NVIDIA expands AI infrastructure capacity through partnerships with Australian data centers
NVIDIA Newsroom
NVIDIA is lining up Australian data centers for up to a 2-gigawatt AI buildout by 2027. It’s a big bet that local compute, not just cloud access, will decide who gets to build AI.
Based on reporting by NVIDIA Newsroom — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
NVIDIA is putting real muscle behind Australia’s AI push. The company said it’s working with a cluster of local cloud and infrastructure partners — Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC and AirTrunk — to expand land, power and shell capacity for NVIDIA DSX AI factories, with a target of up to 2 gigawatts by 2027.
This isn’t just about more racks in more buildings. NVIDIA is supplying the DSX platform itself, plus accelerated computing, networking, software and ecosystem support, while the partner companies operate the facilities. The pitch is that these factories can support multiple generations of NVIDIA compute, which matters if you’re trying to avoid building infrastructure that feels old before it’s even full.
The list of partners shows how broad the effort is. Firmus is growing Project Southgate. Sharon AI says it is deploying up to 68,000 NVIDIA GPUs with Quantum InfiniBand and Spectrum-X Ethernet. IREN is pairing DSX with its own power, land and data center experience, and points to its 800-megawatt Bundey campus in South Australia. Megaport, through Latitude.sh, is aiming at inference demand. ResetData is focused on enterprises, government and researchers that need Australian data residency. CDC, NEXTDC and AirTrunk are all building out high-density, liquid-cooled infrastructure.
There’s also a clear regional angle here. NVIDIA says the expanded capacity should give Australian startups, universities, researchers, enterprises and AI-native companies better access to accelerated computing and Nemotron open models, so they can build models, applications and agents locally. That’s already showing up in named customers: Heidi is using Nemotron 3 Ultra for healthcare work, and Atlassian is applying NVIDIA AI to Rovo’s semantic search.
The message is simple: if Australia wants serious AI capability, it needs serious infrastructure first. NVIDIA is treating data centers as the new strategic layer, and, annoyingly for the people who prefer shiny demos to power cables, that’s probably the right order.
My take — AI-written commentary, not fact-checked reporting
This is the least glamorous part of the AI race, which is exactly why it matters. Everyone loves talking about models; fewer people want to talk about power, cooling and shells full of hardware. Australia is being sold the useful version of AI ambition here: build the boring stuff first, and the clever stuff can follow.
Read more about this at: NVIDIA Newsroom