June 28, 2026, (Inside AI) — Australian AI infrastructure company Firmus Technologies has struck a strategic partnership with Nvidia Corp, the U.S. chip giant, to slash computing costs for emerging AI firms. The deal, announced on Monday in Sydney, aims to democratize access to high-performance AI infrastructure.
The partnership will leverage Nvidia’s advanced GPU technology within Firmus’s sustainable data centers. Firmus specializes in immersion cooling, which reduces energy consumption by up to 45% compared to traditional air-cooled systems. This efficiency is expected to translate into lower operational costs for AI startups and scale-ups.
Financial terms were not disclosed. However, the collaboration signals Nvidia’s deepening push into the Asia-Pacific region, where demand for AI compute is surging. Firmus, founded in 2019, has positioned itself as a green alternative in the infrastructure space, operating facilities powered by renewable energy.
Sustainable Compute Meets Silicon Valley Muscle
The deal pairs Firmus’s eco-friendly hosting with Nvidia’s H100 and GH200 chips, which are the backbone of modern AI training. By integrating these GPUs into its liquid-cooled racks, Firmus can offer competitive pricing without sacrificing performance. Industry analysts note that cooling alone accounts for up to 40% of data center energy costs, making immersion technology a key differentiator.
Read: Nvidia Plans $20 Billion Bond Sale, First in Five Years, to Fuel AI Chip Race
“This partnership will help emerging AI companies access world-class computing power at a fraction of the typical cost,” a Firmus spokesperson said. The company has already attracted clients in biotech, finance, and autonomous systems, sectors where compute expenses can stifle innovation.
Nvidia’s involvement extends beyond hardware supply. The agreement includes joint go-to-market efforts and technical collaboration to optimize AI workloads on Firmus’s infrastructure. This mirrors Nvidia’s strategy of embedding its ecosystem within regional hubs to capture long-term cloud revenue.
Australia’s AI Ambitions Get a Boost
The partnership aligns with Australia’s goal to become a leading AI hub in the Indo-Pacific. Government initiatives like the National AI Centre and investments in sovereign capability have created fertile ground for local infrastructure players. Firmus, based in Sydney, has benefited from tax incentives for green data centers and growing enterprise demand for onshore compute.
Yet challenges remain. Australia’s energy grid faces strain, and critics question whether immersion cooling alone can offset the massive power draw of next-gen AI clusters. Competing providers like NextDC and AirTrunk are also scaling up, though they rely on conventional cooling methods. Firmus’s niche could give it an edge in winning price-sensitive AI startups.
The deal does not include exclusive rights, leaving room for other Australian firms to partner with Nvidia. However, it marks the first time Nvidia has formally backed an immersion-cooling specialist in the region. This could spur similar alliances as the industry grapples with AI’s environmental footprint.
Read: Cloudflare Acquires Ensemble AI Team to Slash Inference Costs
Looking ahead, Firmus plans to expand its data center footprint across Melbourne and Perth by 2027, potentially doubling its capacity. The Nvidia partnership is expected to accelerate those plans, though execution risks tied to supply chain and regulatory approvals persist. For now, the collaboration signals that sustainable AI infrastructure is moving from niche to necessity.
