Nvidia Australia data center expansion is set to reshape the country’s AI infrastructure landscape through a planned 2GW capacity buildout. The initiative brings together major Australian data center operators, cloud providers and infrastructure developers.
Moreover, Nvidia plans to deploy its DSX platform across participating facilities by 2027. The expansion targets surging demand for high-performance computing from AI developers, enterprises, researchers and startups.
Nvidia Australia Data Center Expansion Targets 2GW
The planned buildout could more than double Australia’s existing data center computing capacity. Data Centers Australia estimates current national capacity at approximately 1.6GW.
Nvidia’s target involves up to 2GW of additional AI-related capacity across multiple infrastructure projects. Therefore, the initiative represents a major construction pipeline rather than one standalone facility.
The company is working with Firmus, CDC Data Centers, NEXTDC and AirTrunk. Other participating infrastructure partners include Sharon AI, IREN, Megaport and ResetData.
Furthermore, these companies are expanding land, power availability and data center shells for AI factories. Nvidia will provide accelerated computing, networking, software and DSX reference architecture.
The DSX platform integrates facilities infrastructure, computing, networking and software. It also supports multiple generations of Nvidia accelerated computing infrastructure.
Nvidia Australia Data Center Projects Expand AI Factory Infrastructure
Firmus provides one of the largest components of the broader construction program through Project Southgate. The project targets up to 1.6GW across Australia by 2028.
Firmus previously secured $10 billion in debt financing from funds led by Blackstone and Coatue. The financing supports Project Southgate’s national expansion and AI factory infrastructure.
Project Southgate spans planned deployments across Tasmania, Melbourne, Canberra, Sydney and Perth. Firmus also secured a multi-billion-dollar hyperscale agreement for approximately 18,400 Nvidia GB300 GPUs.
Meanwhile, Sharon AI is developing another significant Nvidia-powered AI facility. Its six-year collaboration covers 72MW of new data centre capacity and up to 40,000 Grace Blackwell GB300 GPUs.
Sharon AI’s broader AI factory capacity could reach 132MW following the deployment. The company expects more than 55,000 Nvidia GPUs across its infrastructure by mid-2027.
CDC already operates more than 550MW across Australia and New Zealand. It also has another 800MW under construction, supporting high-density computing workloads.
NEXTDC is similarly developing liquid-cooled infrastructure for high-density AI factories. AirTrunk is expanding powered shells with direct-to-chip liquid cooling capabilities.
AirTrunk’s role in expanding Australia’s AI-ready data centre capacity also extends to Western Sydney, where its $2.8 billion Sydney data center project is advancing through a proposed A$4.3 billion financing package. The development supports the broader infrastructure expansion needed to accommodate rising AI and hyperscale computing demand across Australia.
Nvidia Australia Data Center Expansion Raises Power Requirements
The construction program also highlights Australia’s growing electricity requirements for AI infrastructure. Transgrid modelling indicates data center demand could reach 3.5GW in New South Wales and the ACT by 2035.
Annual data center electricity consumption could also rise from five terawatt-hours to 34 terawatt-hours by 2035-36. Consequently, developers face increasing pressure to secure reliable and sustainable power supplies.
Nvidia says expanded capacity can support additional power generation projects alongside AI computing growth. However, Australian authorities face concerns about energy consumption, water use and environmental impacts.
The infrastructure could nevertheless strengthen Australia’s position as a regional AI computing hub. It will also give local developers greater access to accelerated computing and Nvidia’s Nemotron open models.
The wider construction push therefore extends beyond server buildings. It includes power infrastructure, cooling systems, fiber connectivity, GPU deployment and specialized technical facilities.
Project Fact Sheet
Project: Nvidia Australia AI data center expansion
Focus: Nvidia Australia data center
Planned AI capacity: Up to 2GW
Target completion: 2027 for the broader Nvidia-supported buildout
Existing Australian capacity: Approximately 1.6GW
Technology platform: Nvidia DSX
GPU technology: Nvidia accelerated computing, including Grace Blackwell GB300
Major locations: Australia-wide, including Tasmania, Melbourne, Sydney, Canberra and Perth
Associated program: Firmus Project Southgate
Southgate target: Up to 1.6GW by 2028
Southgate financing: $10 billion debt facility
Sharon AI capacity: 72MW under its Nvidia collaboration
Sharon AI GPU deployment: Up to 40,000 additional GB300 GPUs
Cooling: Direct-to-chip and liquid-cooling technologies
Connectivity: High-speed fiber and interconnected data center infrastructure
Primary purpose: AI training, inference, research and enterprise computing
End users: Startups, universities, researchers, enterprises and AI-native companies
Sustainability focus: Renewable electricity and advanced water-efficient cooling
Project Team
AI technology provider: Nvidia
AI infrastructure developer: Firmus Technologies
Data center partner: CDC Data Centers
Data center operator: NEXTDC
Hyperscale infrastructure provider: AirTrunk
AI cloud partner: Sharon AI
AI infrastructure partner: IREN
Connectivity partner: Megaport
AI infrastructure provider: ResetData
DSX platform provider: Nvidia
Project Southgate leadership: Firmus Technologies
Southgate infrastructure partner: CDC Data Centers
GPU platform: Nvidia Grace Blackwell
Networking technology: Nvidia Quantum InfiniBand and Spectrum-X Ethernet
Financing partner for Southgate: Blackstone-led funds with Coatue Management support
Leave a Reply