NVIDIA has announced a partnership with eight data center and cloud operators in Australia to build up to 2 gigawatts of new AI computing capacity in the country by 2027. Australia's existing computing capacity sits at roughly 1.6 gigawatts, so the buildout would more than double the national total if it proceeds as planned, with large-scale GPU deployments and new data center construction now underway across the participating companies.
Eight Partners Join the Buildout
The partnership brings together Firmus, Sharon AI, IREN, Megaport (participating through its wholly owned subsidiary Latitude.sh), ResetData, CDC, NEXTDC, and AirTrunk. Australia's existing computing capacity currently stands at roughly 1.6 gigawatts, according to figures from analyst firm DC Byte[2]. The 2-gigawatt figure is not a new total but a target for what the partners plan to add by 2027, which would take the national footprint to more than double its current size. The lineup spans data center operators and network providers alike, reflecting a distributed expansion across multiple companies rather than a single mega-facility.
Large-Scale Investment From Each Partner
Each participating company is pursuing its own buildout. Sharon AI plans to deploy up to 68,000 NVIDIA GPUs alongside a purpose-built network architecture. IREN is developing an 800-megawatt campus in South Australia known as the Bundey campus, which sits at the center of the company's development portfolio[1]. CDC already operates more than 550 megawatts of capacity across Australia and New Zealand, with an additional 800 megawatts under construction. CDC says its facilities run on 100 percent renewable electricity.
NVIDIA's "AI Factory" Vision
Raj Mirpuri, NVIDIA's vice president of global AI clouds and infrastructure ecosystem, said that AI factories turn energy into intelligence, calling it an essential resource for the AI economy. NVIDIA frames the partnership as part of its "full-stack" AI factory strategy, which bundles facilities, compute, networking, software, and reference designs together. Partner executives echoed that sentiment: Firmus's Oliver Curtis described the effort, known as Project Southgate, as a tangible commitment to building Australia's AI future, while Sharon AI's James Manning said Australia has the energy, connectivity, and ambition to lead in AI.
Meeting Surging Demand From Business and Research
The buildout responds to surging demand for high-performance AI computing from startups, universities, research institutions, and enterprises. NVIDIA points to Heidi, a healthcare technology company, and Atlassian, the collaboration software maker, as examples of companies building on its open Nemotron models. Heidi is combining Nemotron with its own clinical knowledge for voice-driven AI, while Atlassian is using the models to advance semantic search in its enterprise search product, Rovo. The infrastructure is also positioned to serve AI compute demand beyond Australia's borders, supporting regional and global needs.
Summary
NVIDIA has partnered with eight Australian data center and cloud operators to build up to 2 gigawatts of new AI computing capacity by 2027. Added to the roughly 1.6 gigawatts already in place, that would more than double the country's total. Large-scale GPU deployments and new facility development are underway across the participating companies, with renewable energy use and healthcare and enterprise AI applications underscoring that the effort goes beyond simple hardware expansion.
