Australia's government released a consultation paper Friday laying out strict conditions for artificial intelligence companies seeking to build massive data centres on Australian soil—a move that signals Canberra won't let tech giants write their own rules.
The framework demands that companies authorized to run large-scale AI training operations here disclose reportable incidents to Australian authorities, reserve computing capacity for local researchers, and ensure their facilities run on renewable energy while minimizing water consumption. It's a significant assertion of democratic oversight over infrastructure that will shape how Australians interact with AI for decades.
Assistant Science Minister Andrew Charlton framed the approach bluntly: "The Australian Government has been clear about the important objective of growing AI capability in Australia. We welcome investment, but on Australia's terms."
The government hopes to legislate these national standards in early 2027. The consultation period closes October 9, leaving less than a month for public and corporate feedback on what could become binding rules across the sector.
What Companies Must Accept
The consultation paper doesn't yet define what constitutes a reportable AI incident, but it asks whether developers should disclose them proactively, on request, or through public statements. That ambiguity matters: it signals the government's intention to maintain transparency requirements that prevent companies from quietly managing problems.
Data centre operators face mandatory renewable energy targets. The paper proposes that facilities must bring forward new renewable generation to offset their electricity demand, backed by gas, batteries, or hydro to maintain grid reliability. Some proposals would require matching consumption with renewable energy annually or during specific periods using renewable energy certificates.
Water management carries equal weight. Centres must minimize consumption, prioritize recycled or non-drinking water for cooling, and pay their share of infrastructure costs. They'll also need to prepare for drought and water disruptions, meet efficiency standards, and report usage transparently.
Companies training advanced AI in Australia face an additional expectation: supporting local research either through direct payments or by reserving computing capacity for Australian researchers and institutions. It's a straightforward mechanism to ensure that Australians benefit from the infrastructure built on their land.
Even data centres already in development but not yet built may be required to comply retroactively with new rules—a provision that prevents companies from racing to finish construction before regulations take effect.
Industry Response
Belinda Dennett, chief executive of Data Centres Australia, welcomed the framework. "The paper asks the right questions, on energy, water, costs and community, and it rightly recognises that this infrastructure has to be built on terms that work for Australians," she said.
Anthropic, which signed a deal this week with Queensland to use part of a proposed $32 billion data centre in the state, also backed the consultation. David Masters, the company's ANZ head of policy, stated: "We believe that Australia should have agency in how AI is used and developed. This includes clear rules on how the associated infrastructure is built out. We intend to build to meet these standards and look forward to engaging constructively in this process."
OpenAI, Google, and Meta did not immediately respond to requests for comment, suggesting the framework may face pushback from the largest players.
Strategic Leverage
The government's approach reflects a deeper calculation about power and influence. Helen Wilson, deputy secretary of the Department of Industry, Science and Resources, told a parliamentary hearing that hosting AI training gives Australia greater influence over the technology itself. "By having training here, that does give us the ability to help shape the development and deployment of AI," she said.
That's the center of the strategy: Australia isn't simply chasing investment dollars. It's negotiating for sovereignty over how AI develops and operates within its borders, ensuring that the infrastructure serving this technology serves Australian interests—environmental protection, research advancement, and public accountability—rather than purely corporate profit maximization.
The government signaled in July 2026 that it intended national laws and rules about AI and its infrastructure. This consultation paper is the concrete expression of that commitment.
Why This Matters:
Artificial intelligence training requires enormous energy and water resources. Without clear standards, companies would externalize those costs onto Australian communities and ecosystems while reaping the benefits of the technology elsewhere. This framework attempts to reverse that equation: it makes AI companies pay for the infrastructure they use and share the computing power they build with local researchers. It also asserts that Australians should have a voice in how AI gets developed—not through market forces alone, but through democratic governance. The consultation period allows public input, though the government's direction is already clear. Whether these standards actually constrain corporate behavior or become performative will depend on enforcement mechanisms and political will when implementation begins in 2027.