AI Datacentres: A Growing Energy & Water Conundrum
The rapid expansion of AI workloads has pushed datacentre energy consumption and cooling water use to new levels. In Australia, where hyperscale facilities are clustering near renewable resources, forecasts show rising electricity demand and mounting pressure on local water supplies for evaporative cooling. That trend risks higher emissions and makes meeting climate targets harder unless new approaches are adopted.
Bridging the Gap with Renewables & Energy Storage
Industry action is visible. Large power purchase agreements and onsite solar projects are common. Organisations such as the Clean Energy Finance Corporation and Data Centres Australia promote low-carbon builds. Yet renewables alone do not solve intermittency or simultaneous peak demand across regions. Without dedicated storage capacity, datacentres still rely on fossil firming or unstable grid imports. Storage provides the firming, frequency response and time-shifted energy required to run AI reliably while keeping marginal emissions low.
Crafting Smart Policy for AI Energy Systems
Policymakers must move from identifying risks to delivering specific frameworks that integrate AI, storage and grid operation. Recommended measures include:
- Storage capacity targets and capacity credit rules that recognise storage paired with datacentres as firm resources.
- Financial incentives and streamlined permitting for colocated storage, grid upgrades and dry cooling technologies that reduce water demand.
- Market access for aggregated datacentre loads and virtual power plants so facilities can provide demand response and grid services.
- Standardised emissions and water-use reporting to prevent overstated green claims in procurement.
- Regulatory sandboxes to trial AI-driven energy management that coordinates forecasting, workload scheduling and storage dispatch for lowest carbon operation.
AI itself can optimise cooling, forecast renewables and shift compute to low-carbon windows, but it requires policy to unlock grid integration and revenue streams for flexibility. A targeted AI energy systems storage policy will turn AI from a driver of demand into a tool for resilient, low-emission infrastructure.
