OpenAI Australia Data Center Switches Cooling Strategy After Recycled Water Plan Fails

OpenAI’s planned mega data center in Australia will rely on a more energy-intensive cooling system after developers abandoned plans to use recycled water, according to a Reuters report published Thursday. The change comes after NEXTDC, OpenAI’s Australian infrastructure partner, failed to secure planning approval for a pipeline that would have delivered treated wastewater to the site.

The proposed S7 facility, a 612-megawatt data center in Sydney, is expected to rank among the largest data centers in the world. Initial plans called for the facility to use treated sewage water to cool its servers, a strategy designed to conserve drinking water while minimizing additional pressure on the local electricity grid. Planning documents submitted in April highlighted the recycled water system as a key sustainability feature.

However, NEXTDC confirmed that discussions with local utilities ended after the necessary pipeline approval could not be obtained. As a result, the company has adopted an alternative cooling solution that eliminates water usage but increases electricity consumption.

Instead of recycled water, the facility will use direct liquid cooling technology, which circulates coolant around high-performance AI chips to remove heat efficiently. While this approach avoids the use of potable water and reduces dependence on water infrastructure, it requires significantly more energy to operate.

The decision underscores the growing challenges facing AI infrastructure developers as demand for high-performance computing accelerates worldwide. Advanced artificial intelligence systems require increasingly powerful processors, making cooling technology a critical factor in data center design, operational efficiency, and environmental impact.

The OpenAI-backed project also highlights the trade-offs between sustainability goals and infrastructure constraints. Although direct liquid cooling is becoming more common for AI-focused data centers due to its effectiveness in managing heat generated by advanced chips, the increased energy demand raises new questions about long-term power consumption and environmental sustainability.

As global investment in AI infrastructure expands, balancing energy efficiency, water conservation, and regulatory approvals will remain a key challenge for developers building the next generation of hyperscale data centers.