GreenSquare AI Campus Australia: Powerful Upgrade
Author: Crazyrites Editorial Team
Date: October 19, 2026
Introduction
Australia is taking another step towards becoming an important centre for artificial intelligence infrastructure. GreenSquare, an Australian data centre company, has brought the first stage of its SYD1 data centre campus in Western Sydney into operation. The project has turned a facility that originally opened in 1995 into a modern site designed to support demanding AI computing workloads.
The development is notable because it combines advanced technology with the reuse of an existing building. Instead of demolishing the old facility and starting again, GreenSquare redesigned the site to support high-density computing. Its first customer, Sharon AI, plans to deploy approximately 8,200 NVIDIA Blackwell Ultra graphics processing units (GPUs) at the facility.
The GreenSquare AI Campus Australia project reflects two growing trends in technology: rising demand for AI computing power and the search for more efficient ways to build digital infrastructure. It also raises important questions about electricity, cooling, water use and the future of data centres.
What Is the GreenSquare AI Campus Australia Project?
GreenSquare’s SYD1 campus is located in Norwest Business Park in Western Sydney, New South Wales. The first stage uses an existing data centre building that dates back to 1995. According to industry reports published on October 19, 2026, the redevelopment was completed in approximately eight months, with major construction work beginning in February 2026.
The GreenSquare AI Campus Australia facility has been redesigned to handle high-density AI computing. These workloads need powerful processors, reliable electricity supplies and advanced cooling systems. Therefore, upgrading an older data centre requires more than installing new computer equipment. The building’s layout and supporting systems must also meet modern requirements.
GreenSquare retained important parts of the existing structure while moving the data halls to the ground floor. This change helped the company make better use of the available space and accommodate higher-density computing without adding structural reinforcement.
The company reported that IT capacity within the original building footprint increased by 275%. In simple terms, the redesigned facility can support substantially more computing capacity within the same building area.
NVIDIA Blackwell Ultra GPUs Take Centre Stage
One of the biggest features of the GreenSquare AI Campus Australia development is its planned deployment of approximately 8,200 NVIDIA Blackwell Ultra GPUs by Sharon AI.
GPUs are specialised processors that can perform many calculations at the same time. This makes them useful for training AI models, running large language models and handling other computing tasks that require significant processing power.
NVIDIA’s Blackwell Ultra platform is designed for demanding AI workloads. The deployment at SYD1 is expected to use NVIDIA HGX B300 infrastructure, making it a significant announced installation of this technology in Australia.
Sharon AI is the first customer for the facility’s initial stage. The company provides AI computing infrastructure to organisations that need access to powerful systems without building and operating every part of the infrastructure themselves.
However, it is important to understand that the planned deployment of 8,200 GPUs does not mean every system is already installed and running at full capacity. The infrastructure is being deployed as part of the campus rollout, and the final operational performance will depend on installation, testing and customer requirements.
Key Details at a Glance
| Feature | Details |
|---|---|
| Company | GreenSquare |
| Project | SYD1 data centre campus |
| Location | Norwest Business Park, Western Sydney, Australia |
| Original facility | Opened in 1995 |
| First stage | Now operational |
| First customer | Sharon AI |
| Planned GPU deployment | Approximately 8,200 NVIDIA Blackwell Ultra GPUs |
| Infrastructure | NVIDIA HGX B300 |
| Redevelopment period | Approximately eight months |
| Reported IT capacity increase | 275% within the original building footprint |
| Proposed next stage | 80 MW |
| Main focus | High-density AI computing |
These details show why the GreenSquare AI Campus Australia development matters beyond a single building upgrade. It combines a large AI hardware deployment with a different approach to data centre construction.
Why Reusing an Old Data Centre Matters
Building a new data centre can require significant construction work, materials, land and time. Reusing an existing facility may reduce some of these demands, provided the building can safely support the new equipment and its technical needs.
GreenSquare chose to retain much of the original building structure rather than demolish it. An independent assessment by engineering consultants TTW reportedly found that the project retained 92% of the existing structural carbon.
The company also reported that the approach, combined with low-carbon concrete, delivered 76% lower structural upfront carbon per square metre than a comparable new-build facility. It reported a 93% reduction in structural upfront carbon per kilowatt of IT capacity compared with a new-build equivalent.
These figures relate to the structural carbon associated with construction, rather than the total environmental impact of running the data centre over its lifetime. Electricity consumption, hardware manufacturing and cooling requirements still matter.
Nevertheless, the GreenSquare AI Campus Australia project offers an example of how existing digital infrastructure may be adapted for newer technology. This approach could become useful in areas where older buildings have suitable locations, power access and structural conditions.
Cooling and Water Use Are Important
AI data centres can produce substantial heat because thousands of processors may operate at high levels of activity. Without effective cooling, equipment may become less efficient or suffer operational problems.
