Nvidia aims for up to 2 GW of AI infrastructure in Australia by 2027

Nvidia announced on Wednesday, September 9, 2026, in Melbourne a partnership with several Australian cloud and data center operators to significantly enhance local computing capabilities dedicated to artificial intelligence. The American company and its partners aim for a deployment that could reach 2 gigawatts by 2027.

The initiative includes Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC, and AirTrunk. These companies will provide and operate the land, power supply, and data center infrastructure necessary for multiple generations of Nvidia’s computing platforms.

Nvidia will contribute its DSX platform, accelerators, networking technologies, software, and technical support. The company describes these facilities as “AI factories,” designed to train and run large-scale models.

According to the company, the announced capabilities will be accessible to startups, universities, researchers, businesses, and developers in Australia. Nvidia specifically mentions Heidi and Atlassian among the organizations likely to use its tools and open models, Nemotron, to develop applications and AI agents tailored to the local market.

Some partners have already initiated large-scale projects. For instance, Sharon AI plans to deploy up to 68,000 Nvidia GPUs, while CDC reports operating over 550 megawatts of capacity in Australia and New Zealand, with an additional 800 megawatts under construction.

A surge in “AI factories”

The announcement illustrates the global acceleration of investments in infrastructure for artificial intelligence. Data centers need more electrical power, suitable cooling systems, and networks capable of connecting thousands of specialized processors.

Firmus is developing Project Southgate, presented as a set of AI infrastructures intended for Australia and the Asia-Pacific region. IREN is also collaborating with Nvidia on large computing centers that combine access to electricity, land, buildings, and GPU deployment.

Reuters confirmed on Thursday, September 10, the agreement between Nvidia and the group of Australian operators, as well as the goal of reaching up to 2 gigawatts of AI-related capacity by 2027.

Capacity potentially reaching 2 GW by 2027

Nvidia believes that Australia has advantages for hosting these infrastructures, including a developed digital sector, energy resources, and technological skills. The partners will operate the facilities, while Nvidia will remain the provider of the computing platform and software.

According to Nvidia, the announced projects are expected to bring the total capacity involved up to 2 gigawatts by 2027.