Australia’s data centre boom: how companies building smarter can benefit through the R&D tax program

A Practical Guide to Data Centre R&D Tax Claims

Australia’s R&D tax incentive (RDTI) program, which provides a meaningful tax offset of 38.5% - 46.5% on eligible R&D expenditure, is a key inducement for the booming development of data centres here. The RDTI provides an important funding mechanism not only for AI infrastructure developers but also the many related industries such as engineering and software companies providing solutions to data centres to resolve problems or improve efficiencies.

Moreover, the recently released guidelines from the Australian government; Expectations of data centres and AI infrastructure developers indicates that many companies will be required to develop new solutions to AI infrastructure problems around energy, impacts and flexibility, requiring extensive R&D.

What sorts of R&D activities are claimable?

Like the RDTI program, experimentation to meet this emerging need is broad, focusing on both the product development and process improvement aspects of experimentation. This ranges from the development of the AI infrastructure itself including novel cooling and water systems, power system resilience, energy efficiency solutions, to related control systems including scheduling architectures and deployment frameworks.

On the process side, R&D is increasingly being undertaken in analytics and optimisation where extensive testing and evaluation is required and where outcomes can’t be predicted without experimentation. These involve the development of systems creating new knowledge rather than small scale iterative development.

Examples of data-centre R&D that commonly supports claims

• Cooling and thermal management: Airflow modelling/validation, liquid cooling trials, heat‑reuse pilots, containment designs, control tuning to reduce hot spots and fan energy
• Power and resilience: UPS operating modes, generator integration, black‑start sequencing, ride‑through optimisation, harmonics mitigation, and protection coordination under new configurations
• Energy and emissions: Demand response/peak shaving strategies, battery dispatch optimisation, renewable integration controls, PUE/WUE improvement methods requiring extensive testing
• Software and controls: DCIM/BMS algorithms, firmware development, workload scheduling to align with thermal/power constraints, anomaly detection models tested in production-like environments
• Operations: Novel commissioning methods, predictive maintenance approaches, and reliability experiments where outcomes are not known without experimentation

What sorts of costs are claimable?

While capital building costs are generally not claimable under Australia’s R&D Tax Incentive and the building itself is excluded, companies can still potentially claim the RDTI on:

• Staff and contractor costs in activities such as designing and testing novel cooling, controls, and energy optimisation etc
• Materials and consumables used in prototypes or trials
• Decline in value (depreciation) of plant and equipment used for R&D activities (for example, test rigs and monitoring equipment) for the extent they are involved in R&D activities

Some indirect costs with a strong nexus to the R&D may also be claimable.

Registering R&D activities

For RDTI purposes, the strongest claims are built around genuine technical uncertainty and a systematic experimental process (hypothesis → test → measure → analyse → conclude). In data centres this can include controlled trials of alternative cooling architectures (e.g., air vs liquid variants), containment approaches, control algorithms (DCIM/BMS), energy optimisation strategies, resilience/failover tuning, or water use minimisation—where outcomes cannot be known or determined in advance without testing.

How much and what type of documentation is required?

One of the key benefits of the RDTI is that it is a self assessed program rather than a competitive grant, but this does not lessen the documentation/substantiation required. In relation to R&D supporting documentation, ‘more is better’. It is a legislative requirement of the program that companies to maintain contemporaneous documentation and AusIndustry (on behalf of Innovation Australia) and the ATO both focus on whether evidence supports activities meeting the definition of core and supporting R&D, and whether expenditure is correctly categorised and apportioned.

Companies need to keep detailed documentation relating to both the technical and expenditure aspects of the claim, at the start, during and at the conclusion of each of the activities, describing the process of each activity as it develops.

Documentation can be varied but it’s important to ensure that the technical unknowns at the outset, including the sources investigated, are recorded. Additionally, evidence might be in the form of project plans, research, measurements taken during the activities, meeting minutes discussing outcomes, reports and photos and these needs to be maintained at all stages of the activities. Cost tracking includes timesheets relating to labour, contractor invoices and an auditable method to apportion shared costs and asset use. The documentation required will vary from activity to activity, requiring a tailored process to track activities and expenditure as R&D occurs.

Common pitfalls in data centre R&D claims

While there are many eligible areas of R&D in data centre infrastructure development, not everything is claimable. Companies need to cordon off Business as Usual activities from any R&D activities and ensure that work that does not have an unknown outcome or technical hypothesis to test, are not claimed. Specifically, areas to watch out for include:

• Treating standard design/construct/commissioning as R&D without a clear technical uncertainty and experimental method.
• Describing “optimisation” at a high level but not capturing test parameters, data, and learnings that demonstrate new knowledge.
• Over claiming assets or shared operating costs (e.g. claiming 100% of depreciation where only a portion relates to trials).
• Relying on after the fact documentation rather than keeping records during the work.
• Confusing building works (generally excluded) with plant/equipment (potentially claimable via decline in value, subject to apportionment).

Key takeaway

Australia’s RDTI can materially reduce the cost of solving hard data centre problems, particularly where teams are running structured trials around energy, cooling, controls and resilience. Both data centre infrastructure developers and companies looking to provide solutions to the industry can both benefit.

When claiming the RDTI, the best outcomes usually come from identifying technical unknowns/hypotheses early, separating business as usual from R&D, and tracking costs and asset use in a way that is simple, consistent, and audit ready.

A structured approach to identifying and documenting R&D can make a material difference to outcomes. FB Rice works with clients across the data centre ecosystem to capture eligible activity and maximise RDTI claims. Get in touch to discuss turning R&D into a tangible asset.