Executive Summary
Australia’s accelerating data‑centre boom—driven largely by hyperscale cloud providers and AI compute demand—is reshaping the nation’s energy architecture, resource allocation, and sovereignty landscape. As global technology giants expand their footprint, data centres are becoming the new ‘pipelines’ of the digital age. However, geography can no longer reliably assume sovereignty, especially as this hyper-scaling infrastructure binds Australia’s land, energy, and water systems to foreign‑owned platforms. This raises urgent questions about who controls—and who benefits from—the infrastructure powering Australia’s digital future.
Australia thus faces a structural paradox: mass adoption of AI technologies with limited sovereign control over the infrastructure enabling them. Heavy reliance on foreign‑owned platforms exposes Australia to external decision‑making, supply‑chain vulnerabilities, and environmental pressures that undermine effective policy control.
Introduction
Data centres have become foundational to national productivity, sovereignty and the digital economy. As one of the largest data-centre markets in the Asia-Pacific, Australia’s digital infrastructure boom represents a defining moment for its future. Given its vast landscape and strategic location, Australia has the capability to become a “data centre capital of the world” (Verrender, 2026).
This growth is not exclusively technological. It is also geopolitical and environmental. Data centres are deeply physical infrastructure: requiring vast areas of land, continuous electricity loads, water for cooling, and proximity to transmission networks. As AI adoption accelerates, these centres increasingly shape Australia’s energy system design, renewable‑integration pathways, and regional development patterns. They bind Australia’s critical resources to the strategic priorities of foreign companies,raising questions about national control, democratic accountability, and long‑term resilience.
Despite the National AI Plan framing the investment as central to a more competitive, productive, and resilient economy, the government has yet to implement tougher and more comprehensive AI legislation. Australia’s data centre boom risks being a classic example of political myopic – mass adoption, with Australia “sleepwalking into an AI crisis” (McIlroy, 2026), as technology outpaces policy.
Analysis
1. New Pipelines Yet No Sovereignty
Australia’s abundance of land, natural resources, clean energy stores, and established data centre construction skills have made it possible for it “to become a data centre capital of the world” (Verrender, 2026). Australia already has approximately 162 data centres, with its biggest project, Mamre Road, projected to become the largest data centre in the Southern Hemisphere (Cumbo, 2026).
Data centres function as critical infrastructure. They anchor Australia’s digital economy and enable AI model training, cloud services, and national‑security capabilities (Taylor-Price, 2025; Ciriello, 2026). Their importance is reflected in the scale of investment; with Microsoft’s AUD 25 billion expansion and Amazon Web Services committing AUD 20 billion by 2029, Australia is second only to the United States in data‑centre investment levels (Ciriello, 2026).
However, unlike traditional infrastructure, Australia is the “host” rather than in control of the majority of its data‑centre assets (Ciriello, 2026). Hyperscale providers includingMicrosoft and Amazon own and operate much of the country’s compute capacity. The benefits from constructing new data centres go to foreign investors, given Australia does not currently have the necessary technology (Taylor-Price, 2025; Verrender, 2026). The government’s framework for AI infrastructure governance currently sits alongside existing non-AI specific laws, making the legitimacy of Australia’s authority over the digital domain ambiguous. As data centres become both a necessary and profitable infrastructure, this risks foreign firms or private bodies capturing the gains, while civil society bears the environmental, social, and democratic costs.
2. Energy Architecture Under Strain
Current policy positions Australia as a regional data centre hub to support AI demand, despite increasing concerns that the long-term environmental burdens are being overlooked by the short-term economic gain (Cumbo, 2026). Experts have noted a familiar pattern: “alongside the $200 billion energy transition investment pipeline, Australia is seeing an investment boom, with the total pipeline approaching 13% of GDP, rivalling the size of the (mid-2000s) mining investment boom” (Verrender, 2026).
While investments currently support short-term economic growth, this surge is reshaping Australia’s long-term energy architecture. Data centres require highly reliable grid connections, uninterrupted electricity and water supplies to keep them running and cool (Cumbo, 2026). Data centres need large amounts of water for cooling. In Australia, data centres currently use about 3.5 billion litres per year in Sydney alone—less than 1% of total demand, with predictions this figure will jump to 25% by 2035 (Cumbo, 2026). The source of water is contentious and data centre developers are not required to show projections on how much water they will use (Kaye, 2025; Cumbo, 2026).
Data centres currently account for 2.8% of electricity consumption on Australia’s east coast, but the head of climate and energy economics warns that “by 2030 we expect that to be 7%, rising to above 10% by the mid 2030s” (Commins, 2026). Yet, data centres are already failing to cover their own emissions with renewable alternatives, thereby holding back any major energy transition initiatives (Commins, 2026; Greenpeace Australia Pacific, 2026). Australia’s Mamre Road project is estimated to generate peak annual grid emissions equivalent to that produced by over half a million petrol cars for a year or all domestic flights within New South Wales in 2023 (Commins, 2026; Greenpeace Australia Pacific, 2026).
