Fujifilm v. Kodak
UPC-CoA-473/2025, decision of 13 July 20261
This Court of Appeal decision is the latest development in a wider UPC dispute between Fujifilm and Kodak concerning Kodak’s Sonora printing plates. A different branch of the dispute has already given rise to landmark decisions2 on long-arm jurisdiction and enforcement, while this branch is now bearing fruit with guidance from the Court of Appeal on procedural issues including partial revocation, defence of dependent claims, and the interplay of EPO central limitation and UPC proceedings, and on substantive issues with further development of the UPC’s approach to inventive step.
At first instance, Fujifilm asserted EP3476616 against Kodak’s Sonora printing plates in Germany and the UK. Kodak responded with a counterclaim for revocation seeking revocation of the patent in its entirety for the territories in which it remained in force before the UPC.
At first instance, the Mannheim Local Division revoked the German part of the patent and dismissed the infringement action in respect of both Germany and the UK as a result.
A key feature of the first-instance decision was the Mannheim LD’s refusal to consider Fujifilm’s defence based on dependent claims and dependent-claim combinations, on the basis that those combinations had not been included in Fujifilm’s application to amend the patent under Rule 30 RoP.
The Court of Appeal held that the Mannheim Local Division was wrong to refuse to consider Fujifilm’s defence based on dependent claims and dependent-claim combinations. The court started from the principle reflected in Article 65(3) UPCA and Article 138(2) EPC that a patent should not be revoked in its entirety if only part of it is invalid. Consequently, the invalidity of an independent claim does not automatically mean that dependent claims are invalid.
The court drew an important distinction between defending claims as granted and seeking to amend claims.
According to the Court of Appeal, Rule 30 RoP applies only where the patentee seeks to amend the wording of the patent or proposes claim combinations that are not already explicitly encompassed by the granted claims. It does not apply where the patentee simply defends dependent claims or claim combinations already contained within the granted claim structure.
The court therefore held that a patentee need not file a Rule 30 application merely to defend dependent claims, and combinations already encompassed by granted claims can be relied upon without amendment proceedings.
However, the court also imposed an important qualification. If a patentee wishes to rely on combinations that do not directly and specifically emerge from the granted claim structure—for example because claims depend on “any preceding claim”—the patentee must clearly identify which combinations it wishes to defend, and the number of such combinations must remain reasonable.
After the Mannheim LD’s decision to revoke the patent, Fujifilm applied for and obtained a central limitation before the EPO under Article 105b EPC. The amended claim 1 corresponded to the combination of granted claims 1, 2 and 7, which Fujifilm had argued for at first instance.
The Court of Appeal accepted Fujifilm’s reliance on the central limitation granted by the EPO during the appeal. The court emphasised that the EPO’s limitation decision had retroactive effect and therefore altered the patent as if it had always existed in the limited form.
Kodak argued that the limitation should be disregarded and suggested that seeking a central limitation during UPC proceedings was procedurally problematic. The court rejected those arguments. It noted that the limited claim corresponded to subject matter that had already formed part of Fujifilm’s case at first instance. Kodak had also been informed during the appeal that the limitation request had been filed and was expected to be granted before the oral hearing in the UPC appeal. The court therefore saw no procedural unfairness and no abuse of process.
While the decision does not say why Fujifilm requested central limitation of the patent before the EPO, one likely explanation is that Fujifilm wished to put beyond doubt its ability to rely on the combination of these claims in the appeal proceedings, in view of the fact that this claim combination was not included in a specific request to amend the patent in the first-instance proceedings. However, since the Court of Appeal has now stated, as discussed above, that it is not necessary to submit a specific request to amend the patent where the proprietor wishes to rely on a combination of dependent claims already encompassed by the granted claims, the central limitation may have ultimately been unnecessary.
Nevertheless, the court did not answer the broader question of how a central limitation interacts with Rule 30 RoP, because in its view Rule 30 was not engaged where the patentee was merely relying on combinations already contained within the granted claims. The decision therefore leaves scope for future argument about whether, and in what circumstances, EPO central limitation proceedings can affect the claims considered in UPC appeal proceedings.
The judgment also contains important guidance for revocation actions.
The court emphasised that a claimant seeking revocation cannot merely ask for revocation of the patent “in its entirety” and leave everything else to the court. Under Rule 25 RoP, the claimant must identify the extent of the revocation sought and sufficiently substantiate the attack for each claim covered by the request.
