NVIDIA is working with Australian NVIDIA Cloud Partners and AI infrastructure providers to expand the country’s capacity for AI data centers, with plans designed to support multiple generations of NVIDIA DSX™ AI factories and a potential buildout of up to 2 gigawatts by 2027. The effort is focused on securing the physical foundations AI systems require: land, power and data center shell capacity. as reported by Nvidianews Nvidia ## Building capacity around AI factories
The announcement positions Australia’s data center ecosystem as a key part of NVIDIA’s response to rising demand for AI compute. Rather than describing a single facility, NVIDIA outlined a broader collaboration with a growing group of partners that can provide the sites and infrastructure needed to host AI systems over multiple hardware generations.
The planned capacity is intended for NVIDIA DSX AI factories, NVIDIA’s term for infrastructure designed to support AI workloads. The announcement does not identify the individual Australian partners involved or specify how the total capacity will be distributed across sites.
Why land, power and shell capacity matter
AI infrastructure depends on more than processors and software. Data center land provides room for development, power supplies the computing systems, and shell capacity refers to the physical data center space prepared to house the technology. For a practical look at image generation, Harper is a useful comparison.
By focusing on these elements together, NVIDIA and its partners are preparing the underlying capacity before multiple generations of AI infrastructure are deployed. The approach also gives Australian NVIDIA Cloud Partners a role in expanding access to NVIDIA-based compute within the country’s existing data center ecosystem.
The proposed scale—up to 2 gigawatts by 2027—signals an infrastructure plan measured in power capacity rather than a single system or installation. However, the announcement does not provide a detailed construction schedule, identify specific sites or state how much capacity has already been secured.
What to watch next
The central question is how quickly the planned capacity becomes operational. NVIDIA has described the initiative as an expansion of land, power and shell capacity, but the available announcement does not state when individual facilities will open or how customers will access the resulting compute. The ai hardware story also surfaces in Japan Partners With NVIDIA to Accelerate..., adding another angle.
The partnership also highlights the importance of local infrastructure ecosystems in national AI development. Australia’s growing demand for AI compute will depend not only on NVIDIA technology, but also on the ability of cloud and infrastructure partners to provide suitable facilities at the scale required.
The 2027 target will therefore be an important marker for whether the proposed buildout progresses from infrastructure planning to deployed AI factory capacity.