3. Sovereignty Risks: Foreign Ownership and Policy Dependence
Australia has proven it can get capital. Its capability to establish who controls and benefits from the data centres remains unclear. Digital sovereignty is necessary for competitive advantage in data centre development. Yet, as a global data centre hub, external decision‑making remains a key concern for Australia, especially sinceAustralia lacks a hyperscale domestic tech stack, foreign companies determine where data centres are built, how energy is sourced, and what models are trained (Taylor-Price, 2025).
In response to these concerns, the Albanese Government has begun articulating expectations for data centres, emphasising that they must ‘deliver for Australians’ and operate within national‑interest frameworks (Department of Industry, Science and Resources (DISR), 2025; Chin, 2026). The government has also signalled a new AI reform agenda, including consultations on data‑centre regulation and energy‑system planning for digital loads (DISR, 2025). However, Australia’s proposed AI strategy contradicts existing developments. Growing professional and public concerns highlight a “minimal strategy” despite mass adoption, with policy lagging behind technological change (Budde, 2026). Despite its National AI Plan claiming to be part of a broader effort to revitalise Australian industry and ensure the benefits of technological progress remain domestic (DISR, 2025), current investments and lack of its own hyperscale tech stack risks deepening Australia’s dependence on foreign‑owned infrastructure and losing the ability to shape its own AI future.
4. An Alternative Approach
With the most comprehensive AI governance framework globally, Europe has demonstrated that the key difference is philosophical: while Australia relies on balancing existing and foreign legal frameworks, Europe proactively writes new AI-specific rules (Wondrasek, 2026). It is not voluntary nor a guidance but rather hard law, with penalties up to EUR 35 million or 7% of global annual turnover for the most serious violations (Miller, 2026). Europe also instituted an extraterritorial reach; its risk-based framework applies to any provider placing AI systems on the EU market, regardless of origin. Australia, by contrast, shares similarities to the market-driven model that dominates the U.S., including regulatory norms and governance models that don’t explicitly protect Australian citizens, institutions, or democratic accountability (Budde, 2026). Rather than comprehensive legislation, Australia currently regulates AI through existing legislation, voluntary standards, and self-assessment against current legal principles (AIRiskAware, 2026).
Europe’s Digital Services Act (DSA) presents a more democratic and privacy focused framework that explicitly recognises that technological infrastructure molds public discourse, economic power, and civil society (European Commission, 2022). The DSA recognises the necessity of a public-interest governance and prioritises creating a digital space that empowers citizens and consumers by strengthening the protection of their fundamental rights online (European Commission, 2022).
Policy Recommendations
Australia’s current approach risks locking in a corporate model that prioritises private incentives over national interests, strains energy and water systems, and limits democratic control over AI development. Drawing on insight from Australia’s economic, environmental, and democratic priorities, this brief offers several recommendations for Australia to strengthen its AI governance.
- Shift from reactive to proactive: Governments and policymakers need to be proactive by writing prescriptive rules specifically for AI and digital development. Australia has no AI-specific transparency obligations that require platforms to disclose when content is algorithmically generated, relying instead on general consumer law. The establishment of a national AI system registry, requiring platforms and providers to publish system purpose, risk classification, audit summaries, accuracy ranges and limitations, would improve public trust and support democratic accountability.
- Establish a tiered, risk-based regulatory structure: Australia’s current AI framework is voluntary and its conformity approach is self-assessment against existing legal principles. Risk tiers for AI systems establish a more easily regulated structure for AI governance. Likewise, shifting from a technology-neutral to AI-specific approach for conformity assessment necessitates implementing a third-party human oversight body. Financial penalties should also reflect the tiered structure, rather than reliance on guidance from existing consumer law penalties.
- Promote community-aligned AI infrastructure: Large‑scale facilities reshape regional development patterns and compete with agricultural or conservation priorities. Regarding long-term sustainability, these pressures highlight the need for environment-specific governance frameworks that recognise data centres as resource‑intensive industrial assets—not merely digital infrastructure. Plans should demonstrate clear public-interest value prioritising development that aligns with community needs. Smaller, regionally governed data centres could support essential services without imbalancing the energy grid.
- Extraterritorial obligations for foreign AI providers: To strengthen sovereignty and ensure data centres offer consistent protection and benefit Australian citizens, reforms necessitate requiring foreign AI providers and investors to meet Australian standards and to reduce regulatory arbitrage. Investing in sovereign capability, owned and operated by Australians, would give a competitive advantage and maintain control over their critical data and technology, protect national interests, and avoid dependence on foreign corporations.
