Quick answer: The R&D Tax Incentive reaches the experiments inside a cleantech project — electrolyser efficiency trials, novel cell or membrane chemistry, degradation and failure-mode work whose outcome cannot be known in advance — not the plant that houses them. Standard construction, commissioning, procurement and commercial production are generally supporting or excluded, and where the plant produces a tangible saleable output the feedstock rules (Subdiv 355-H) can apply, but only on their specific conditions. You self-assess; the money is in isolating the core activities from the commercial build.
20 July 2026 — the 2026–27 Federal Budget proposed R&DTI changes that would apply to income years starting on or after 1 July 2028. These reforms are proposed, not current law, and the current legislation continues to apply until then.
If you are a founder, project developer or CFO staring at a first-of-a-kind electrolyser or a 200 MWh battery, the tempting logic goes like this: the whole thing is unprecedented, first in Australia, nobody has built exactly this before — so surely the whole capex is R&D.
It is the single most expensive assumption in cleantech. The R&D Tax Incentive does not test whether your project is novel. It tests whether specific activities have an outcome that a competent professional could not have known in advance, and it strips out the commercial plant, the routine engineering and the value of anything you sell.
Over-claim the whole build and you invite a review; scope it properly and a genuinely experimental core can be defensible. This article is written from the perspective of a Registered Research Service Provider. It is general information, not tax advice, and eligibility must be self-assessed.
The Short Answer: The Experiment Qualifies, the Plant Doesn't
The R&DTI splits your spend into three buckets, and getting that split right is the whole game.
Core R&D activities — experimental activities whose outcome cannot be determined in advance on the basis of current knowledge, and that proceed through a systematic progression of work for the purpose of generating new knowledge.
Supporting R&D activities — activities directly related to a core R&D activity. A stricter dominant-purpose test applies to activities that produce, or are directly related to producing, goods or services, and to activities supporting excluded core activities.
Excluded and commercial work — standard construction, commissioning, procuring commercial equipment, routine scale-up and commercial production. These are generally not experiments, although some may qualify as supporting activities depending on the facts.
The key distinction: a first-of-a-kind demonstration plant is a vehicle for experiments. It is not, merely because it is new, a core R&D activity in its entirety.
What the R&DTI Actually Tests: The “Unknown Outcome” Bar
The statutory test is not “is this new to us?” or “is this the first in Australia?” It is whether the outcome of the activity could have been known or determined in advance by a competent professional applying existing knowledge.
If a competent electrolyser engineer, reading the current literature and vendor data, could reasonably predict the result, it is not core R&D — however expensive or novel-feeling the build.
A genuine technical unknown — for example, a hypothesis about membrane durability under variable renewable input that current knowledge cannot resolve.
A systematic progression of work — designed experiments, instrumentation, recorded observations, evaluation and conclusions that inform the next iteration.
Generation of new knowledge — the purpose is to learn something not already reasonably available, not merely to prove a known design works at your site.
This is the Australian test under Division 355. It is self-assessed and hypothesis-driven. The excluded-core rules in s 355-25(2) are separate from the supporting-activity dominant-purpose test in s 355-30(2), and they should not be conflated.
Green Hydrogen: Which Electrolyser Activities Are Core, and Which Are Supporting?
Hydrogen projects are unusually rich in candidate core activities — and unusually easy to over-claim, because the balance-of-plant can dwarf the experimental cell work in dollar terms.
Likely candidate core activities — self-assess against the unknown-outcome bar
• Experiments to raise stack or cell efficiency beyond what current data predicts.
• Novel catalyst, membrane or electrode chemistry development and testing.
• Stack degradation, contamination and failure-mode studies under real, variable renewable input.
• Control-strategy experiments where the response to intermittent power is a genuine unknown.
Typically supporting or excluded
• Balance-of-plant piping, compression, gas handling and water treatment performed as standard engineering.
• Standard EPC construction and civil works.
• Commissioning and reliability runs that confirm a known design performs.
• Routine scale-up of a design already demonstrated at pilot stage.
The dividing line is the hypothesis. Building a compression skid to a known specification is engineering. Running an instrumented experiment because you genuinely cannot predict how a new membrane degrades under Australian solar variability is a candidate core activity.
Grid-Scale Batteries: Novel Chemistry and Degradation Modelling vs Buying and Installing Cells
Grid-scale storage compresses the same lesson. Procuring commercial LFP packs and integrating them to a known design is unlikely to be core R&D on its own, no matter how large the installation — although some work may qualify as a supporting activity depending on the facts.
Activity
Likely classification
(self-assess)
Novel cell or electrolyte chemistry development
Candidate core
Thermal and degradation modelling with a genuine unknown outcome
Candidate core
BMS or control-algorithm experiments for an unresolved technical problem
Candidate core
Procuring commercial battery packs
Not core; may be supporting on the facts
Civil works, containerisation and grid connection to a standard design
Not core; supporting or excluded on the facts
Standard integration and commissioning
Not core; supporting or excluded on the facts
The test is identical: is there a technical unknown that a systematic experiment is designed to resolve? Degradation modelling can be core when the outcome genuinely cannot be predicted, and merely descriptive when you are confirming vendor curves.
The Feedstock and Saleable-Output Trap
The moment a plant produces something with market value — hydrogen that is sold or electricity exported to the grid — the R&DTI feedstock provisions in Subdivision 355-H may become relevant.
The feedstock adjustment does not simply cut the notional deductions. Where the statutory conditions are met, it includes an additional amount in assessable income. Those conditions broadly require a tangible product of the R&D activities, notional deductions for feedstock inputs transformed or processed by those activities, and that output being supplied to another entity or applied to the claimant's own use.
