Australia sits at a crossroads where its ambition to become a renewable energy leader collides with a new digital export opportunity. Past hopes such as large-scale green hydrogen and marquee subsea cable projects have stalled or remained speculative, but energy-hungry “AI factories” - massive data centres hosting advanced language models - are suddenly very real, very well funded and very interested in setting up shop in Australia.
Right now, the country has a modest but fast-growing data centre sector built over the past two decades to serve domestic users from government and business storage to streaming, email and video calls. In the last five years, capacity has more than doubled to around 1.5 gigawatts as global cloud platforms expanded local footprints. Global AI and cloud companies backed by heavyweight investors and attracted by Australia’s security partnerships and access to US-made chips see this as a now-or-never moment to lock in large sites and long-term energy arrangements outside an increasingly stretched US grid.
The catch is electricity. Australia has not yet built the large-scale storage, transmission and firming capacity needed to match its solar and wind potential, and coal power is now expected to stay online roughly a decade longer than earlier forecasts. That is before adding the enormous load of AI-focused data centres, with some individual projects topping 500 megawatts and too big to rely solely on the existing grid. In response, policymakers are quietly exploring whether tech giants could help finance or underwrite more than $100 billion in new renewable generation, batteries and transmission well beyond their own usage, in exchange for planning certainty, faster approvals and national coordination.
If that bargain works, AI data centres could accelerate grid upgrades, revive a slow investment pipeline for large-scale renewables and build a new export category where Australia sells “compute” or AI processing capacity into the broader Asia Pacific region and beyond. Consulting analysis suggests an AI hub strategy could add around $80 billion to GDP by 2030 and support roughly 100,000 jobs across construction, operations and supply chains, although those estimates depend on how quickly projects move and how communities respond to large industrial “sheds” and how national rules evolve on issues such as data use and copyright for AI training.
Major independent data centre developers are already planning or building sites an order of magnitude larger than anything operating a few years ago, often in locations with spare grid capacity or near unused substations. Some in the sector argue for locating these AI factories close to renewable energy zones in states such as NSW, Victoria and Queensland while others point to regional industrial hubs like the Hunter Valley, Geelong and Wollongong as a better compromise between power access, construction cost, workforce availability and low latency links to major cities. Either way, the next generation of facilities looks set to bundle their own big batteries, dedicated transmission lines and possibly gas-fired backup, which will blur the line between data infrastructure and energy infrastructure.
What happens next seems to depend on whether Canberra decides to treat AI data centres as strategic national assets or just another commercial land use. Global tech providers are pushing for faster planning pathways and more permissive rules on training models with local content while policymakers are weighing up tax revenues, employment and export potential against higher energy demand and regulatory trade-offs. If Australia leans in, AI factories could become the country’s next big export story after resources, and if it hesitates those same projects may simply land in competing Asia Pacific markets racing to host the world’s AI workloads.