Bibliography
AIRiskAware. (2026). EU AI Act vs Australia: Two Approaches to AI Governance and What It Means for Your Organisation. [online] Available at: https://airiskaware.com/insights/eu-ai-act-vs-australia-ai-governance-comparison.
Budde, P. (2026). Australia’s AI paradox: Mass adoption, minimal strategy. [online] Independent Australia . Available at: https://independentaustralia.net/business/business-display/australias-ai-paradox-mass-adoption-minimal-strategy,20664.
Budde, P. (2026). Energy and sovereignty will decide Australia’s AI future, not algorithms. [online] Independent Australia. Available at: https://independentaustralia.net/business/business-display/energy-and-sovereignty-will-decide-australias-ai-future-not-algorithms,20998.
Chin, M. (2026). Microsoft expands AI footprint in Australia with $18 billion investment. [online] CNBC. Available at: https://www.cnbc.com/2026/04/23/microsoft-expands-ai-footprint-in-australia-with-18-billion-investment.html.
Ciriello, D.R.F. (2026). To Avoid US-Style AI Backlash, Australian Data Centres Need Collective Sovereignty, Not Corporate Extraction. [online] Australian Institute of International Affairs. Available at: https://www.internationalaffairs.org.au/australianoutlook/to-avoid-us-style-ai-backlash-australian-data-centres-need-collective-sovereignty-not-corporate-extraction/.
Commins, P. (2026). Thirsty and power hungry: Australia is in the middle of a datacentre boom – but are they good for the economy? [online] the Guardian. Available at: https://www.theguardian.com/technology/2026/jun/22/australia-ai-datacentre-boom-environmental-impacts-climate-how-sustainable-is-it-artificial-intelligence.
Cumbo, B. and The Conversation Digital Storytelling Team (2026). Can Australia build one of the world’s largest data centres? [online] The Conversation. Available at: https://stories.theconversation.com/can-australia-build-one-of-the-worlds-largest-data-centres/.
Cybersecurity and Infrastructure Security Agency CISA. (2026). Five Eyes Cyber Security Agencies Statement | CISA. [online] Available at: https://www.cisa.gov/news-events/news/five-eyes-cyber-security-agencies-statement.
European Commission (2022). The Digital Services Act. [online] European Commission. Available at: https://digital-strategy.ec.europa.eu/en/node/15288/printable/pdf.
Greenpeace Australia Pacific (2026). Energy vampires: The AI data centres draining Australia. [online] Greenpeace. Available at: https://www.greenpeace.org.au/greenpeace-reports/ai-data-centres-draining-energy-australia/.
Kaye, B. (2025). In Australia, a data centre boom is built on fvage water plans. [online] Reuters. Available at: https://www.reuters.com/sustainability/land-use-biodiversity/australia-data-centre-boom-is-built-vague-water-plans-2025-09-15/
Maiden, S. (2026). Trump’s terrifying tech move shocks Albo. [online] news.com.au. Available at: https://www.news.com.au/technology/online/internet/anthony-albaneses-cryptic-reveal-of-ai-reform-agenda-amid-spy-warning/news-story/748f9774beeecd7a33efa619635d1b08.
McIlroy, T. (2026). Australia ‘sleepwalking’ into AI crisis and ‘tech bro free-for-all,’ says Greens senator. [online] the Guardian. Available at: https://www.theguardian.com/technology/2026/jun/23/david-pocock-urges-albanese-to-stop-tech-companies-training-ai-models-using-australian-content.
Miller, M. (2026). Five Eyes Just Issued an AI Warning. The EU Already Made it Law. Where Does That Leave Australia & New Zealand? [online] Insicon Cyber. Available at: https://insiconcyber.com/blog/five-eyes-ai-call-to-action-eu-ai-act-anz-governance
Ministers for the Department of Industry, Science and Resources. (2026). An Australian approach to AI: Expectations for data centres that deliver for Australians | Ministers for the Department of Industry, Science and Resources. [online] Available at: https://www.minister.industry.gov.au/ministers/timayres/media-releases/australian-approach-ai-expectations-data-centres-deliver-australians.
Taylor-Price, R. (2025). Sovereign Digital Infrastructure. Will Australia Seize the Moment? [online] InnovationAus.com. Available at: https://www.innovationaus.com/sovereign-digital-infrastructure-will-australia-seize-the-moment/
Verrender, I. (2026). In the great data centre boom, will the benefits flow offshore again? [online] Australian Broadcast Company (ABC). Available at: https://www.abc.net.au/news/2026-06-07/data-centre-boom-will-australia-benefit/106764460
Wondrasek, J.A. (2026). How Australia’s AI Regulation Compares to the EU AI Act, US Approach, and Other International Frameworks. [online] softwareseni. Available at: https://www.softwareseni.com/how-australias-ai-regulation-compares-to-the-eu-ai-act-us-approach-and-other-international-frameworks/