Nvidia Deploys AI Infrastructure in Australia, Up to 2 GW by 2027
Nvidia has announced a collaboration with a group of Australian data center partners to deploy infrastructure dedicated to artificial intelligence across the country.
The agreement aims to host multiple generations of its DSX computing platform, with capacity potentially reaching 2 gigawatts by 2027.
Collaboration with Eight Local Infrastructure Partners
Nvidia indicated it is collaborating with several Australian cloud partners and infrastructure providers, among which Firmus, Sharon AI, Iren, ResetData, Megaport, CDC, Nextdc and AirTrunk.
The objective of this collaboration is to expand land, energy and building capacity ("shell") intended to host multiple generations of Nvidia DSX computing infrastructure.
According to the announced distribution, participating providers will operate the "AI factories", with Nvidia supplying the DSX platform, accelerated computing, networking, software and support from its ecosystem.
Capacity of 2 Gigawatts Targeted by 2027
The deployment covers capacity potentially reaching 2 gigawatts by 2027, in response to growing demand in the country for AI-dedicated computing.
Several partners have specified their own projects. Sharon AI plans the deployment of up to 68,000 Nvidia GPUs connected via Nvidia Quantum InfiniBand networks and Nvidia Spectrum-X Ethernet. Iren mentions its Bundey campus of 800 megawatts in South Australia.
CDC indicated it operates more than 550 megawatts of capacity in Australia and New Zealand, with 800 megawatts additional under construction.
DSX and Open Models for Local Players
The DSX platform is presented by Nvidia as a comprehensive infrastructure covering facilities, computing, networking, software and reference architectures, compatible with the CUDA ecosystem.
This capacity is intended to provide Australian players with expanded access to accelerated computing and Nvidia Nemotron open models.
Organizations such as Heidi and Atlassian use these open models to develop applications, notably in healthcare for Heidi and for semantic search of the Rovo tool at Atlassian.
