Sustainable Compute & Digital Infrastructure Systems
Designing computing systems and infrastructure within energy, water and thermal limits.
Artificial intelligence and automation are rapidly increasing demand for digital infrastructure, making energy, water and thermal constraints binding at system scale. IID works across compute and infrastructure systems to reduce resource intensity and enable coordinated, low-carbon digital capability.
Resource-Constrained Compute: Designing Within Physical Limits
The sustainability of digital systems is fundamentally shaped by the physical resources required to perform computation.
For decades, advances in computing have been enabled by treating electricity, water and cooling capacity as effectively unconstrained inputs. This has supported rapid growth in compute capability—particularly in large-scale AI systems—but has also driven increasing energy use, thermal loads and infrastructure requirements.
As AI and automation drive continued growth in compute demand, these assumptions are no longer valid. Energy, water and thermal limits are becoming binding constraints on how and where digital infrastructure can be deployed.
These dynamics directly shape the scale, location and resource intensity of data centre infrastructure.
A new paradigm is emerging in which computation is designed with explicit awareness of physical constraints. This includes:
Processor architectures optimised for performance under power constraints, including ARM and other RISC-based systems
Specialised accelerators that reduce energy per operation for AI and data-intensive workloads
Software and algorithmic optimisation to minimise compute intensity and memory movement
Distributed, edge and on-device AI that reduces reliance on centralised infrastructure
Hardware–software co-design, integrating system architecture, software and workload optimisation
Workload orchestration and scheduling aligned with energy availability and system constraints
These approaches reduce the physical footprint of computation and reshape the scale and form of infrastructure required to support digital systems.
IID is engaging with partners across the global compute ecosystem to support this transition, including work on software optimisation and open-source frameworks to improve AI performance on emerging low-power architectures such as Apple Silicon.
Infrastructure Under Constraint: Coordinating Energy, Water and Land Use
Even with improvements in compute design, large-scale digital infrastructure remains essential.
Data centres are a new class of industrial infrastructure underpinning artificial intelligence, advanced manufacturing and digital services. Their expansion is now interacting directly with constrained physical systems:
- Electricity networks with limited capacity and long lead times for augmentation
- Water systems under increasing pressure in climate-constrained environments
- Land-use systems competing across industrial, residential and environmental priorities
These constraints are no longer theoretical. In major development corridors, they are already shaping where and how digital infrastructure can be deployed.
However, these systems are typically planned and regulated independently. As a result, cumulative impacts are not well coordinated and infrastructure investment can become reactive, fragmented and inefficient.
A Precinct-Scale Approach to Data Centre Development
The impacts and opportunities associated with AI and hyperscale data centre development arise at the level of urban clusters and regional hubs, where multiple facilities share electricity, water, land and thermal systems alongside other industrial and urban users.
IID advances a precinct-scale approach that enables:
Coordination of electricity demand with generation, storage and network capacity, including behind-the-meter clean energy solutions
Integration of recycled water and climate-appropriate cooling strategies across multiple facilities and other industrial sectors
Deployment of shared infrastructure, including microgrids and precinct-scale circular thermal systems
Differentiation between metropolitan clusters and regional digital–energy– industrial hubs aligned with resource availability
This approach shifts infrastructure planning from project-by-project assessment toward integrated, system-level coordination.
The result is more efficient use of energy and water resources, reduced infrastructure risk and improved alignment between digital infrastructure growth, industrial development and long-term environmental objectives.
IID Capability in Sustainable Compute and Infrastructure Systems
The UNSW Institute for Industrial Decarbonisation (IID) operates as a pre-commercial systems integration layer, connecting research, industry and government to address complex, cross-sector industry transformation challenges.
IID’s role is to translate system-level problems into discovery research initiatives, coordinated translation programs, pilot demonstration projects and investment opportunities, drawing on UNSW’s depth in engineering, science and business.
Research Capability
UNSW brings together one of Australia’s most comprehensive research capabilities supporting sustainable compute and digital infrastructure.
IID coordinates this capability across faculties, schools and specialised centres to enable development of novel IP and industry relevant solutions.
-
Research spanning computing architectures, software optimisation and AI systems, with a focus on reducing the energy and thermal intensity of computation. This includes:
Energy- and resource-constrained computing architectures, including ARM and RISC-based systems and specialised AI accelerators
Software and algorithmic optimisation, improving performance per watt through efficient code, memory use and workload design
AI systems and distributed computing, including edge, enterprise and on-device AI that reduces reliance on centralised infrastructure
Workload orchestration and scheduling, aligning compute demand with system constraints such as energy availability and thermal limits
Open-source and platform-level optimisation, including frameworks to improve AI performance on emerging hardware platforms such as Apple Silicon
This work focuses on reducing the physical footprint of computation while maintaining performance, reshaping the scale and form of digital infrastructure required.
