Nvidia Announces 2GW AI Data Center in Australia: Is NVDA's Next Growth Wave Going Global?

TradingKey - Nvidia (NVDA) recently announced that it will collaborate with Australian cloud computing and data center companies including Firmus, CDC, NEXTDC, and AirTrunk, aiming to build up to 2GW of AI data center capacity by 2027. The project will utilize Nvidia's DSX platform, with partners responsible for data center construction and operation, while Nvidia provides GPU computing, networking, software, and overall AI infrastructure technology.

Australia's current data center computing capacity is approximately 1.6GW. If the new plan is fully implemented, it would more than double existing capacity. Nvidia stated that the additional infrastructure will serve Australian local enterprises, AI startups, universities, and research institutions, while also catering to AI computing demand from the Asia-Pacific region and globally.

Australia is not the only market where Nvidia has recently expanded its AI infrastructure on a large scale; others include South Korea, Japan, and Europe.

In July this year, Nvidia and SK Group announced a long-term AI partnership plan exceeding $500 billion, under which SK Telecom plans to build a 2GW AI data center utilizing Nvidia Vera Rubin chips and SK Hynix HBM4, scheduled to begin operations in 2027. Nvidia also plans to invest $1 billion to acquire a stake in Naver, which is collaborating with Brookfield to expand the Sejong AI data center in South Korea, targeting 200MW by 2028.

In Japan, Nvidia announced in July this year a partnership with government-backed Noetra to build national physical AI infrastructure, planning to deploy 27,500 Rubin GPUs and 13,750 Vera CPUs, with a data center capacity of approximately 140MW. The project will primarily serve physical AI applications such as robotics, manufacturing, logistics, and healthcare, with construction expected to start in 2027 and operations to begin in 2028.

In Europe, Nvidia is advancing AI factory construction in the UK alongside partners including CoreWeave, Microsoft, and Nscale, with plans to deploy approximately 120,000 Blackwell Ultra GPUs and total UK data center investment reaching up to £11 billion. Meanwhile, the company is also pushing forward sovereign AI and industrial AI infrastructure in France, Germany, and Italy, with the German project planning to deploy around 10,000 Blackwell GPUs for manufacturing simulation, digital twins, and robotics scenarios.

For Nvidia, as next-generation chips such as Rubin enter their deployment cycle, its customer base is expanding from major U.S. cloud hyperscalers to foreign governments, telecom companies, data center operators, and local cloud service providers. If global 'sovereign AI' and AI data center construction continue to accelerate, Nvidia's next phase of growth momentum may no longer stem solely from chip upgrades, but rather from the continuous expansion of global AI infrastructure capacity.