Feedstock adjustment — where the conditions are met, an amount is added to assessable income by reference to the feedstock revenue and related notional deductions. It is not triggered merely because a plant produces something saleable.
Clawback or recoupment on grants — Government support can create a separate clawback interaction that changes the net value of the offset, but what matters is whether what you received is a qualifying government recoupment rather than which agency provided it — ARENA and state cleantech programs typically award grants, while the Clean Energy Finance Corporation generally provides finance or investment rather than grants.
The practical discipline: separate experimental runs from production runs in both the engineering plan and the records. If you cannot distinguish them on paper, neither can a reviewer.
Worked Example: A First-of-a-Kind Electrolyser Demonstration
Consider a hypothetical $40 million green-hydrogen demonstration project. The figures below are illustrative only and are not a promise of any benefit.
Core R&D
A designed programme testing a new membrane and stack configuration whose degradation behaviour under variable solar input is genuinely unknown: instrumentation, experimental runs and analysis. This is likely to be a modest slice of total project spend.
Supporting R&D
Activities directly related to those experiments, such as a purpose-built test rig. Where an activity produces or is directly related to producing goods or services, or supports an excluded core activity, it must satisfy the dominant-purpose test in s 355-30(2).
Non-core
EPC construction, compression, water treatment, commissioning and routine scale-up. These are not core R&D and are typically the majority of the build, although some may qualify as supporting activities depending on the facts.
Feedstock-adjusted output
Where the conditions in Subdivision 355-H are met, hydrogen produced and supplied, or applied to the claimant's own use, may bring an amount into assessable income under the feedstock rules rather than reducing the notional deductions directly.
Even in a genuinely novel plant, the claimable core is usually a minority of total capex — and that is the correct, defensible result.
Offset Rates and Thresholds
The refundable offset is the company tax rate plus 18.5 percentage points — 43.5% for a 25% base-rate entity — and is available to eligible companies with aggregated turnover under $20 million that are not controlled by one or more income-tax-exempt entities.
The non-refundable offset is the company tax rate plus an intensity premium: an additional 8.5 percentage points on notional deductions within the 2% R&D-intensity threshold and 16.5 percentage points above it. Unused non-refundable amounts are generally carried forward rather than paid as cash.
Important: the $150 million figure is an R&D expenditure threshold, not a cap. Notional deductions above that level generally attract an offset at the company tax rate rather than the intensity premium.
Getting It Right: Registration, Records and Where an RSP Fits
AusIndustry registers the activities; the ATO administers the offset claimed in the company's tax return. Activities must generally be registered within 10 months after the end of the income year.
Eligible entity — broadly, the claimant must be an eligible company. A company exempt from income tax is not an eligible R&D entity.
Minimum expenditure — the usual threshold is $20,000, although eligible R&D conducted through an RSP may be claimable below that threshold.
Contemporaneous evidence — hypotheses should be stated before the experiment, results recorded as work proceeds, and expenditure mapped to specific core activities.
Ignition Research is a Registered Research Service Provider (RSP000047) based at Lot Fourteen in Adelaide. An RSP can supply research capability and help structure and substantiate the experimental core. It does not approve, pre-approve or guarantee a claim. Eligibility remains self-assessed.
Frequently Asked Questions
Q: Does building a green hydrogen demonstration plant qualify for the R&D Tax Incentive?
A: The plant itself generally does not. What may be eligible are the experimental activities inside it — efficiency, chemistry and degradation work whose outcome cannot be known in advance. Standard construction, commissioning and production are supporting or excluded. Each activity must be self-assessed.
Q: Is grid-scale battery storage R&D eligible, or just the construction cost?
A: Construction and procuring commercial packs are unlikely to be core R&D on their own. Novel cell chemistry, genuine-unknown thermal or degradation modelling, and control-algorithm experiments may be candidate core activities. Installing cells to a known design is generally not core R&D.
Q: What cleantech activities are excluded, supporting or non-experimental?
A: Standard EPC construction, civil works, balance-of-plant to known specifications, procuring commercial equipment, commissioning, routine scale-up and commercial production. Some may qualify as supporting activities if directly related to core R&D and, where required, undertaken for the dominant purpose of supporting that core activity.
Q: How does selling the hydrogen or electricity affect a claim?
A: It may trigger a feedstock adjustment, but not automatically. Subdivision 355-H applies only where its specific conditions are met. Where it applies, an amount is added to assessable income rather than directly reducing the notional deductions. Grants may separately create clawback interactions.
Sources & Further Reading
Overview of the R&D Tax Incentive — business.gov.au
Apply to register with the R&DTI — business.gov.au
Rates of the R&D tax incentive offset — ato.gov.au
Refundable and non-refundable offsets — ato.gov.au
Proposed reforms to the R&D Tax Incentive — business.gov.au
Income Tax Assessment Act 1997, Division 355 — legislation.gov.au
Related: R&D Tax Incentive for renewable energy
Related: What does not qualify
Scoping a hydrogen or battery project is a scoping problem before it is a tax problem. The defensible claim is the one where the experimental core is cleanly separated from the commercial plant, on paper, before construction begins. Ignition Research helps structure and substantiate the R&D without over-reaching on the whole capex.
Note: this article describes the current rules. Changes proposed in the 2026–27 Federal Budget for income years starting on or after 1 July 2028 are not yet law.
This article is general information from a Registered Research Service Provider about the R&D Tax Incentive. It is not tax, legal or financial advice; eligibility depends on your circumstances and you should self-assess and seek your own advice.
