Firmus Technologies announced it raised approximately A$500 million (about US$327 million) to ramp up development of its AI-data-centre business in Australia.
The raise is one of the latest large investments in infrastructure aimed at meeting rising demand for AI compute and training capacity. The company is backed by Nvidia and domestic investor Ellerston Capital.
The funds will support what Firmus calls Project Southgate, a plan to establish high-capacity, renewable-powered AI data centres with total capacity reaching up to 1.6 gigawatts by 2028.
The build-out will involve partnerships with data-centre specialist CDC Data Centres and aims to serve global AI-model training workloads in a region with favourable energy and regulatory conditions.
Firmus already raised A$330 million earlier this year. Combined with the new capital, the company’s valuation has reportedly climbed significantly, and the investment positions it as a regional contender in the AI-infrastructure arms race.
Company executives describe the build-out as enabling “fast, cost-effective scaling” of compute that aligns with Australia’s renewable-energy ambitions.
Funding And AI Infrastructure
The AI industry is under pressure to scale hardware and energy systems rapidly. Large-language-model training demands both high-density GPU clusters and locations capable of supporting heavy power loads, cooling, networking and sustainability factors.
Firmus’ approach combines these requirements in a renewable-energy context, tapping Australia’s potential as a clean-power region for next-generation data centres.
Nvidia’s involvement adds strategic weight. The chipmaker supplies much of the hardware used in AI training and may leverage its stake or partnership with Firmus to secure additional capacity or influence infrastructure design.
For global AI firms and cloud providers, more infrastructure-rich locations offer options outside the U.S. and Europe.
What To Monitor As The Plan Unfolds
Key indicators will include how quickly Firmus moves from funding to facility construction, how it secures renewable-energy contracts and how it attracts major customers for the new centres.
Also worth watching: whether the company files for an initial public offering next year as some analysts expect, and how it manages the technical challenges of large-scale AI-data-centre operations.
The broader takeaway: as AI accelerates, not just models but the physical infrastructure they run on is becoming strategic. The outcome of Firmus’ rollout will offer insights into how regions beyond traditional tech hubs may compete in the AI-hardware domain.
