The growth of artificial intelligence (AI) is driving profound transformation in global data infrastructure and energy demand. The integration of AI into a range of commercial and industrial products, from refrigerators to virtual assistants, is placing mounting pressure on both domestic and foreign energy networks. With global data centre demand projected to surge 165% by 2030, concerns around the adequacy of existing legal frameworks, energy access, and environmental sustainability have been raised.[1] This article will explore the significant energy demands of AI, the capacity for the NEM and national transmission infrastructure to adapt, and international responses to this evolving demand-side challenge.
What’s the issue with AI’s energy demands?
Data centres are already significant energy consumers, and the rapid uptake of artificial intelligence – particularly generative AI – is accelerating this trend. AI requires immense computational power, placing growing demands on both existing and new data centres. As AI products and services become further embedded in the economy, data centre energy consumption will continue to rise.
Globally, the scale of this demand is striking. For example, ChatGPT alone is estimated to use 39.98 GWh of electricity and 148.28 million litres of water daily for cooling.[2] Meeting the worldwide appetite for AI is projected to require $5.2 trillion in capital expenditure by 2030.[3]
In Australia, data centre development is accelerating. AI-intensive sectors such as finance and telecommunications are driving a substantial surge in demand for high-performance data infrastructure.[4] These facilities require uninterrupted, highly reliable electricity, placing disproportionate pressure on local energy distribution systems. Data centres already consume approximately 4–5% of Australia’s electricity,[5] with projections suggesting this could rise to 8% by 2030.[6]
As the scale and complexity of data centres expand, so too do the regulatory challenges surrounding AI deployment. In response, the Federal Government has introduced the Voluntary AI Safety Standards, encouraging companies to self-regulate across all stages of AI development and implementation.[7] While a positive step, the voluntary nature of these standards leaves room for inconsistency and non-compliance. Introducing mandatory regulations would likely be more effective in ensuring responsible, industry-wide adoption of AI and its supporting infrastructure.
Meeting AI’s growing energy demands will require major investment in both infrastructure and policy reform. Yet AI is evolving rapidly, outpacing the development of domestic and global energy systems. Without timely and consistent investment, energy prices may surge – ultimately leaving consumers to bear the cost of the AI revolution.
International responses
United States of America
The United States has adopted a multi-pronged strategy to manage rising AI-related energy demand, focusing on investment, regulation, and research. Through the Inflation Reduction Act and CHIPS and Science Act, the federal government is directing billions of dollars toward grid modernisation, clean energy expansion, and semiconductor manufacturing critical to AI development.[8] The Department of Energy is also conducting assessments on improving data centre efficiency and developing performance standards in light of growing AI workloads.[9] However, the continued viability of these federal programs remains uncertain under the Trump administration, which has expressed a desire to repeal the Biden-era AI and clean energy initiatives.[10] As a global hub for AI, the United States’ regulatory volatility will influence the industry’s future growth.
China
China’s energy generation development and policy has shifted significantly as AI becomes a cornerstone of the digital economy. China has responded to escalating AI energy demand by implementing its “Eastern Data, Western Computing” policy. This police relocates energy intensive computing centres to western provinces with abundant renewable resources.[11] Also, new data centres must meet stringent energy efficiency and carbon intensity benchmarks, with projects failing to comply denied approval. The policy is designed to balance regional development, reduce pressure on eastern urban grids (which trend toward higher residential concentration), and align AI infrastructure with China’s broader dual-carbon goals: peak emissions before 2030 and carbon neutrality by 2060.[12]
Ireland
Ireland is a case study on AI regulation and infrastructural lag. The nation is facing an unprecedented surge in energy consumption from AI and data centre operations. In 2024, data centres accounted for 21% of the nation’s total electricity use, overtaking all urban residential consumption.[13] This rapid growth has raised concerns about energy affordability, grid reliability, and compliance with European climate targets. In response, Ireland’s energy regulator has proposed a policy requiring data centres to match their energy load with equivalent new power generation.[14] This approach shifts the burden of infrastructure expansion – across generation, transmission, and storage – onto developers, effectively making data centre approval contingent on securing additional supply.[15]
Looking Forward
The energy demands of AI are no longer a future challenge but a present and escalating reality. Countries at the forefront of AI adoption are beginning to experiment with policy tools to avoid grid instability, rising costs, and environmental backsliding. Internationally, a clear trend is emerging as governments shift responsibility for infrastructure upgrades onto data centre developers. While the regulatory mechanisms vary, the common goal is to integrate AI into energy systems without compromising grid security or climate targets.
