The Operator Landscape Is Consolidating Around AI-Ready Capacity
Australia’s data centre market has long been dominated by a mix of domestic specialists and global hyperscalers. That structure is now being reshaped by AI. CBRE’s 2026 report identifies AI uptake as the defining force in the next phase of growth, increasing the scale, infrastructure requirements, and investment intensity of data centre development. The operators that thrive will not simply be those with the most land or the largest sheds, but those that can deliver high-density, liquid-cooled, power-secured capacity at scale.
AirTrunk’s A$75 Billion AI Bet
AirTrunk, Australia’s largest data centre owner, has signaled the scale of its ambition with plans to spend A$75 billion on AI infrastructure in Australia over the next decade. The company has secured a A$4.25 billion loan to fund its expansion, including the SYD3 400 MW data centre campus in Western Sydney, targeted for completion by mid-2028. AirTrunk is also expanding AI-ready powered shells across the region, incorporating direct-to-chip liquid-cooling infrastructure designed to support high-density AI workloads. The message is clear: the next generation of data centres is being designed around GPU density, not general-purpose cloud racks.
NextDC and the AI-Ready Buildout
NextDC has similarly positioned itself at the center of the AI infrastructure cycle. The company has developed AI-ready data centres and is issuing convertible notes to fund capital expenditure aligned with AI demand. In September 2026, NVIDIA announced partnerships with eight Australian cloud and infrastructure companies—including NextDC, AirTrunk, CDC, and Firmus—targeting a buildout of up to 2 GW of AI-related capacity by 2027. These partnerships are not merely supply agreements; they represent a deepening integration between chip architecture, cooling design, and data centre operations.
Liquid Cooling and the Grid Connection Dividend
The shift to AI workloads has a direct bearing on how operators use secured grid capacity. Schneider Electric’s analysis notes that cooling efficiency directly affects how much power is available for compute: lower facility overhead can increase the share of a fixed connection available to IT, making PUE and reliable compute per secured megawatt critical capacity measures. In Western Sydney, where Transgrid has already allocated about 1.5 GW of existing network capacity to signed agreements, the ability to stretch every secured megawatt through liquid cooling and efficient electrical architecture becomes a competitive advantage.
The Hyperscaler Shadow
Global hyperscalers continue to anchor demand. Microsoft has committed A$25 billion (US$18 billion) to build data centres and digital resources in Australia to support AI development, including a facility at Kemps Creek in Western Sydney. AWS has contracted 990 MW of renewables across 20 projects to support its A$20 billion Australian data centre investment. For domestic operators and their infrastructure partners, the hyperscaler presence is both a demand signal and a competitive pressure: it validates the market while raising the bar for power procurement, sustainability credentials, and delivery speed.
The operator story in Australia is therefore not simply one of expansion. It is a story of re-engineering—around AI density, liquid cooling, renewable energy procurement, and grid-constrained capacity. The companies that master these variables will define the next decade of Australian digital infrastructure.