AI Inquiry Atlas

An unofficial, independent guide to the public records of the Joint Select Committee on Artificial Intelligence. Not connected with the Parliament of Australia or the committee. Its makers also made a submission (no. 481; see About). Summaries and scores are written by AI and can be wrong: check the original before relying on them.

Submission 288

Data Centres Australia

Submitter group: Industry bodies.

An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.

Open in the interactive guide

Summary

Read the original submission for the full text.

Argues that sovereign AI capability depends on domestic data centres, presenting economic, jobs and export projections. Recommends a national compute ambition, faster approvals and grid connections, workforce support, precinct planning with early community involvement, and non-duplicative national standards. Argues existing frameworks protect households from infrastructure costs, while citing analysis that wholesale electricity pressure depends on supply entering alongside new demand.

Key points

Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.

  1. Data Centres Australia argues that sovereign AI capability requires sovereign AI infrastructure hosted under Australian jurisdiction.

    Our position is straightforward. Australia does not get sovereign AI capability without sovereign AI infrastructure. Models run on compute, compute sits in data centres, and data centres sit in a jurisdiction. The jurisdiction writes the rules of access, and captures the value chain.

  2. Describes projections that operational data centre capacity will more than double to 3.9GW by end-2030, requiring around $42 billion investment.

    Operational capacity is projected to more than double to 3.9GW by the end of 2030, requiring around $42 billion in total investment.

  3. It argues that compute can be an export industry, citing modelling that would require substantial additional construction investment.

    Capturing the full APAC and global export opportunity would require up to $52 billion in construction investment by 2030 on top of that required for domestic demand.

  4. It argues that Australia is losing ground to other markets on approval speed and power availability.

    The two dimensions where Australia is losing ground are speed to build (time through the approval process) and power availability. State planning approval processes and environmental assessment requirements are extending project timelines relative to faster-moving markets.

  5. It argues that existing Australian regulatory frameworks are built to prevent households bearing data centre costs.

    A recurring objection to building this infrastructure is that households will pay for it. The Australian regulatory framework is built specifically to prevent that, and it is materially different from the frameworks in the overseas markets that get cited in the debate.

  6. It argues that worker shortages are the next most likely barrier to capturing the opportunity.

    Worker shortages are the next most likely thing to stop Australia capitalising on this opportunity. Mandala estimates a further 8,300 workers will be needed for data centre jobs by 2030

Recommendations

Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.

  1. Allocate policy responsibility where it can be acted on, recognising that co-location data centres do not own or operate the compute hardware in their facilities.

    This distinction matters in setting policy, because co-location data centres do not own or operate the compute hardware within their facilities and responsibility should fall where it can be actioned.

  2. Respect the water–energy trade-off in policy, recognising water-cooled systems’ generally greater energy efficiency.

    Water and energy involve a genuine trade-off that policy needs to respect. Water is more than 3,000 times better than air at transferring heat, so water-cooled systems are for the most part more energy efficient.

  3. Support precinct planning coordinating energy, water and transport, setting clear amenity expectations and involving councils and communities early, with local benefit discussed at project outset.

    Precinct-level planning that coordinates energy, water and transport provision, sets clear expectations for amenity outcomes and brings councils and communities in early would give residents more certainty and developers a clearer path, and it would put the conversation about local benefit at the start of a project rather than at the end.

  4. Recognise in the Committee's report that domestic data centre capacity is the physical precondition for data sovereignty, Australian-trained domain-specific models and security of critical systems, and cannot be acquired by regulation alone.

    Recognise in the Committee's report that domestic data centre capacity is the physical precondition for data sovereignty, for Australian-trained domain-specific models, and for the security of critical systems, and that this capability cannot be acquired by regulation alone.

  5. State publicly that Australia intends to be the Asia Pacific region's hub for AI infrastructure and back it with a measurable capacity target.

    Australia should state publicly that it intends to be the Asia Pacific region's hub for AI infrastructure, and back that with a measurable capacity target.

  6. Support statutory assessment timeframes for nationally significant digital infrastructure, specialist assessment teams and Commonwealth-state coordination of processes, consistent with the Productivity Commission's 2025 recommendations on priority infrastructure approvals; the submitter says this does not require lowering assessment standards.

    Support statutory assessment timeframes for nationally significant digital infrastructure, specialist assessment teams, and coordination across Commonwealth and state processes, consistent with the Productivity Commission's 2025 recommendations on priority infrastructure approvals. This does not require lowering assessment standards.

  7. Fix grid connection timeframes and support anticipatory investment: progress National Electricity Rules flexibility for large load growth, consistent transmission/distribution load treatment, and joint utility–government precinct-hub planning for shared infrastructure.

    Fix grid connection timeframes and support anticipatory investment. Progress the reforms the network businesses have identified, including greater flexibility under the National Electricity Rules to respond to large load growth, consistent treatment of load across transmission and distribution, and joint utility and government planning to designate precinct- based hubs where shared infrastructure can serve multiple facilities.

  8. Recognise data centre operations as a defined trade pathway, fund regional and metropolitan training in shortage occupations, and set migration settings to fill specialist gaps while the domestic pipeline builds.

    Recognise data centre operations as a defined trade pathway, fund regional and metropolitan training programs in the shortage occupations, and set migration settings that let the sector fill specialist gaps while the domestic pipeline builds.

  9. Include AI compute in Australia's export and trade strategy and consider marketing the low-emissions grid as a competitive advantage to customers facing Scope 2 obligations.

    Include AI compute in Australia's export and trade strategy alongside green iron and other embedded-energy exports, and consider how Australia's low-emissions grid can be marketed as a competitive advantage to customers facing Scope 2 obligations.

  10. Adopt forecasting approaches that account for phantom demand, duplicate applications and ramp-up rates, and work with industry and networks on a consistent national methodology.

    Adopt forecasting approaches that account for phantom demand, duplicate applications and ramp-up rates, and work with industry and networks on a consistent national methodology.

  11. Make national data centre standards non-duplicative: build them on the March 2026 expectations with industry, and ensure they do not duplicate obligations under the National Electricity Rules, NGER, the SOCI Act or state planning frameworks.

    Make the national data centre standards non-duplicative. Build the standards agreed by National Cabinet on the Government's March 2026 expectations, develop them with industry and ensure they do not duplicate obligations already imposed under the National Electricity Rules, NGER, the Security of Critical Infrastructure Act 2018 or state planning frameworks.

Read the original

Open in the interactive guide