NVIDIA and eight Australian partners plan a 2-gigawatt AI infrastructure buildout

NVIDIA is backing a major expansion of Australia's AI data center capacity, partnering with eight local infrastructure providers on a buildout that could more than double the country's existing footprint. The company announced the collaboration on September 9, 2026, targeting up to 2 gigawatts of new capacity by 2027.

What's new

The eight partners named are Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC, and AirTrunk — spanning cloud providers, colocation operators, and networking firms across Australia's data center ecosystem. Each will build and operate AI factories on NVIDIA's DSX platform, expanding "land, power and shell capacity designed to host multiple generations of NVIDIA DSX AI factories, supporting the nation's growing demand for AI compute — with up to a 2-gigawatt buildout by 2027."

Raj Mirpuri, NVIDIA's vice president of global AI clouds and infrastructure ecosystem, framed the buildout in terms of energy conversion: "AI factories turn energy into intelligence — the essential resource of the AI economy," he said.

The scale is significant relative to the market: Australia's entire existing data center footprint is roughly 1.6 gigawatts today, meaning a 2-gigawatt addition would more than double national capacity. One partner, Sharon AI, alone plans to deploy up to 68,000 GPUs as part of the buildout.

Context

The expansion builds on NVIDIA's DSX full-stack AI factory platform, which pairs NVIDIA's accelerated computing and networking hardware with software and ecosystem support for operators. NVIDIA says local organizations are already putting the resulting capacity to use: companies including Heidi and Atlassian are using NVIDIA's open Nemotron models to build region-specific AI applications in healthcare and enterprise search.

The announcement is part of a broader pattern of "sovereign AI" infrastructure deals NVIDIA has struck with national and regional partners throughout 2026, as governments and local providers push to keep AI compute — and the data that runs through it — within their own borders rather than depending entirely on US hyperscale clouds.

Why it matters

A 2-gigawatt commitment is a meaningful bet on sustained AI demand in a single mid-sized national market, and it signals that the capital cycle behind AI data centers is broadening beyond the US and a handful of hyperscalers to national ecosystems built around local power, land, and regulatory advantages. For NVIDIA, spreading DSX deployments across a diverse set of regional partners — rather than a single hyperscaler — reduces customer concentration risk and seeds new demand for its full hardware-and-software stack in a market it does not yet dominate.