Kyro Card Capital: patent eligibility under the new examination guidelines

IP Australia rewrote section 5.6.8.6 of its Patent Manual of Practice and Procedure (PMPP) on 3 July 2026. This is the version that currently stands and further revision in the immediate future looks unlikely. The new text cites Aristocrat throughout. CQMS Pty Ltd v Joy Global Surface Mining Inc [2026] APO 19 heard by Delegate Andrew Burgess was the first published decision under it. Kyro Card Capital Pty Ltd [2026] APO 31, decided on 22 October 2026 by Delegate Dr W.E. Guinea, is the second major decision.

So much has been written and said about the Aristocrat decisions, which is covered elsewhere. The Aristocrat trilogy was on its face a sharp rebuke to a large number of Federal Court decisions from the last two decades that had shaped Patent Office practice on computer-implemented inventions. What is worth noting is that the Patent Office was not a passive observer in litigation but very often involved as an active counterparty in to unsuccessful patent applicants, often filing successive appeals against decisions when they went in favour of patent applicants. So to some extent, the Patent Office could be said to have shaped the law that has emerged from the courts. Aristocrat set aside a lot of former patent law on patent eligibility in Australia that had developed in these cases in favour of 'established principles' that had always applied to all subject matter, and was more generous to patent applicants.

Kyro is informative and worthy of attention as it further indicates how the Patent Office will be applying the current PMPP 5.6.8.6 of 3 July 2026, which can be expected to be influential of further Patent Office hearings and indeed day-to-day examination practice from examiners seeking to follow examination guidance.

Kyro decision

The application (AU 2023322214) claims a payment processing system for a product purchase reward scheme. An EFTPOS terminal stores product IDs against prices, reads barcodes through an optical scanner, totals the purchase, takes the card and sends product-level transaction data to a reward scheme server. The point is to run product-level rewards without access to the retailer's locked-down POS system.

The applicant asked for both an original claim set and later amendments made 3 June 2025 to be decided (refer paragraph[4]). These amended claims restricted the scope of the claims by adding significant further detail relating to a reward rules database, a rule controller and payment controller, a refund controller, and one-way hashing of the card's primary account number (PAN), so that the card presented for a refund can be matched to the original purchase (at [41]).

The Delegate found that claim 1 in both sets, and most dependent claims, are not a manner of manufacture (at [189]). The claims requiring PCI compliance were sent back for further examination, because it was unclear whether they fail for patent eligibility, for lack of support or enablement under s 40, or for lack of utility (at [190]). He also flagged the Square Register and PCI DSS v4.0, which lists one-way hashing of PANs, as relevant to inventive step (at [191]–[194]).

Old reasoning, new authority

The delegate's statement of principle sets the tone. He sets out RPL Central at [96]–[98] as the "considerations useful" for computer-implemented business methods, including that the invention "must lie in that computerisation" (at [87]). He then says the principles of RPL Central and Research Affiliates "assist in determining where the substance" of such inventions resides (at [88]).

He also quotes the Full Court in Aristocrat 2025, which said only that the results in those cases were correct (at [90]). An endorsed outcome is not endorsed reasoning. Kyro takes the reasoning, and PMPP 5.6.8.6 supports this view.

There is certainly some novelty creep evident in the patent eligibility analysis. The Manual says the analysis "does not involve an assessment of whether what is claimed is novel or inventive". Yet the Delegate's reasoning on the hashing integers is that hashing is known. At [164], he tells the applicant that its arguments about the novelty of hashing in this context belong to novelty. The same is true of his own reasoning. His closing remarks then reach for the Square Register and PCI DSS v4.0, which go to inventive step (at [191]–[194]).

Also, at [144], the delegate says that a device distinguished from others of its kind only by the information it contains "cannot possibly" be a manner of manufacture. That is close to a rule of exclusion. This is hard to square with the claim the Full Court allowed in Aristocrat, as the next section shows.

Two claims, side by side

The delegate invited this comparison himself. At [182], he sets out claim 1 of the patent the Full Court found patentable in Aristocrat 2025. At [184], he finds no interdependence "in the sense meant" there. Put the claims side by side and it is hard to see where the difference lies.

