UPC warns NVIDIA on conditional appeals in ParTec patent dispute

The Unified Patent Court’s Court of Appeal has reinforced an important procedural point for defendants relying on conditional patent revocation counterclaims: if an infringement claimant appeals a finding of non-infringement, the defendant should file its own conditional appeal to preserve the revocation issue in case the infringement finding is overturned.

The guidance came in an order concerning the continuing dispute between ParTec AG and NVIDIA GmbH and NVIDIA Corporation over European patent EP 3 743 812. The Court of Appeal granted NVIDIA’s application for reinstatement of its right to appeal after it missed the applicable deadline, but the case highlights the need for a defendant to protect its conditional revocation counterclaim through a separate conditional appeal.

The order was issued on September 28 in NVIDIA GmbH and NVIDIA Corporation v. BF exaQC AG and ParTec AG, UPC-CoA-140/2026.

Conditional revocation counterclaim

ParTec sued NVIDIA before the UPC’s Munich Local Division in March 2025, alleging infringement of EP 3 743 812, titled “Application runtime determined dynamical allocation of heterogeneous compute resources.” NVIDIA filed a counterclaim seeking revocation of the patent. The counterclaim was made conditional on the infringement action succeeding.

The Munich Local Division subsequently found that NVIDIA did not infringe the patent. Because the condition attached to NVIDIA’s revocation counterclaim had therefore not been fulfilled, the Local Division did not decide the counterclaim. ParTec appealed the non-infringement decision, but NVIDIA did not file its own appeal within the applicable period.

NVIDIA later applied for reinstatement of its right to appeal. ParTec did not oppose the application, and the Court of Appeal granted it.

The procedural issue is significant because the conditional revocation counterclaim does not simply disappear when the first-instance Court finds no infringement.

Separate appeal needed for revocation

The Court of Appeal had already addressed the issue in its July 2026 order in Emboline v. AorticLab.

In that case, the Court confirmed that a counterclaim for revocation may be made conditional on a finding of infringement. It also held that infringement proceedings and a revocation counterclaim constitute separate actions under Article 32(1) of the UPC Agreement.

Consequently, where the infringement claimant appeals a finding of non-infringement, the counterclaimant has a legitimate interest in appealing the first-instance court’s decision not to decide the revocation counterclaim.

The Court specifically held that the counterclaimant may lodge that appeal conditionally on the infringement claimant lodging an appeal. If the infringement claimant does not appeal, the condition is not fulfilled, and the conditional appeal is treated as not having been lodged, with no appeal fee becoming due. If an infringement appeal is filed, the conditional revocation appeal becomes operative, and the applicable appeal fee must be paid.

The NVIDIA proceedings illustrate the practical importance of that mechanism. Because NVIDIA did not initially file such an appeal, it subsequently had to seek reinstatement of its appeal rights after ParTec challenged the non-infringement finding.

What happens if infringement is found on appeal?

The procedural distinction matters because the revocation counterclaim remains conditional during the appeal. If the Court of Appeal overturns the first-instance finding of non-infringement and finds infringement, the condition attached to the revocation counterclaim is fulfilled.

The Court of Appeal has previously indicated that, in those circumstances, it will generally decide the revocation counterclaim itself rather than refer it back to the Court of First Instance. That follows from Article 75(1) of the UPC Agreement and Rule 242.2(b) of the Rules of Procedure.

This means that a defendant relying on a conditional revocation counterclaim needs to consider the appeal position even where it prevailed on infringement at first instance.

A finding of non-infringement may mean that the validity issue does not need to be decided immediately. But if the infringement claimant appeals, the defendant cannot necessarily wait to see what happens before protecting its own position on validity.

A developing UPC procedure

Conditional revocation counterclaims have become an increasingly important procedural mechanism at the UPC because they allow a court to avoid deciding patent validity where it concludes that the asserted patent has not been infringed.

The Court of Appeal confirmed in Emboline v. AorticLab that such conditional counterclaims are permissible. It also clarified that the condition is inter-procedural because it links two separate actions—the infringement action and the revocation counterclaim. That distinction is important when determining whether a party must file a separate appeal or can rely on a cross-appeal.

The Court has also made clear that a conditional revocation counterclaim remains pending after a first-instance finding of non-infringement for as long as the decision remains capable of being appealed. The reference to a finding by “the Court” ordinarily encompasses both the Court of First Instance and the Court of Appeal.

The NVIDIA order therefore provides a practical reminder for UPC litigants: a defendant that makes its revocation counterclaim conditional should consider filing a conditional appeal if the infringement claimant challenges a first-instance finding of non-infringement.

For NVIDIA, the Court’s grant of reinstatement means the company can pursue its appeal despite missing the original deadline. But the procedural history demonstrates why parties should address the appeal status of a conditional revocation counterclaim at the outset of the appellate stage, rather than assuming that the counterclaim will automatically follow an appeal of the infringement decision.

← Back to News