Firmus signs long-term contract with unnamed ‘global hyperscale customer’
Make us a preferred source
Make us a preferred source on Google
The news: AI factory company Firmus has signed a long-term contract with an unnamed ‘global hyperscale customer’ for dedicated AI infrastructure capacity at Project Southgate’s first deployment in Australia.
The numbers: The contract is a multi-billion-dollar commitment for approximately 18,400 Nvidia GB300 graphics processing units (GPUs) to be deployed at the Melbourne facility.
The context: It represents the second large-scale customer secured for Project Southgate, which is being developed as a multi-site AI factory program across Australia.
Firmus said the contract marks “the arrival of hyperscale AI factories in Australia for the first time, positioning the country on a path to become a major producer and exporter of AI tokens globally”.
Firmus has committed more than $300 million to its domestic supply chain, establishing advanced manufacturing capacity to support up to 1.5 gigawatts of AI facility delivery per year.
This first deployment forms part of the broader Project Southgate build-out, backed by a USD10 billion ($14 billion) debt facility led by Blackstone.
Firmus is also progressing with construction of a greenfield AI factory campus in Tasmania, which will host approximately 36,800 Nvidia GB300 GPUs, due for completion in late 2026.
What they said: “This agreement is a defining milestone for Firmus and for Australia’s role in the global AI infrastructure landscape,” said Firmus co-CEO Tim Rosenfield.
“As the second major contract secured under Project Southgate, it reflects growing market conviction in our approach. A commitment of this scale from one of the world’s leading technology platforms validates what we’ve been building AI Factories that are co-designed from model to grid, where every decision in power, cooling and network is engineered for production performance at hyperscale,” he said.
“These AI Factories are purposely engineered to minimise resource use and maximise performance, with the specific output to generate the best Total Cost of Operation per AI token for our customers.”
The source: Firmus media release