The delegate distinguishes Aristocrat this way. There, the new game produced "a new gaming machine that was inherently improved or altered in its essential functionality". Here, the terminal is altered only by the data stored on it (at [144]).

Interestingly paragraph 144 reads as follows: 

It is not clear to me that the analogy with Aristocrat 25 is of assistance to the Applicant.  In Aristocrat 25 there was a fundamentally new operation of the gaming system by virtue of the new game brought into being via the interaction between the various physical integers; that is, as best understood there was a new gaming machine that was inherently improved or altered in its essential functionality as a gaming machine.  The same cannot be said of the payment processing system in the present matter.  While the notional alteration or improvement is said to come from the EFTPOS terminal in terms of alleged PCI compliance, the only way the EFTPOS terminal of the payment processing system is altered is by virtue of the data stored on it – a device that is distinguished from other devices of the same kind merely by the information it contains cannot possibly provide a manner of manufacture simply by storing new or different information.  Nor does merely integrating a payment system with a reward scheme, as claimed provide for the sort of alteration that provides for patentability.  In my view this is analogous to Rokt and Encompass both of which are characterised by the sending of data between computerised devices and the processing and generation of data by general purpose computers.  Hence the interdependence per se of the features concerned does not bring into being a relevantly changed or altered payment system where that interdependence itself merely comprises standard or known uses of the technology involved in the payment system.

What changes the Aristocrat EGM is game code and symbol data held in memory. The configurable symbols and feature game are rules of play. The PMPP itself describes the improvement as one relating to player engagement. If new rules in memory made the EGM a different machine "in a general sense", and patent eligible, it is not clear why an EFTPOS terminal that now scans, prices, itemises and runs rewards without a POS system is not patent eligible.

The Kyro claim if anything has more physical engagement. The scanner feeds product IDs to the memory lookup, which drives the display and the payment, which drives the server and its rules. A refund depends on the terminal reading the card again, and the server hashing that PAN and matching it against its log. That chain is as interdependent as the EGM of Aristocrat. The Delegate's own observation that the terminal is "non-standard" (at [184]) points the same direction, yet somehow with a different conclusion.

The remaining difference is what the rules are about: a game in one case, a reward scheme in the other. NRDC, which is the foundation decision on patent eligibility in Australia, does not draw a line between subject matter. Overall it is difficult to reconcile the result here with the 'established principles' of NRDC, as endorsed by Aristocrat and especially the High Court disposition.

Why advice is hard to give

Patent applicants in Australia will often have claims that 'look like' those in Aristocrat: hardware that operates in new and unique ways under direction of a new architecture and animated by new software. An applicant with a claim like this now gets two different answers: the law pulling for eligibility and the PMPP and now the Patent Office pulling in the other direction towards patent ineligibility:

• From the law. Take NRDC and the Aristocrat trilogy as the guide, and characterise the claim as a whole. The Kyro claim is an artificial state of affairs with a useful result, much like the claim the Full Court allowed in Aristocrat.
• From the Patent Manual of Practice and Procedure. Take the PMPP and the CQMS and Kyro decisions applying the PMPP. The same claim fails, on reasoning drawn from RPL Central, Rokt and Repipe that do not follow from Aristocrat.

This divergence of opinion creates significant uncertainty for applicants, and makes advice difficult for advisors. The honest position is that the claim may well be patent eligible at law, but could well be refused in examination. And an internal appeal in the Patent Office to a hearing would seem unlikely to change the result. Getting to a favourable result on patent eligibility may mean an appeal to the Federal Court, with the cost and delay that brings. It may be a little while until a patent applicant sufficiently invested in the outcome is willing to make that move.

References

1. Kyro Card Capital Pty Ltd [2026] APO 31
2. CQMS Pty Ltd v Joy Global Surface Mining Inc [2026] APO 19‍
3. Aristocrat Technologies Australia Pty Ltd v Commissioner of Patents [2026] HCADisp 15
4. Aristocrat Technologies Australia Pty Ltd v Commissioner of Patents [2025] FCAFC 131‍
5. Aristocrat Technologies Australia Pty Ltd v Commissioner of Patents [2022] HCA 29
6. IP Australia, Patent Manual of Practice and Procedure, 5.6.8.6, as updated 3 July 2026