Nvidia announced strategic partnerships with Australian cloud partners and AI infrastructure providers to build up to 2 gigawatts of AI computing capacity by 2027. The buildout could extend Nvidia's reach beyond chip sales into a broader infrastructure role, though power availability and utilization remain open risks.
Nvidia announced strategic partnerships on September 9 with Australian NVIDIA Cloud Partners and AI infrastructure providers, working toward an AI infrastructure buildout of up to 2 gigawatts by 2027. The initiative expands the availability of land, power, and data center shell capacity built to host multiple generations of Nvidia's DSX AI factory infrastructure.
Expanding the Nvidia ecosystem
The Australian buildout could benefit Nvidia beyond the initial sale of GPUs. Nvidia will provide its DSX platform, accelerated computing, networking, software, and ecosystem support to the emerging network of AI factories, and DSX is also compatible with Nvidia's CUDA ecosystem. As enterprises, universities, government agencies, and startups increasingly build AI workloads around Nvidia's architecture, the company could strengthen CUDA's position as the underlying software platform for AI development and deployment.
This initiative therefore highlights Nvidia's evolution from primarily a chip supplier toward a broader full-stack AI infrastructure platform. Rather than simply selling GPUs into existing data centers, Nvidia is increasingly helping shape the infrastructure in which its products will be deployed.
A global AI factory expansion
Australia is only one part of Nvidia's broader global infrastructure strategy. The company is expanding AI factory capacity across North America, Europe, and Asia as governments, cloud providers, enterprises, and other organizations seek greater access to domestic and regional AI computing resources. In Japan, Nvidia is developing a 140-megawatt AI factory expected to use Rubin GPUs and Vera CPUs for applications including robotics, manufacturing, healthcare, and digital twins.
In Europe, Nvidia is also working with partners on new AI supercomputers across multiple countries. Each additional AI factory potentially creates demand not only for Nvidia GPUs, but also for its CPUs, networking products, software, and broader infrastructure architecture.
Power and utilization risks remain
However, the biggest challenge to Nvidia's global infrastructure ambitions may not be demand — it could be electricity. A 2 GW buildout requires enormous amounts of reliable power, and delays in grid connections, transmission infrastructure, land development, construction, or regulatory approvals could push back Australia's planned timeline.
There is also a utilization risk. Infrastructure operators, and indirectly Nvidia, are betting that AI workloads will keep expanding rapidly enough to justify large-scale AI factories. If Australian enterprises, government agencies, research institutions, and startups do not consume AI compute as quickly as expected, operators could face underutilized GPU clusters and weaker returns, which could in turn slow additional AI factory deployments.
Hedge funds stay bullish
Institutional investors remain broadly bullish on Nvidia's long-term AI opportunity. According to Insider Monkey data, 285 hedge funds held Nvidia shares at the end of the second quarter, compared with 275 in the previous quarter. Fisher Asset Management increased its Nvidia position by 3% to approximately $18.19 billion, while AQR Capital Management increased its stake by 18% to approximately $7.48 billion.
Short interest also remains relatively low: approximately 1.23% of Nvidia's shares were sold short as of August 14, representing roughly 285.96 million shares. The limited bearish positioning suggests investors continue to view Nvidia as one of the primary beneficiaries of the global AI infrastructure spending cycle.
Source: Insider Monkey
Trading involves risk.
Risk Disclaimer