To address this challenge, the GreenSquare AI Campus Australia facility incorporates liquid-to-chip cooling. This system transfers heat away from processors through liquid cooling equipment, helping support high-density computing.
GreenSquare has also described a closed-loop cooling system designed to reduce water use. According to company information reported by industry publications, the system is expected to operate without evaporative cooling for approximately 96% of the year. Rainwater harvesting is expected to offset around 10% of annual cooling water requirements.
The facility has projected power usage effectiveness (PUE) of 1.20 and water usage effectiveness (WUE) of less than 0.1 litres per kilowatt-hour. These are projected performance figures, not proof of long-term results under every operating condition.
PUE measures the total energy used by a data centre compared with the energy used by its IT equipment. A lower figure generally indicates that less additional energy is required for supporting systems such as cooling and power distribution.
WUE helps measure water use against IT energy consumption. Both measures are useful, but actual results can change with weather, workload and operating practices.
What Does This Mean for Australia’s AI Industry?
The GreenSquare AI Campus Australia project comes as businesses increasingly use AI for customer support, software development, research, data analysis and other tasks. These services require computing infrastructure that can process large amounts of information reliably.
A larger supply of AI-ready data centre capacity could help cloud providers serve customers in Australia and the wider Asia-Pacific region. It may also support organisations that prefer to keep certain data and computing workloads within Australia, depending on their contracts, technical setup and legal requirements.
Furthermore, projects like SYD1 can create work for construction teams, electrical specialists, engineers and data centre technicians. The redevelopment involved partners including Erilyan for construction and delivery, and Fredon for electrical infrastructure and commissioning.
However, a single project cannot meet every future AI computing need. Demand will also depend on electricity supply, network connections, hardware availability, customer investment and the cost of running large-scale systems.
Expansion Plans and Future Growth
The first stage is only part of GreenSquare’s broader plan for the SYD1 site. Industry reporting says the New South Wales Government’s Investment Delivery Authority has endorsed support for a proposed second stage of 80 megawatts.
If developed as planned, this expansion would increase the campus’s potential capacity and support further growth in AI computing. Nevertheless, the proposal should not be confused with a completed development. Construction schedules, approvals, financing and power availability will influence what happens next.
The GreenSquare AI Campus Australia project also highlights a wider challenge for infrastructure companies. They must build capacity quickly enough to meet customer demand while keeping facilities reliable, cost-effective and responsible in their use of resources.
Although AI hardware attracts most of the public attention, the buildings, electrical systems, cooling equipment and network connections behind that hardware are equally important.
Challenges and Risks to Watch
Despite its promise, the GreenSquare AI Campus Australia project faces issues common to large AI data centres.
First, electricity demand is a major consideration. Thousands of advanced GPUs require a dependable power supply, and additional computing capacity can place greater demands on local energy infrastructure.
Second, cooling performance must remain reliable during periods of high demand and changing weather. Liquid cooling can support dense computing, but it also requires careful system design, maintenance and monitoring.
Third, hardware and AI systems change quickly. Operators must plan for upgrades and future customer needs without making existing infrastructure unnecessarily expensive or difficult to adapt.
Finally, environmental claims should be assessed using clear measurements over time. Retaining an existing building can reduce construction-related carbon, but the full environmental picture also includes electricity sources, equipment replacement and day-to-day operations.
These challenges do not remove the value of the project. Instead, they show why technical performance and responsible resource use must develop together.
Conclusion
GreenSquare’s SYD1 campus is an important development in Australia’s growing AI infrastructure sector. By repurposing a facility that opened in 1995, the company has created a modern data centre designed for high-density computing and a planned deployment of approximately 8,200 NVIDIA Blackwell Ultra GPUs by Sharon AI.
The GreenSquare AI Campus Australia project also demonstrates how reusing existing buildings may help increase computing capacity while reducing some construction-related environmental impacts. Its cooling systems and expansion plans will be important parts of its future performance.
At the same time, the long-term value of the campus will depend on reliable operations, responsible energy and water use, and the needs of customers. As AI adoption continues, projects like SYD1 may contribute to Australia’s digital economy while reminding the industry that innovation works best when efficiency and sustainability receive equal attention.
Sources:
- Data Centre Magazine, “GreenSquare: Repurposing a 1995 Facility Into an AI Campus,” October 19, 2026.
- Data Center Dynamics, “GreenSquare launches first phase of data center campus in Sydney, Australia,” October 19, 2026.
- GreenSquare official announcement, “GreenSquare signs long-term AI infrastructure customer at SYD1,” September 21, 2026.
- GreenSquare, “GreenSquare’s SYD1: Building more capacity with less carbon,” June 15, 2026.
