In its judgment of July 9, 2026 (X ZR 71/24 – Kombinationsgerät), the German Federal Court of Justice (BGH) revoked a European patent that was undone by its own priority document. The decision carries a brief headnote on utility model law – but its real lesson lies in the constellation in which that headnote became decisive.

The patent at issue, EP 2 078 113, concerns a combination device for tasks such as snow blowing, mowing or scarifying, in which the housing can be rotated around the respective working tool and fixed in different operating positions. It was filed on November 7, 2007, claiming priority from a German utility model of November 7, 2006. The utility model had already been entered in the register on October 11, 2007 – almost four weeks before the European application was filed.

The patent failed on priority. The utility model application disclosed only a configuration in which housing and working tool rotate coaxially. Claim 1 of the patent, by contrast, merely required the housing’s axis of rotation to run parallel to that of the working tool. In the Court’s view, such a generalization was not foreshadowed in the earlier application; a parallel but non-coaxial arrangement would generally have required further design changes. Nor did it help that the utility model claim referred to a rotatable housing without specifying the axis: leaving the orientation open does not disclose any particular one.

With priority lost, the utility model itself became full prior art – because, as the headnote holds, a utility model application is available to the public from the date of its entry in the register. Since the coaxial embodiment of the utility model falls within the broader claim of the patent, the BGH found a lack of novelty. The same mechanism is likely to apply to patent applications having effect in Germany that serve as the priority basis, even where they are published only after the subsequent filing: as earlier rights, they are relevant to novelty only – but novelty is precisely where a generalizing subsequent filing fails if the earlier application discloses an embodiment falling within the broader claim.

On the priority issue, the decision departs in substance from G 1/15, in which the Enlarged Board of Appeal largely defused the problem of self-collision with one’s own priority application. Under G 1/15, partial priority may not be refused for a claim encompassing alternative subject-matter by virtue of a generic expression, insofar as that subject-matter is disclosed directly and unambiguously in the priority document. The Court takes a narrower view: claim 1 does not encompass alternatives, but provides only for the parallel arrangement; the fact that it also covers the coaxial arrangement makes no difference. Priority could only be claimed for a version of the claim limited to the coaxial configuration. That is the all-or-nothing approach that G 1/15 abandoned for generic expressions. Had the case been assessed in line with G 1/15, the coaxial subset would have been entitled to priority, and the patent would at least not have failed for lack of novelty over the utility model.

For practice, this means: anyone generalizing in a subsequent filing relative to the priority application should not rely on partial priority, at least in German nullity proceedings, and should always have an auxiliary request ready in litigation that is limited precisely to the subject-matter covered by the priority. Where patent applications serve as the priority basis, withdrawing them before publication can also prevent them from becoming prior art in the first place.