Key Takeaways:
An Australian neocloud operator just secured 40,000 of Nvidia's most advanced GPUs through a revenue-sharing model that could reshape how mid-tier AI infrastructure providers access cutting-edge silicon.
Nvidia is extending its AI infrastructure financing model to Australia, signing a six-year compute collaboration with Sharon AI that will deploy 72 megawatts of data center capacity equipped with up to 40,000 Grace Blackwell GB300 GPUs.
"This strategic compute collaboration with Nvidia marks a pivotal moment in Sharon AI's mission to deliver sovereign, large-scale AI compute infrastructure," James Manning, co-founder and chief executive officer at Sharon AI, said.
The deal brings Sharon AI's total AI factory capacity to 132MW, of which 102MW is already contracted to end customers. The company expects to have more than 55,000 total Nvidia GPUs deployed by mid-2027. Under the revenue-sharing and credit-support structure, Sharon AI sells Nvidia-powered cloud services while Nvidia earns both standard product revenue and a share of the cloud revenue on the supported capacity.
The collaboration gives Sharon AI a capital-efficient path to scale while providing Nvidia with a recurring, usage-linked earnings stream — a model that accelerates adoption among customers who historically lacked access to capital-intensive AI infrastructure. Sharon AI shares surged 25% on the announcement.
How the Revenue-Sharing Model Works
The agreement departs from traditional GPU procurement, where cloud providers buy chips upfront and recoup costs through service margins. Instead, Nvidia shares in the upside of Sharon AI's cloud business, aligning incentives around utilization rather than unit sales. For Sharon AI, a company with a market capitalization of roughly $2.5 billion, this structure reduces the upfront capital burden of deploying infrastructure that would otherwise require hundreds of millions in hardware spending. For Nvidia, it creates a recurring revenue stream tied to actual compute consumption — a hedge against any future slowdown in direct GPU sales.
The 72MW deployment will use Nvidia's DSX AI factory design, a reference architecture optimized for dense GPU clusters. By comparison, a typical hyperscale data center hall runs 15-20MW, meaning this single project represents roughly four to five halls of dedicated AI compute. The GB300, Nvidia's latest Blackwell-generation GPU, succeeds the H100 and offers significant improvements in FP8 training throughput and memory bandwidth, according to Nvidia's published specifications.
Australia's Sovereign AI Ambitions
The collaboration positions Sharon AI as a key enabler of sovereign AI infrastructure in Australia, a market where governments and enterprises are increasingly seeking domestic compute capacity to reduce reliance on overseas cloud providers. Sharon AI, headquartered in Sydney, is a certified Nvidia Cloud Partner and already operates existing infrastructure within Australian data centers.
The deal also signals growing AI infrastructure investment in the Asia-Pacific region, where countries from Japan to Singapore are racing to build domestic GPU capacity. Australia's relatively low energy costs and established data center hubs in Sydney, Melbourne, and Perth make it an attractive location for AI factories, though the country faces constraints in grid capacity and construction timelines for new facilities.
Investor Implications
For Nvidia, the collaboration extends a financing strategy the company has deployed selectively with CoreWeave and other neocloud operators in the US and Europe. Each such deal locks in multiyear GPU demand while diversifying Nvidia's revenue beyond hyperscaler customers like Amazon, Microsoft, and Google, which are also developing their own AI chips. Nvidia shares rose 2.2% on the day, reflecting investor comfort with the model.
For Sharon AI, the agreement provides a clear growth trajectory: 132MW of total capacity with 77% already contracted. The remaining 30MW of uncontracted capacity represents potential upside if demand continues to outstrip supply in the Australian AI cloud market. The company's ability to secure 40,000 GB300 GPUs — among the most sought-after AI hardware globally — suggests strong confidence from Nvidia in Sharon AI's execution capability.
This article is for informational purposes only and does not constitute investment advice.
