TechNewsReel
Live

Nvidia Partners With Eight Australian Firms for 2GW AI Infrastructure Buildout

The chipmaker aims to deploy massive 'AI factory' capacity by 2027 to bolster Australia's sovereign compute capabilities.

TechNewsReel Newsroom · September 10, 2026

Nvidia has launched a strategic collaboration with eight Australian cloud and data center providers to deploy up to 2 gigawatts (GW) of AI factory capacity by 2027. The initiative establishes a domestic foundation for high-density compute, ensuring that AI training and inference occur within Australian borders.

To achieve this scale, Nvidia is utilizing its DSX platform, which integrates accelerated computing, networking, and software across local facilities. The partnership involves a consortium of eight industry players: AirTrunk, NEXTDC, CDC, ResetData, Megaport, IREN, Sharon AI, and Firmus. Specific deployments are already taking shape; Sharon AI plans to integrate up to 68,000 NVIDIA GPUs supported by Spectrum-X Ethernet and Quantum InfiniBand, while IREN is implementing the DSX architecture at its 800-megawatt Bundey campus in South Australia.

The Shift Toward AI Factories

This expansion comes as global AI workloads transition from general-purpose cloud computing to specialized "AI factories." These facilities are engineered specifically to handle the extreme power and cooling demands of modern large-scale models. For Australia, the move addresses a critical need for data sovereignty. By building domestic capacity, the country reduces its reliance on overseas compute hubs, which often introduce significant latency issues that can hinder the performance of real-time AI applications.

Strategic Implications for the Region

Nvidia's approach signals a broader strategy to decentralize AI compute by embedding its full-stack architecture into existing local ecosystems rather than constructing its own proprietary data centers. This allows the company to scale its hardware and software footprint rapidly while empowering local enterprises to build sovereign AI tools.

Industry players are already leveraging this ecosystem. Atlassian is utilizing Nemotron open models for enterprise search, while Heidi is applying similar technology to healthcare applications. By providing the underlying infrastructure, Nvidia enables these firms to develop and run sensitive applications without moving data across international borders.

Future Outlook

As the 2027 deadline approaches, the primary focus will be the physical rollout of the 2GW capacity and the integration of the DSX platform across the eight partner sites. The success of the initiative will likely be measured by the speed at which Australian firms migrate their workloads to these local factories and the subsequent reduction in latency for domestic AI services. Observers will be watching to see if this model of partner-led infrastructure expansion is replicated in other regions seeking digital sovereignty.

Sources

Get a notification when a big story breaks. A few a day at most — no spam.