Nvidia Expands Global AI Empire in Australia
Nvidia has made a significant move to expand its global AI empire by partnering with Australian cloud providers to build up to 2 gigawatts of AI computing capacity by 2027.
The partnership involves several companies, including IREN and Sharon AI, who will host Nvidia's DSX infrastructure in their facilities. This infrastructure combines hardware, networking, and software that is compatible with Nvidia's CUDA ecosystem.
Nvidia argues that software improvements can make the infrastructure more productive and durable, allowing operators to treat compute capacity as an investable, revenue-generating asset.
The Australian AI factory network will support multiple Nvidia chip generations rather than a single product cycle. Demand for the computing capacity is expected to come from companies building models and applications locally, including Atlassian which plans to use Nvidia's open models, datasets, and tools to customize multimodal applications and enhance its workplace AI platform.
The announcement demonstrates how Nvidia can influence infrastructure decisions long before chips are installed, potentially encouraging customers to remain within the company's ecosystem when upgrading equipment or expanding capacity.