Nvidia and eight Australian cloud and data-center companies are working toward an AI-infrastructure buildout of up to 2 gigawatts by 2027. The companies plan to use Nvidia’s DSX platform in facilities run by the local providers, according to a September 9 company announcement.
The partners named are Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC and AirTrunk. Nvidia said it would supply computing, networking, software and ecosystem support, while the operators would run the AI factories. The 2GW figure is a target for the combined buildout, not capacity already in service. Reuters reported the expansion on September 10.
Operators outline different roles
Firmus is expanding Project Southgate, its Australian AI-factory program. The company has also announced a separate deal to supply OpenAI with computing capacity from two planned sites in Malaysia. Nvidia’s new announcement concerns the Australian ecosystem; it does not say that the Malaysian sites are included in the 2GW target.
Sharon AI said it is deploying up to 68,000 Nvidia GPUs with Quantum InfiniBand and Spectrum-X Ethernet networking. The announcement does not specify how many of those processors have been installed, the timetable for the full deployment or Sharon AI’s share of the proposed 2GW. IREN described its 800-megawatt Bundey campus in South Australia as part of its development portfolio using Nvidia’s AI-factory architecture. That campus figure is a project capacity, not an additional confirmed amount to add to the coalition-wide target.
CDC said its facilities across Australia and New Zealand currently operate more than 550 megawatts of capacity, with a further 800 megawatts under construction. Its existing and in-construction figures describe CDC’s broader portfolio; Nvidia did not say all of that capacity would be dedicated to DSX AI factories. NEXTDC cited its liquid-cooled facilities and cloud interconnections, while AirTrunk described plans to expand AI-ready powered shells with direct-to-chip liquid cooling.
Megaport said its Latitude.sh subsidiary would extend access to accelerated computing for inference workloads through the company’s network. ResetData described an AI-cloud offering for customers requiring Australian data residency. These roles range from providing land and powered facilities to delivering computing and connectivity, rather than a single construction project under Nvidia’s ownership.
Target and delivery details remain open
Nvidia describes DSX as a full-stack platform spanning facility designs, computing, networking and software. The company says broader domestic access to AI compute could serve startups, universities, researchers and enterprises. It also named Australian companies Heidi and Atlassian as users of its Nemotron models and other AI tools. Their model-development work is distinct from commitments to buy capacity at the proposed facilities.
The announcement does not provide a site-by-site power allocation, aggregate investment amount, financing breakdown or construction schedule for each operator. It also does not identify the share of the 2GW that already has grid connections, planning approvals, contracted customers or installed Nvidia systems. The proposed timetable therefore depends on work by several independent operators and the availability of power and supporting infrastructure.
Reuters noted that Australia’s current data-center capacity is about 1.6GW, citing industry group Data Centres Australia. That countrywide estimate is a separate measure from Nvidia’s proposed AI-focused buildout and should not be read as a direct before-and-after count of operational compute. Nvidia said its facilities could support both Australian users and wider Asia-Pacific demand.
Firmus already has a substantial regional footprint, including its planned Indonesian expansion, but Nvidia’s September announcement is centered on Australian capacity. The companies have not disclosed whether any Southeast Asian project would count toward the 2GW Australian target.
Featured image credit: NVIDIA
Australia’s Firmus raises $2B from Blackstone, Coatue, NVIDIA, to expand AI factories in APAC