Australia's Energy Operator Warns AI Data Centers Risk Grid Stability

The Australian Energy Market Operator (AEMO) has issued a stark warning: the rapid growth of AI data centers could create serious stability risks for the country’s national electricity grid. According to AEMO modelling reported by The Australian, a single fault near western Sydney could cause about 1,500 MW of data center load to disconnect in milliseconds by 2030. The Victoria–New South Wales interconnector could also reach a critical stress point by 2028 if multiple facilities disconnect simultaneously. Data centers consumed roughly 4 TWh of electricity across the National Electricity Market in FY2025—about 2.2% of total grid demand—and under AEMO’s Step Change scenario, that could climb to 12 TWh (6%) by 2029-30. To better track the sector, AEMO began forecasting data centers as a standalone demand category in August 2025, and the Australian Energy Market Commission proposed new ride-through requirements for large loads in March 2026.

The warning is not a call to halt AI development but highlights a coordination problem: developers need to engage grid planners before locking in site selection, leases, and construction schedules. For cloud buyers and CIOs, power constraints could dictate where providers expand AI capacity and how quickly services become available. Operators must be ready to show realistic demand profiles and explain load flexibility during grid stress. For APAC infrastructure investors, grid coordination now sits beside land cost, tax incentives, connectivity, chip supply, and construction timelines in site due diligence. The pressure extends beyond Australia, as large projects like Meta’s first India AI data center deal with Reliance reflect the region’s broader compute hunger. Singapore’s Green Data Centre Roadmap offers a contrast by tying new capacity to efficiency standards, while Australia’s warning shows how quickly AI infrastructure can become a grid-planning problem.