NEXTDC and OpenAI to Build Major AI Campus in Sydney
NEXTDC signed a memorandum of understanding with OpenAI to develop a hyperscale AI campus and GPU supercluster at its Eastern Creek site, sending the data centre operator's shares sharply higher. The project is part of OpenAI's Australia programme and marks a major push into local AI infrastructure while prompting scrutiny over electricity demand and legal issues tied to model training.
NEXTDC Ltd said it had signed a memorandum of understanding with OpenAI to collaborate on the development of a hyperscale artificial intelligence campus and a graphics processing unit supercluster at its S7 site in Eastern Creek, Sydney. The announcement on Thursday triggered a strong market reaction, with shares of the Queensland headquartered data centre operator rising as much as 10.9 percent to A$14.90 by 2312 GMT, their highest level since November 13.
The Eastern Creek site was purchased last year for nearly A$353 million, or about $233.37 million, and NEXTDC has indicated the location could ultimately support up to 550 megawatts of capacity. The scale of the planned build reflects the rapid expansion in demand for specialised data centre capacity to train and run large language models and other AI systems, a trend that has driven hyperscale investment worldwide.
The collaboration sits under OpenAI's Australia programme, part of the company’s stated economic blueprint to accelerate AI adoption and expand infrastructure investment across the country. For NEXTDC, landing a major international AI developer as an anchor client offers an opportunity to convert rising demand into long term revenue, supporting its recent decision to raise capital expenditure guidance. The company last week increased its FY26 capex range by A$400 million to between A$2.2 billion and A$2.4 billion as it expands inventory to meet new contracts.
"This is a significant step for NEXTDC," said Michael McCarthy, chief executive of Australia and New Zealand at trading platform Moomoo. Industry observers said the presence of an anchor tenant such as OpenAI can provide revenue certainty that underpins the financing and construction of major projects of this scale.
Analysts and policy experts cautioned that the deal brings with it several challenges. High electricity consumption associated with GPU superclusters raises questions about grid capacity, long term operating costs and the environmental footprint of concentrated AI compute. There is also ongoing debate over the legal implications of training large language models, particularly around copyright and the use of proprietary content in model training data.
The project could deliver substantial local economic benefits if it proceeds to construction, including jobs in construction and operations and increased activity for local suppliers, but planners and regulators will closely scrutinise its impact on power networks and community concerns about industrial scale infrastructure.
The announcement underscores how global AI developers are racing to secure physical capacity close to major markets, and how national strategies for AI infrastructure are drawing partnerships between international technology firms and regional operators. For NEXTDC, the MoU represents a strategic pivot toward meeting the specialised needs of AI customers. For Australia, it poses choices about how to balance rapid technology investment with energy, legal and social considerations as the country positions itself in a shifting global AI ecosystem.
