public consent is also part of the geopolitical equation. Communities experience the data centre boom through electricity prices, water availability, construction, land use, and decisions about who gets access to constrained resources. The federal government has explicitly made social licence a foundation of its expectations for developers, with projects that best meet requirements relating to the national interest, energy, water, employment, and innovation receiving priority in Commonwealth assessment processes. As the sector expands, public support may increasingly depend on whether communities believe the benefits justify the resources being used, and whether those benefits remain in Australia. Foreign ownership could sharpen that debate, particularly if overseas companies are seen to be receiving preferential access to electricity, water, or land while households and other industries face constraints and costs are borne locally.
These critical dependencies of computing, energy, and public consent are inextricably connected and cannot be managed separately. Measures intended to secure electricity supply can deepen exposure to concentrated technology supply chains. Tighter foreign-investment controls can reduce access to the capital needed for expansion. Greater domestic processing can create new environmental and community concerns. Decisions intended to strengthen one part of the system may change risk elsewhere.
Organisations are adopting AI on assumptions about continued access to models, computing capacity, cloud platforms, electricity, and digital infrastructure. Those assumptions increasingly depend on decisions outside of Australia’s control, like relationships between governments, export-control regimes, concentrated supply chains, and the commercial decisions of a small number of foreign companies.
For organisations, this means treating access to AI as a strategic dependency. Many are building expected efficiency and productivity gains into their plans, often without examining how dependent those gains are on particular providers, models, cloud platforms, or computing infrastructure that are outside of their control. They need to understand where that reliance is concentrated, watch for geopolitical and commercial developments that could change access, cost, or capability, and preserve options to shift providers, models, or ways of working if necessary. The objective is to ensure that an external disruption does not undermine the productivity assumptions on which the organisation’s strategy depends.
Australia’s data centre bet may deliver significant investment and productivity gains. Whether it does will depend partly on energy, water, planning, and community consent. It will also depend on geopolitical conditions that Australia cannot assume will remain constant. Governments and organisations need to understand which critical dependencies underpin their plans, how those dependencies could change, and what options they will need if they do.
