Nvidia는 호주 이해관계자들과 협력하여 Asia-Pacific 지역에 2GW AI 컴퓨팅 캠퍼스를 구축합니다.

Nvidia and eight Australian cloud and data-center operators are targeting a combined AI-infrastructure buildout of up to 2 gigawatts by 2027, using Nvidia's DSX full-stack platform across their facilities. The partners—Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC, and AirTrunk—announced the initiative on September 9. Nvidia will supply computing, networking, software, and ecosystem support, while the operators will run the AI factories. The 2GW figure represents a target for the coalition-wide buildout, not currently operational capacity.

The eight operators are taking distinct roles in the expansion. Firmus is expanding Project Southgate, its Australian AI-factory program, and has separately announced a deal to supply OpenAI computing capacity from two planned Malaysian sites; Nvidia's announcement focuses on the Australian ecosystem and does not indicate whether the Malaysian sites count toward the 2GW target. Sharon AI is deploying up to 68,000 Nvidia GPUs with Quantum InfiniBand and Spectrum-X Ethernet networking, though the announcement does not specify installation progress, deployment timeline, or Sharon AI's share of the total. IREN describes its 800-megawatt Bundey campus in South Australia as part of its portfolio using Nvidia's AI-factory architecture—a project capacity figure rather than a confirmed addition to the coalition target. CDC operates more than 550 megawatts across Australia and New Zealand currently, with another 800 megawatts under construction; Nvidia has not stated that all of this capacity will be dedicated to DSX AI factories. NEXTDC cited its liquid-cooled facilities and cloud interconnections, while AirTrunk outlined plans to expand AI-ready powered shells with direct-to-chip liquid cooling. Megaport's Latitude.sh subsidiary will provide accelerated computing access for inference workloads through its network, and ResetData is offering an AI-cloud service for customers requiring Australian data residency.

The announcement leaves critical details unresolved. Nvidia has not disclosed site-by-site power allocation, total investment, financing breakdown, or construction schedules for each operator. It also does not identify how much of the 2GW target already has grid connections, planning approvals, customer contracts, or installed Nvidia systems. Delivery depends on work by several independent operators and the availability of power and supporting infrastructure.

Australia currently has approximately 1.6GW of data-center capacity, according to Data Centres Australia. Nvidia's proposed 2GW AI-focused buildout should not be read as a direct before-and-after comparison, as the figures measure different infrastructure cohorts. Nvidia said the facilities could serve both Australian users and broader Asia-Pacific demand. The company also named Australian firms Heidi and Atlassian as users of its Nemotron models and other AI tools, though their model-development work is separate from commitments to purchase facility capacity.