Nvidia eyes 2 gigawatts of AI infrastructure in Australia by 2027

Nvidia's Australia announcement fits a pattern the company is building worldwide: instead of only selling GPUs into existing data centers, it is increasingly helping design and supply the infrastructure where those chips will run. The Australian partnerships with NVIDIA Cloud Partners and AI infrastructure providers are meant to expand land, power, and data center shell capacity, which Nvidia says is designed to host multiple generations of its DSX AI factory infrastructure. That approach could create demand not just for GPUs but for Nvidia's networking, software, and CUDA ecosystem.

The Australia push sits alongside other regional efforts. In Japan, Nvidia is developing a 140-megawatts AI factory expected to leverage Rubin GPUs and Vera CPUs for robotics, manufacturing, healthcare, and digital twins. In Europe, it is working with partners on new AI supercomputers across multiple countries. Each project could deepen reliance on Nvidia's architecture among enterprises, universities, government agencies, and startups.

Whether the Australian buildout reaches its 2027 target will likely hinge on local partners securing land and power, and on whether customers adopt DSX and CUDA at scale. If they do, Nvidia's position as a full-stack AI infrastructure platform could strengthen further; if not, the initiative may remain one of several regional bets rather than a new growth engine.