-
Research spanning electricity systems, on-site energy infrastructure and hybrid supply architectures for energy-intensive digital and industrial systems. This includes:
Electricity market modelling and transmission planning, including Renewable Energy Zones and integration of large, flexible loads under uncertainty
Industrial-scale renewable generation, including solar and hybrid generation systems designed for co-location with digital and industrial infrastructure
Behind-the-meter generation and storage systems, including co-located renewables, battery storage and on-site firming technologies
Advanced energy storage technologies, including long-duration storage such as redox flow batteries and integration of heterogeneous storage systems
Battery lifecycle and recycling systems, including end-of-life battery recovery, reuse and circular supply chains
Power electronics and advanced energy conversion systems, enabling efficient integration of renewable generation, storage and high-density electrical loads
Islandable microgrids and hybrid energy systems, combining grid-connected and on-site supply to improve resilience and reduce network constraints
Digital grid and control systems, including real-time monitoring, optimisation and coordination of distributed energy resources and flexible loads
Flexible load management and demand shaping, enabling data centres and industrial systems to respond dynamically to electricity availability, price and carbon intensity
Integration of compute and energy systems, including workload scheduling and control strategies aligned with energy system conditions
Precinct-scale energy system design, coordinating multiple users and infrastructure assets within shared energy environments
This work focuses on enabling resilient, low-carbon energy supply for digital infrastructure while reducing pressure on transmission networks and improving overall system efficiency. This includes co-location of renewable generation and digital infrastructure to improve energy utilisation and reduce transmission constraints.
-
Research spanning water infrastructure, advanced cooling technologies and thermal management under climate-constrained conditions. This includes:
- Recycled and non-potable water systems for industrial-scale cooling
- Discovery research in cooling technologies, including fluids, materials and heat transfer systems
- Low-water and hybrid cooling architectures suited to high-temperature environments
- Thermal load shifting and heat recovery, including integration with adjacent industrial processes
- Groundwater and geo-exchange systems, including ground-source heat pumps and subsurface thermal management
- Precinct-scale thermal networks enabling coordinated cooling and heat reuse across multiple facilities
This work supports the development of cooling strategies that are adapted to Australian climate conditions while reducing pressure on potable water systems and improving overall infrastructure efficiency.
-
Use of AI, digital twins and advanced computing systems to improve the performance of energy- and emissions-intensive industrial sectors. This includes:
Discovery and development of novel digital and control methods, generating new approaches to improving productivity while reducing energy use and emissions
Automation and real-time optimisation of industrial operations, improving performance, reducing waste and enabling continuous system improvement
Electrification of industrial processes, supported by modelling and control of variable renewable electricity supply
Digital twins and process modelling, enabling system design, testing of low-carbon configurations and reduction of capital risk
Integration of compute with physical systems, enabling predictive control, automation and system-level optimisation
Coordination of industrial loads with energy systems, aligning production with renewable energy availability and system constraints
This work improves how industrial systems are designed and operated in practice, increasing productivity while reducing energy use, emissions and cost.
-
Computing platforms for integrated modelling, digital twins and decision-support systems . This includes:
AI-supported modelling platforms, integrating energy, water, land-use and infrastructure datasets
Digital twins of precincts and infrastructure systems, enabling scenario testing and system optimisation
Probabilistic modelling of demand and infrastructure utilisation, including uncertainty in large, staged loads such as data centres
Cross-system optimisation tools, linking compute, energy and water systems within unified analytical environments
Decision-support systems for planning and policy, enabling coordinated infrastructure sequencing and investment
This capability supports a shift from fragmented, project-level assessment toward integrated, system-level planning and coordination.
Strategic Focus Areas
Advancing hardware–software co-design and distributed architectures that reduce energy, water and thermal intensity at source.
Coordinating electricity, water and land-use systems across major development corridors and regional hubs.
Designing shared infrastructure including microgrids, recycled water networks and circular thermal systems.
Supporting development of nationally significant computing infrastructure for research, industry and government.
Aligning data centre development with renewable energy and clean industry development in regions such as the Hunter and Illawarra.
Developing integrated modelling platforms to support cross-system planning, scenario testing and decision-making.
Policy Engagement and Impact
IID is actively engaged in shaping digital infrastructure policy and planning.
In its submission to the NSW Legislative Council Inquiry into Data Centres, IID identified the need for coordinated, system-level approaches to managing digital infrastructure growth, including precinct-scale planning, integration of energy and water systems and development of sovereign computing capability.
This work reflects a broader shift from treating digital infrastructure as isolated assets toward managing it as a critical, interconnected system within the economy.
Partner with IID
IID works with government, industry and research partners to design and deliver next-generation compute and infrastructure systems.
We are particularly interested in partnerships that:
- Develop novel technologies, methods and system architectures, generating new intellectual property in areas such as compute efficiency, cooling systems, energy integration and industrial optimisation
- Address system-level challenges across compute, energy, water and infrastructure
- Enable precinct-scale pilots and demonstrations to validate new approaches in real-world settings
- Advance resource-constrained, low-carbon computing systems
- Support development of sovereign AI and digital capability
To explore collaboration opportunities, contact the UNSW Institute for Industrial Decarbonisation.