This requirement was closely linked to the court’s analysis of the dependent claims. The court explained that the UPC is not required to examine every conceivable dependent claim combination on its own initiative. Rather, the parties define the scope of the dispute. Once the patentee identifies the dependent claims or combinations it wishes to defend, the revocation claimant must engage with those claims and provide a properly substantiated attack.
The judgment therefore promotes procedural clarity on both sides: patentees must identify which dependent claims or combinations they rely on and revocation claimants must explain why those particular claims are invalid.
One of the more interesting aspects of the Court of Appeal’s inventive step analysis is in the discussion of “pointers” to secondary prior art documents. The court emphasised that, where an inventive step attack relies on combining one document, representing a “reasonable starting point”, with a secondary document, there must be a reason why the skilled person would have consulted the secondary document.
The court held that the reason for consulting a secondary document must be found in the starting document itself, considered together with the skilled person's common general knowledge. Significantly, it stated that the motivation to combine cannot be derived from the content of the secondary document itself. In the court's view, if there is no pointer directing the skilled person towards a particular document, it is impermissible to rely on the content of that document to justify why it would have been consulted. Doing so would amount to hindsight because it assumes that the skilled person already knows where the solution can be found.
The court summarised the position as follows:
If there is no pointer to combine a reasonable starting point with another document, a motivation to take the next step in the direction of the invention cannot be derived from the content of such a combination document itself.
Applying this principle, the court rejected Kodak's reliance on a series of secondary references directed to image visibility and surface characteristics. Although those documents contained teachings potentially relevant to the missing claim features, the court found no reason why the skilled person, starting from the cited starting documents, would have consulted them. The starting documents already disclosed their own approaches to improving image visibility, i.e. the objective problem, principally through modifications to the image-recording layer, and therefore provided no prompt to seek a solution in the properties of the aluminium substrate surface. In the absence of such a pointer, the contents of the secondary documents could not be used to establish obviousness.
This reasoning appears somewhat stricter than the approach often encountered before the EPO. Under the EPO's problem-solution approach, it is common to ask whether the skilled person, faced with the objective technical problem, would have consulted the secondary document. The answer may be based on factors such as the similarity of the technical field, the nature of the problem addressed by the secondary document, or the skilled person's general search for solutions. By contrast, the UPC's analysis places greater emphasis on identifying a concrete reason, arising from the starting point and common general knowledge, for consulting the secondary document before its contents are taken into account.
That said, the decision should not necessarily be read as establishing a fundamentally different legal test from that applied by the EPO. The requirement to avoid hindsight and to demonstrate that the skilled person would, rather than merely could, have made the combination is well established in EPO case law. The significance of the decision lies more in the court's explicit separation of two questions: (i) why the skilled person would have consulted the secondary document; and (ii) what the skilled person would have learned from it. According to the court, the second question cannot be used to answer the first.
Ultimately, Fujifilm succeeded because the Court of Appeal accepted that the invention lay in a carefully balanced combination of interacting parameters that improved image visibility without sacrificing press life and other performance characteristics. While Kodak identified prior-art documents disclosing individual aspects of the claimed printing plate precursor, it failed to show why the skilled person would have been motivated to combine those teachings so as to arrive at the claimed combination. In the court’s view, the inventive-step attack relied on hindsight: there was no convincing pointer in the prior art leading the skilled person to modify the known printing plate precursors in the manner claimed, and the patent as limited was therefore upheld as inventive.
The decision therefore suggests that parties before the UPC should devote greater attention to identifying and articulating a clear pointer to any secondary reference relied upon in an inventive step attack. Where no such pointer can be shown, there is a real risk that the court will dismiss the proposed combination as the product of hindsight, regardless of how closely the secondary document may appear to address the claimed solution.
Fujifilm succeeded in overturning the first instance decision that the patent was invalid. The Court of Appeal found the patent in this case to be valid and infringed by Kodak. Interestingly, the outcome in this case is essentially the inverse of the outcome in the other Fujifilm v. Kodak case1, where the patent was found to be valid and infringed at first instance, but on appeal was found to not be infringed. The Kodak products accused of infringement by Fujifilm were the same in both cases, so after appeal the parties are in essentially the same situation as they were after the first instance judgments.