FY26 Year in Review — Link R&D Advisory | R&D Tax Incentive Consulting Australia

Connecting Innovation to Incentives

Home
About
Services
Pricing
R&D Knowledge
Practical Reflections

The R&D Tax Incentive in FY26: What Changed | Link R&D Advisory

You are here: Home Practical Reflections FY26 Year in ReviewThe R&D Tax Incentive in FY26: What Changed, and What Only Looks Like it Changed

The 2026 financial year was an unusually busy year for the R&D Tax Incentive. Between a landmark review, a Budget reform package, a long-running Tribunal decision and new guidance on AI, there is a lot to absorb.

There is also a lot of commentary treating all of it as though it carries the same weight. It does not. The most useful thing a claimant can do right now is separate what is law from what is an announcement.

What is actually law.
The tobacco and gambling exclusions have been brought in. Amendments to the Income Tax Assessment Act 1997 excluding R&D activities related to tobacco and gambling are finalised and in force. They apply to income years starting on or after 1 July 2025, which means FY26 claims.

The changes or exclusions are narrow:

Gambling-related R&D may remain eligible only where it is undertaken for the sole purpose of minimising harm to a person or to the Australian community.
Tobacco-related R&D may remain eligible only where it relates to the development of a therapeutic good aimed at minimising harm from tobacco use, or is intended to stop the ingestion or transfer of nicotine into the human body.

Two points that can get lost in summaries. First, the test is ‘sole’ purpose, not dominant purpose. That is a materially higher bar, and it is easy to misstate. The second is that the exclusions extend to supporting activities. This is deliberate to prevent excluded core activities re-entering the program through supporting-activity characterisation.

The drafting focuses on the nature and purpose of the activity rather than the label on the business. Companies in adjacent sectors (payments, analytics, platform technology, harm-minimisation research, cessation therapeutics) should not assume the exclusions are only the problem of other companies. If any part of your R&D touches these products or services, document the purpose of the activity properly and test it against the legislation rather than against intuition.

What is proposed, and not yet law.
Of course, the 2026–27 Budget reform package. Announced 12 May 2026. If legislated as proposed, it applies to income years commencing on or after 1 July 2028, two or three claims away. At the time of writing there is no exposure draft and no formal consultation on these measures.

The proposals:

Core R&D offset rates increase by 4.5 percentage points.
Expenditure on supporting R&D activities ceases to be eligible.
The expenditure cap rises from $150 million to $200 million.
The minimum R&D expenditure threshold rises from $20,000 to $50,000.
The aggregated turnover threshold for the refundable offset rises from $20 million to $50 million.
Refundable offsets are limited to a company's first ten years of operation.
The intensity threshold for the higher rate reduces from 2% to 1.5%.

Two of these dominate client conversations, and for good reason.

The ten-year limit on refundability. This is the one that changes the shape of an early-stage funding model. A loss-making company past the threshold could retain the offset and lose the cash refund, and for pre-revenue businesses, the refund is the benefit. The offset is an accounting outcome because the refund is working capital. A key design question remains unanswered around whether the ten-year test runs off the age of the R&D entity or the corporate group. Until that is resolved, companies with legacy holding structures cannot model their position with any confidence.

The removal of supporting activities. For a lot of claims, supporting activities are a meaningful share of eligible expenditure. Removing them raises the rate on a narrower base. Whether any given claimant is better or worse off depends entirely on their core-to-supporting mix, which is not a question that can be easily answered.

The ‘Ambitious Australia’ Strategic Examination of R&D panel delivered its final report in FY26, recommending a segmented program: a premium start-up stream, an SME and scale-up stream tied to growth outcomes, and reworked settings for large firms including removal of the expenditure cap. The Budget picked up some of this, refined some, brought in new, and left the rest. The unadopted recommendations are a reasonable guide to where pressure for further reform will come from, but they are not policy, and should not be planned around.

What has not changed at all.
Fundamentally, the eligibility criteria in Division 355 have not moved, and the evidentiary expectations have if anything hardened.

Ultimate Vision Inventions Pty Ltd v Industry, Innovation and Science Australia (Taxation) [2025] ARTA 1813 was decided in September 2025, concluding a dispute that began with a 2017 decision about activities conducted in 2014 and 2015. The procedural history alone is instructive, an earlier Tribunal decision was overturned by the Full Federal Court and remitted for re-hearing, and the matter still took until 2025 to resolve. If a dispute goes badly, it does not go badly quickly.

The substantive lessons are the familiar ones, which is precisely the point:

Contemporaneous records of hypotheses, experiments, results and analysis. Not reconstructions.
Contracting arrangements documented precisely, particularly where related entities or research service providers are involved.
Activities delineated annually. Undocumented activities in one year undermined the claimed improvements built on them in the next, the failure compounded across years.
Innovation is not experimentation. Iterative development and prototyping, however sophisticated, do not satisfy the statutory test without genuine uncertainty at the level of scientific or technical principle.

DISR's AI guidance.
DISR AI guidance was released as a subset of its software guidance that applies the same discipline to a newer context. Using a new model or technique does not make an activity eligible. Routine logging and dashboards, data cleaning by standard methods, testing with known outcomes, parameter tuning by well-understood methods and integration of known outputs through predefined logic are all unlikely to be core activities. Investigating a fine-tuning strategy not previously applied in a comparable environment, or testing an architecture where it is genuinely unknown whether it will work in the target domain, may be. The question is the same as it has always been: was the outcome unknown, and was an experiment necessary to find out?

What this all means for FY27.
The Budget measures are two years out and unlegislated. That is not a reason to ignore them, but it is a reason not to restructure around them yet.

The more productive use of FY27 is the work that pays off regardless of which version of the program arrives:

1. Get the core/supporting split right now. If supporting activities disappear in 2028, you need to know today what proportion of your claim they represent. Most companies do not.
2. Fix documentation practices while there is no deadline pressure. Every enforcement signal of the past year points the same direction. Contemporaneous records are not a nicety.
3. Test the tobacco and gambling exclusions if there is any chance they touch you. These apply to the claim you are preparing now, not in 2028.
4. Plan towards the ten-year position. Where the entity-versus-group question is decisive for you, watch for the exposure draft.

As items flow year by year on the focus, this year more so the program is being retargeted but aspects around evidence standard is consistent.

This article was posted on 15 July 2026. This article is general information, not advice, and reflects the position at the time of writing. The Budget measures described are proposals and are not law. If you would like to discuss how any of this affects your claim, please get in touch.

Book an Initial Consultation

harley@linkrd.com.au

0492 933 806