The rapid expansion of datacentres to power artificial intelligence risks undermining climate targets and straining energy grids, according to economist Nicki Hutley. With over 10,000 active datacentres worldwide and numbers expected to triple, the environmental and economic costs are mounting.
Globally, datacentre investment is projected to reach US$7tn, yet governments are largely taking a laissez-faire approach. In Australia, where 286 centres are active or planned, energy and water consumption is set to triple by 2030. The Queensland government has even resisted federal efforts to phase out fossil fuels for datacentres, raising emissions concerns.
While AI promises benefits like improved medical diagnostics and energy grid optimisation, the costs are often overlooked. Datacentres produce massive waste heat, which in warm climates like Australia exacerbates heatwaves. The Australian Prudential Regulation Authority has warned banks of cybersecurity risks from AI itself.
Economic benefits may also be overstated. Most datacentre equipment is imported, limiting direct GDP gains, and job creation beyond construction is minimal compared to sectors like manufacturing. Hutley argues datacentres should face the same cost-benefit scrutiny as traditional infrastructure projects.
Without regulation, the AI boom could slow the transition to net zero and increase energy bills for households. As the UK also grapples with datacentre expansion, policymakers must weigh the benefits against the substantial environmental costs.