In Australia, a more coherent, enforceable framework is necessary. Current state-federal fragmentation, coupled with voluntary AI guidelines, lacks the clarity and enforceability needed to support long-term planning. Looking ahead, policy reform will likely focus on mandatory emissions reporting, enhanced environmental approval standards, and greater integration of data centre load into national transmission planning. Australia may also consider following international models by linking new data centre approvals to infrastructure contributions or renewable procurement. If the energy transition is to remain viable in an AI-driven economy, legal frameworks will need to evolve just as quickly as the technologies they aim to govern.
[1] Goldman Sachs, ‘AI to drive 165% increase in data center power demand by 2030’ (Web Page, 4 February 2025) <https://www.goldmansachs.com/insights/articles/ai-to-drive-165-increase-in-data-center-power-demand-by-2030>.
[2] Ian Wright, ‘ChatGPT Energy Consumption Visualised’ (Web Page, 17 February 2025) <https://www.businessenergyuk.com/knowledge-hub/chatgpt-energy-consumption-visualized/>.
[3] McKinsey & Company, ‘The cost of compute: A $7 trillion race to scale data centers’ (Web Page, 28 April 2025) <https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers>.
[4] Bronwyn Cumbo, ‘Australia is set to get more AI data centres. Local communities need to be more involved’ (Web Page, 8 July 2025) <https://theconversation.com/australia-is-set-to-get-more-ai-data-centres-local-communities-need-to-be-more-involved-259799>.
[5] Sherif Andrawes, et al, ‘The rising energy demand from data centres: A global perspective’ (Web Page, 28 April 2025) <https://www.bdo.com.au/en-au/insights/natural-resources-energy/the-rising-energy-demand-from-data-centres-a-global-perspective>.
[6] Ibid.
[7] Australian Government, ‘Voluntary AI Safety Standard’ (Web Page, 5 September 2025) <https://www.industry.gov.au/publications/voluntary-ai-safety-standard>.
[8] Rhett Buttle, ‘Assessing the Impact of the Chips and Science Act and the Inflation Reduction Act’ (Web Page, 16 August 2023) <https://www.inc.com/rhett-buttle/assessing-impact-of-chips-science-act-inflation-reduction-act.html>.
[9] US Department of Energy, ‘Recommendations on Powering Artificial Intelligence and Data Center Infrastructure’ (Report, 30 July 2024) 3 <https://www.energy.gov/sites/default/files/2024-08/Powering%20AI%20and%20Data%20Center%20Infrastructure%20Recommendations%20July%202024.pdf>.
[10] Simmone Shah, ‘How Trump Is Trying to Undo the Inflation Reduction Act’ (Web Page, 28 February 2025) <https://time.com/7262600/how-trump-is-trying-to-undo-the-inflation-reduction-act/>.
[11] Andrew Stokols, ‘Energy and AI Coordination in the ‘Eastern Data Western Computing’ Plan’ (2025) 25(4) China Brief, 6 – 13.
[12] Ibid 6.
[13] Jillian Ambrose, ‘Ireland’s datacentres overtake electricity use of all urban homes combined’ (Web Page, 23 July 2024) <https://www.theguardian.com/world/article/2024/jul/23/ireland-datacentres-overtake-electricity-use-of-all-homes-combined-figures-show>.
[14] Dan Swinhoe, ‘Ireland’s energy regulator proposes policy requiring data centers to match load with new power generation’ (Web Page, 19 February 2025) <https://www.datacenterdynamics.com/en/news/irelands-energy-regulator-proposes-policy-requiring-data-centers-to-match-load-with-new-power-generation/>.
[15] Ibid.
