NVIDIA is further extending its business to the underlying infrastructure behind the AI craze. The company plans to add up to 2 gigawatts of AI data center capacity in Australia by 2027, with partners including several local cloud service and infrastructure companies. This scale is already close to and may exceed the total current data center capacity in Australia.
Partners include cloud and infrastructure companies.
The enterprises participating in this construction plan include Firmus, IREN, NEXTDC, AirTrunk, CDC, Sharon AI, ResetData, and Megaport. The project will utilize the NVIDIA DSX platform to provide computing power support to AI laboratories, startups, large companies, universities, and developers.
From the disclosed information, this is not just a simple expansion of a single data center, but rather a centralized deployment aimed at meeting the computing needs of the region AI. If the project progresses as planned, NVIDIA's infrastructure presence in Australia will extend beyond chips and servers to include networks, software, and data center systems as well.
The scale of expansion exceeds the local existing base.
The report mentions that Australia currently has a data center capacity of about 1.6 gigawatts. Based on the goal of reaching up to 2 gigawatts, the additional capacity that NVIDIA is promoting this time will exceed the country's current data center capacity base.
This also reflects that NVIDIA's role is changing. Previously, the market saw it more as a supplier of GPU, but with the acceleration of the construction of the AI cluster, the company is increasingly involved in complete infrastructure solutions, including accelerated computing platforms, network connections, and software systems.
Electricity is becoming the bottleneck in the next phase.
Compared to the chips themselves, power supply is becoming a more realistic constraint for the expansion of AI data centers. Large-scale AI clusters not only require processors but also depend on grid access, transformers, cooling systems, and a continuous and stable power supply capacity.
The International Energy Agency predicts that global data center electricity consumption will increase from approximately 485 terawatt-hours in 2025 to around 950 terawatt-hours by 2030, with data centers that primarily use AI likely experiencing even faster growth. As the demand for computing power continues to rise, so does the importance of electricity, cooling, and physical data center resources.
For NVIDIA, this means that the business model is shifting from simply selling chips to a more comprehensive AI infrastructure approach. If the Australian project is successful, it will become another concrete example of this transition.
