Australia is positioning itself as a potential hub for artificial intelligence (AI) computing, leveraging its abundant renewable energy resources amid growing global demand for data center capacity. While currently ranked in the middle tier worldwide with approximately 1.4 gigawatts (GW) of AI compute capacity installed, the country is witnessing heightened interest from developers eager to capitalize on its advantages.
This week, AI company Anthropic signed its first agreement to lease space in an Australian data center, anchoring a large-scale A$30 billion (US$21 billion) project near Brisbane. Meanwhile, in New South Wales—the nation's most populous state—ongoing discussions focus on plans that could add as much as 13 GW of data center capacity, a figure exceeding the state’s current daily electricity consumption. According to the country’s grid operator, data centers might consume up to 12 percent of Australia’s electricity supply by 2050, a significant increase from the current 2 percent.
Despite this potential, the development of new data centers has encountered resistance in many regions, with community objections delaying or blocking projects. This phenomenon has been dubbed “BANANAs,” an acronym for “build absolutely nothing anywhere near anything,” reflecting public opposition to large-scale infrastructure developments. Such challenges are not unique to Australia, where developers must balance the promise of economic growth against local concerns.
The economic benefits of expanding data center infrastructure could be substantial. Analysts at the Commonwealth Bank of Australia estimate that achieving these expansions would require investments approaching A$100 billion by 2030. With the country's broader consumer economy experiencing signs of strain, data center growth offers an alternative avenue for economic activity.
Australia's renewable energy capacity underscores its appeal. A 2024 study assessing 220 countries concluded that the energy required by data centers by 2050 would constitute less than 1 percent of Australia's vast solar and wind potential. Renewable electricity in Australia is also among the most cost-effective globally, with solar generation costs around $40 per megawatt-hour, comparable to rates in Saudi Arabia and approximately 35 percent cheaper than in the United States, according to the International Renewable Energy Agency.
The strategy involves channeling renewable power into data centers, which would then sell AI compute services internationally. While this model suits tasks such as training machine learning models—which involves extensive but less latency-sensitive computation—it may not support applications requiring real-time responsiveness, such as autonomous vehicles.
Australia is concurrently undertaking a significant transition from coal-fired electricity to renewables. By expanding data center infrastructure to meet rising AI demand, the country aims to distribute the financial and logistical challenges of this energy shift more broadly. In this way, Australia’s abundant green energy resources could become a key asset in the global AI ecosystem.
