Zerra DC has filed plans for a very large data center campus in Queensland’s Western Downs region. The August 24 market source describes the proposed Western Downs Digital Park as a 1.44 GW development on a 725.5 hectare site around 37 kilometers northwest of Dalby. Data Center Dynamics reported on August 24, 2026 that the project could be delivered across four phases and sits close to a major substation and several gas and renewable generation assets.
Current reporting uses different capital cost figures depending on the assumed scope of full buildout, so the power scale is the more useful anchor for understanding the project. At 1.44 GW, the proposed campus belongs in the category of infrastructure that must be planned alongside regional energy systems rather than treated as another commercial building. Sensaka’s guide to data center construction cost shows why projects of this size depend on much more than server procurement.
Site selection is increasingly about access to an energy ecosystem
The proposed site is close to the Braemar substation and several nearby generation facilities. That geography is central to the project story. AI data centers can require such large blocks of electricity that developers are increasingly looking for locations where generation, transmission and land can support expansion together.
A large grid connection on paper is not enough. Developers need to understand how much capacity can be delivered, when it can be energized and what upgrades are required. They also need backup and power quality strategies that can handle sudden workload changes from dense GPU systems.
Sensaka’s data center power calculator works at rack scale, but the principle is identical: planning should be based on usable electrical capacity with continuous load headroom rather than a headline maximum.
Phased construction is a response to infrastructure uncertainty
Data Center Dynamics reported that the campus is planned in multiple phases. That is a practical way to align building construction with customer commitments, power availability and equipment delivery. A 1.44 GW campus is unlikely to appear as one fully operational block on a single day.
Phasing also reduces some financial risk. Developers can bring capacity online as infrastructure becomes available rather than committing the entire capital program before demand is proven. The tradeoff is operational complexity because each phase can have different equipment generations, cooling designs and network requirements.
This is where AI era data center capacity planning becomes important. The useful unit is deployable capacity after power, cooling, network and operational constraints are considered together. A future phase may have land reserved while still lacking the electrical or thermal systems needed to support actual AI racks.
Cooling design will decide how much of the electrical capacity becomes compute
Large AI campuses convert enormous amounts of electricity into heat. The proposal has been reported as using primarily air cooling with closed loop water systems and other measures. The final thermal architecture will matter because cooling overhead changes how much of the site’s electrical capacity can be delivered to IT equipment.
High density accelerator deployments may also push individual halls toward more advanced liquid cooling even if other parts of the campus remain air cooled. That creates a mixed facility where operators have to monitor different thermal systems and understand how each affects capacity.
Cooling is also connected to local resource questions. Water use can influence public support and permitting, while more electrically intensive cooling can increase demand on the grid. No design eliminates every tradeoff.
Gigawatt scale campuses are becoming regional infrastructure
A project measured in more than one gigawatt affects more than its owner. It can influence transmission planning, generation investment, local construction markets, water infrastructure and community expectations. It may also attract suppliers and other digital infrastructure around the site.
That makes transparency important. Developers need credible schedules for how quickly power demand will ramp. Utilities need to know which phases are committed. Communities need understandable information about water, noise, land use and employment. Investors need to distinguish between announced capacity and capacity likely to generate revenue.
The Western Downs proposal is another sign that the AI infrastructure race is expanding beyond traditional data center hubs. The winning locations may be those that can combine land, power, permitting and operating capability in one development path. At 1.44 GW, Zerra’s proposal is best understood as both a data center project and an energy infrastructure project.
Originally published on the Sensaka blog.
