Addressing the application of the smallest salable patent-practicing unit analysis, as well as the impact of a patent’s standards essential status on apportionment and the relevance of prior licenses between the parties, the U.S. Court of Appeals for the Federal Circuit vacated and remanded a damages award, explaining that the smallest salable unit is not mandatory for every apportionment analysis, especially when there are relevant prior licenses between the parties. Commonwealth Scientific and Industrial Research Organisation v. Cisco Systems, Inc., Case No. 15-1066 (Fed. Cir., Dec. 3, 2015) (Prost, C.J.).

In the early 1990s Commonwealth Scientific and Industrial Research Organisation (CSIRO) was granted a patent that covered the IEEE 802.11a wireless standard. CSIRO committed to license the patent on reasonable and non-discriminatory (RAND) terms for the 802.11a standard, but refused to commit to RAND terms for later variations of the standards. Between 1996 and 2005 CSIRO agreed to various licenses with Cisco at royalty rates based on total sales volume and per unit prices. In July 2011 CSIRO sued Cisco asserting the patent. After Cisco stipulated not to contest infringement or validity, the district court awarded CSIRO $16.2 million in damages. Cisco appealed.

The Federal Circuit started by re-affirming that damages must “reflect the value attributable to the infringing features of the product, and no more.” The Court emphasized that to be reliable, apportionment models must be “sufficiently tied to the fact of the case,” but rejected Cisco’s argument that all apportionment models must begin with the smallest salable patent-practicing unit.

In terms of the district court reliance on the parties past licenses, the Federal Circuit found that because the parties’ prior negotiations centered on the asserted patent, apportionment was already built into the analysis despite the licenses being based on total sales volume and per unit prices. The Federal Circuit explained that damages models based on comparable licenses remained reliable, even where the licenses are not based on the smallest salable patent-practicing unit. Thus, the Court found error in the district court’s conclusion that the parties’ prior licenses for the same patent were not relevant to the damages analysis. Rather, the Court explained that as an actual agreement between the parties that was contemporaneous with the hypothetical negotiation, entered before the newest 802.11 standard was established and focused on the chips at issue, was pertinent and should be taken into account.

The Federal Circuit next turned to the “unique considerations” applicable to apportionment in the context of a standards essential patent. The Court restated the Ericsson test, requiring that the patented feature be apportioned from all the unpatented features reflected in the standard, as well as that the royalty is based only on the patented feature, not any value added by the standard’s adoption of the patented feature. The Court rejected CSIRO’s argument that Ericsson only applied to patents encumbered with RAND obligations, explaining that Ericsson explicitly applies to both encumbered and unencumbered standards essential patents. Thus, the Court concluded that the district court erred in failing to account for the value of the asserted patent accruing from the adoption of the new 802.11 standard, requiring that the damages award be vacated and the case remanded for a new determination of a reasonable royalty.

Practice Note: Damages analysis is usually highly fact-specific. The overarching principle here is that the Federal Circuit will allow courts to retain the necessary flexibility to ensure that patentees are adequately, but not overly, compensated for innovation.