AI Data Centers Test Limits of Energy and Democracy

The energy and environmental impact of AI and how it undermines democracy – Image for illustrative purposes only (Image credits: Pexels)

Australia now ranks as the world’s second-largest destination for data-centre investment, trailing only the United States. The pace of construction has accelerated so quickly that some planned facilities would draw as much electricity as an entire small city. At the same time, operators have begun exploring new gas-fired plants to guarantee supply, a step that directly conflicts with national targets for cutting emissions and expanding renewables.

Rising Electricity and Water Use Outpace Climate Claims

Recent analyses show AI-related electricity demand could climb eleven times higher by 2030 compared with 2023 levels if current trends continue unchecked. A February 2026 assessment found that three-quarters of industry assertions about AI delivering major climate benefits lack supporting evidence, and no verified examples emerged of consumer tools such as ChatGPT or Gemini producing substantial, measurable emissions reductions.

The gap between promised gains and actual performance matters because the physical build-out itself requires more extraction, more manufacturing and more grid capacity. Chip production alone recorded a four-and-a-half-fold rise in emissions in a single recent year, according to supply-chain tracking in East Asia. These figures illustrate how the infrastructure race is locking in higher resource use rather than offsetting it.

Local Communities Push Back Against New Facilities

Residents near Perth blocked a proposed three-storey, 120-megawatt data centre after concerns surfaced over culturally significant land. In New Brunswick, New Jersey, city officials removed data centres from redevelopment plans following public meetings that highlighted noise, water strain and higher electricity costs. San Marcos, Texas, saw its council reject a similar project by a 5-2 vote after more than one hundred residents spoke against it.

Further afield, South Dublin County Council called for a national moratorium or strict renewable-energy conditions on new centres. In the United Kingdom, campaigners secured the right to challenge a 90-megawatt facility after authorities acknowledged an approval error. These cases show a consistent pattern: communities absorbing immediate costs while distant corporations capture the financial returns.

Corporate Concentration and Political Influence Grow

Nvidia reported annual revenue of 215.9 billion dollars, with data-centre and AI chips now accounting for the vast majority of sales. Its chief executive has described the current expansion as the largest infrastructure project in human history. Greenpeace East Asia rankings placed the company last among peers for decarbonising its supply chain, noting heavy reliance on fossil-powered manufacturing in Taiwan and South Korea.

Amazon recorded more than 77 billion dollars in 2025 profits while reducing its workforce by roughly 30,000 positions amid stepped-up AI spending. Political contributions from leading AI executives, including a one-million-dollar donation to a presidential inauguration fund, have coincided with expanded government contracts for surveillance and defence applications. Such developments concentrate decision-making power in fewer hands and raise questions about accountability.

Paths Toward Accountable Technology

A workable alternative would require new facilities to run exclusively on additional renewable energy and to publish full details of energy, water and material footprints. Governance rules would need to include community consent, limits on monopoly control and clear audit trails so that data and models remain subject to local oversight rather than remote extraction.

Under these conditions, AI tools would be deployed only where they demonstrably serve public needs without exceeding simpler, less resource-intensive options. Privacy protections, freedom from mass surveillance and safeguards against manipulation would form core design requirements. The result would be systems that expand shared capabilities instead of shifting costs onto communities and future generations.