en.Wedoany.com Reported - On June 12, Australian AI cloud service provider Sharon AI announced a six-year AI infrastructure computing power cooperation agreement with Nvidia. The two parties will add 72 megawatts of data center capacity in Australia and deploy Nvidia's DSX AI Factory architecture. According to the agreement, the project can be expanded to 40,000 Grace Blackwell GB300 GPUs in the future, providing large-scale accelerated computing resources for AI startups, commercial clients, and university research institutions. After the agreement takes effect, Sharon AI's total AI factory capacity will increase to 132 megawatts, of which 102 megawatts are already supported by end-customer contracts. The company expects to deploy over 55,000 Nvidia GPUs by mid-2027.

This collaboration elevates Australia's AI computing power construction to a higher scale. Sharon AI is one of the Neocloud enterprises that have emerged in recent years, focusing not on traditional general-purpose cloud services but on providing infrastructure around GPU clusters, AI training, inference tasks, and enterprise-level computing power leasing. With the rapid growth of large models, agent applications, scientific computing, image generation, and enterprise AI systems, GPU resources have become a core cost for developers, research institutions, and industry clients. The additional 72 megawatts of capacity means Sharon AI will gain more space for power supply, server rooms, and cluster deployment, also providing more locally available computing resources for AI applications in Australia. For universities and startups, local AI cloud services can reduce cross-border latency and help address some data compliance and sovereign AI needs.

The DSX AI Factory architecture in the agreement is a reference design direction from Nvidia for large-scale AI data centers. This architecture emphasizes integrating GPUs, networking, storage, software, power supply, cooling, and operations management into a unified system, enabling data centers to be built and expanded like AI factories. AI factories differ from ordinary server rooms, requiring higher computing density, power demands, liquid cooling capabilities, network bandwidth, and cluster scheduling. Any single bottleneck can affect training and inference efficiency. Sharon AI's decision to deploy this architecture in Australia indicates that its expansion goal is not to add servers piecemeal but to build a comprehensive platform capable of continuously handling large model training, enterprise inference, and high-performance computing tasks.

The Grace Blackwell GB300 GPU corresponds to Nvidia's next-generation AI computing platform. A future deployment scale of up to 40,000 GPUs will position Sharon AI as a regional-level AI infrastructure provider. For the Australian market, such clusters can support AI application development in finance, healthcare, mining, energy, education, media, and the public sector, and provide a more stable high-performance computing environment for university research institutions. Australia's industrial structure has a high proportion of mining, energy, and engineering services, and these industries are increasingly using AI for geological modeling, equipment maintenance, production optimization, risk prediction, and automated decision-making. Local computing power supply capabilities will affect the pace of related application advancement.

This deal also reflects that global AI infrastructure competition is shifting from model companies to computing power operators. Over the past two years, tight supply of Nvidia GPUs has driven the rapid growth of a group of enterprises focused on AI cloud services, which meet market demand through large-scale procurement, long-term leasing, and customized cluster services. Sharon AI's six-year cooperation arrangement, with 102 megawatts of capacity supported by end-customer contracts, shows that its expansion is not solely based on long-term concepts but has already secured some actual demand. For Nvidia, supplying GPUs, networking, and AI factory architecture to regional AI cloud service providers can expand its enterprise customer coverage in the Asia-Pacific region and allow more small and medium-sized enterprises and research institutions to access high-end GPUs through cloud services.

Australia's AI computing power construction still faces pressures from electricity, funding, and delivery. AI factories require stable power supply, high-standard cooling, low-latency networking, professional operations teams, and sustained capital expenditure. The additional 72 megawatts of capacity involves not just server procurement but also site access, grid support, liquid cooling systems, rack density, security management, and customer onboarding timelines. Sharon AI expects to deploy over 55,000 Nvidia GPUs by mid-2027, meaning that over the next year or so, equipment delivery, data center construction, customer migration, and service stability will become critical milestones.

This six-year agreement pushes Sharon AI to the forefront of AI infrastructure construction in Australia. As demand for computing power from AI startups, enterprise clients, and university research institutions continues to rise, localized AI factories will become part of the digital economy and sovereign AI capabilities. Whether Sharon AI can complete GPU deployment as planned and convert 132 megawatts of capacity into stable revenue and sustainable operations will determine its position in the Asia-Pacific AI cloud services market. For Australia, the significance of such projects lies not only in increasing the number of GPUs but also in providing infrastructure support closer to usage scenarios for local AI research and development, industry model training, and enterprise intelligent transformation.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com