Context: In Unified Patent Court (UPC) infringement actions, there is almost always some kind of invalidity defense, most frequently in the form of a revocation counterclaim.
What’s new: Today the UPC’s Hague Local Division (LD) decided (PDF) Winnow v. Orbisk, an infringement case over a food waste management patent. The patent was deemed invalid as granted, and the auxiliary requests for which there were infringement allegations on the table failed as well. All that was left of the patent-in-suit, EP3198245 (“A system and method for monitoring food waste”), was a narrowed (but not alleged to be infringed) version. The court clarified what implications this pattern has for the allocation of costs. For the most part, the unsuccessful plaintiff bears the costs, but 15% of the costs of the revocation-related proceedings were imposed on the defendant for unsuccessfully opposing an auxiliary request outside the scope of the infringement action.
Direct impact: Winnow will probably appeal its loss. For now, Orbisk is off the hook and receives a reimbursement of 100% of its costs related to the infringement proceedings and 85% of the costs related to the revocation proceedings. The parties had previously agreed upon the cost level, but it was for the court to allocate them based on the outcome.
Wider ramifications: A headnote clarifies that a defendant to an infringement claim will be considered to have prevailed 100% on its counterclaim for revocation even if the patent survives in a form for which the patentee never alleged infringement, as the objective of the counterclaim will have been achieved in full. But if the defendant overshoots and attacks claims or auxiliary requests for which there was no infringement allegation in the case, then it must bear the costs to the extent those additional invalidity contentions are rejected.
Transition to freemium model: Please note that this article would normally be available only to subscribers. But we do recognize that during the main vacation season in UPCland, many readers would find it hard to make internal decisions on a premium membership, and we didn’t want to withhold this article from them. Even the organizations that already have a master account may not be able to create subaccounts quickly these days. That said, if you haven’t reached out yet, please contact us by email (sales at this website’s domain) sooner rather than later. If you express an interest, we’ll promptly provide you with a provisional account (which comes with no financial obligations; the decision can still be made next month).
The path to the decision on the revocation counterclaim starts with a video that Winnow itself published ahead of the patent’s priority date. It was apparently undisputed that all claim limitations apart from the last one, which involved a server not seen in the video, were disclosed by the video. The court then combined the video with common general knowledge to fill in for the last claim limitation, concluding that there was no inventive step, and turned to the auxiliary requests.
The first auxiliary requests were designed to salvage the patent in a form where there would have been infringement allegations in the case. They failed, partly for lack of an inventive step and partly due to added matter (lack of support in the written description for the term “erroneous categorization”).
Auxiliary request 3 succeeded because it added a claim limitation the court deemed to involve an inventive step and for which it found sufficient support in the description. However, Winnow never even tried to argue that Orbisk infringed the patent in that form. Therefore, no infringement analysis had to be performed in the first place.
Cost allocation
Thanks to a stipulation, the court did not have to resolve a dispute over the reasonableness of the parties’ costs. The pretrial stipulation left the door open to raising certain arguments at the oral hearing, but that did not happen. What the court was called upon to determine, however, was the allocation of the undisputed costs.
The court was clearly guided by the notion that, in other words, the defensive purpose of a revocation counterclaim in the infringement context is to dispose of the infringement claim, which is consistent with the philosophy of the Court of Appeal in a recent decision on security (June 20, 2025 ip fray article). Anything that goes beyond, such as challenging unasserted claims or auxiliary requests involving unasserted claims, is no longer part of the infringement defense. It becomes an offensive action that seeks to do away with patent claims the infringement defendant does not have to deal with at this point. That is obviously allowed, but it has cost implications.
A revocation counterclaimant optimizing for a full cost reimbursement would therefore have to refrain from opposing auxiliary requests with respect to which the patentee did not allege infringement, and focus only on those patent claims that must be defeated in order to dispose of the main action (the infringement claim).
The court’s reasoning is compelling. The patentee provokes a revocation counterclaim only to the extent it actually alleges infringement. If it just wants to salvage the patent in a form too narrow to pose a threat to the defendant in the present case, then the revocation counterclaim is no longer co-extensive with the infringement claim. If it succeeds, fine. If not, the counterclaimant bears the costs.
Prior to this clarification some parties may have thought (and until there is appellate case law, may still think) it prudent to attack all auxiliary requests so as not to be considered to have succeeded on only part of the revocation counterclaim, which would also have cost implications. The Hague LD’s reasoning makes sense, though.
Great writing and layout
The administration of justice is not a beauty contest, but is worth noting that in terms of layout, this decision is at a level with a recent decision by the Nordic-Baltic Regional Division (RD) (item 3 of our August 3, 2025 UPC Roundup).
It is exceptionally easy to read. One can understand everything without even having to look up any external documents. Everything that is relevant is quoted, and pictures are included as well. Also, the way the court juxtaposes the two parties’ arguments on an issue-by-issue basis is clearly preferable over the traditional German judgment structure where several paragraphs in a row, and often multiple pages, summarize only one party’s position, often without indicating (other than by means of headlines) whether the court is expressing its own holdings and findings or merely paraphrasing those of a given party. That can lead to confusion, and some could even try to take advantage of it by quoting out of context to mislead.
It would be desirable for the UPC to develop a uniform judgment style in terms of both layout and structure. This decision would be a good starting point unless there was a disagreement on whether different text colors should be used at all (a question on which reasonable people can, of course, disagree). President Grabinski described the UPC as a startup. It is increasingly a grownup, and at some point it may want to think of a consistent “corporate” design with common templates.
Court and counsel
Panel: Presiding Judge Edger Brinkman, Judge Margot Kokke, Judge Petri Rinkinen (Helsinki, Finland) and Technically Qualified Judge Alessandro Sanchini.
Counsel for Winnow: Bristows’s Gregory Bacon, Myles Jelf and Eden Winlow.
Counsel for Orbisk: Hoyng Rokh Monegier’s Roeland Grijpink, Theo Blomme, Niels Zelders and Anouschka Heemskerk.