Each week, we succinctly summarize the preceding week of Federal Circuit precedential patent opinions. We provide the pertinent facts, issues, and holdings. Our Review allows you to keep abreast of the Federal Circuit’s activities – important for everyone concerned with intellectual property. We welcome any feedback you may provide.
– Joe Robinson, Bob Schaffer, Lindsay Henner, Parker Hancock and Tinh Nguyen
Federal Circuit Review – Issue No. 63-02.
Deliberate Copying Bars Laches Defense; Invalidity Challenge Formulated for Litigation that is Not Objectively Unreasonable Can Prevent Enhanced Damages for Willful Infringement.
Carnegie Mellon Uni. v. Marvell Tech. Group, Ltd., No. 2014-1492, 2015 U.S. App. LEXIS 13622 (Fed. Cir. Aug. 4, 2015) (Before Wallach, Taranto, and Chen, J.) (Opinion for the court, Taranto, J.). Click Here for a copy of the opinion.
Carnegie Mellon University (CMU) sued Marvell Technology Group for infringing two patents related to a detector for reading data stored on hard-disk drives. A jury returned a verdict for CMU, and awarded $1.17 billion as a reasonable royalty for infringement, based on a royalty of 50 cents per chip. The district court increased the damages by 23% for willful infringement and entered a verdict of $1.54 billion in damages, plus a continuing royalty of 50 cents per chip. The district court further denied Marvell’s defense that CMU’s delay in bringing suit should bar pre-suit damages under laches.
The patents arise from the result of research by Drs. Kavcic and Moura at CMU, who developed a detector for reading hard-disk drives that improved a prior method invented by Viterbi. In 1997, they filed a provisional patent application, and in May of 1998 they published a corresponding paper. The two patents in suit issued in 2002. In 2001, Marvell became aware of this work and began developing the technology for its detector chips. Part of this process included a simulated detector chip, which they named the “Kavcic Viterbi” simulator. Marvell used the simulator to design a “sub-optimal” version of the Kavcic Viterbi technology, which they called “KavcicPP.” Marvell’s next generation of chips embodied exactly the technology Kavcic proposed in his 1998 paper.
On appeal, Marvell argued that it developed a “suboptimal” solution, different from CMU’s primary embodiment in the patents, and similar to the Worstell prior art. Thus, if the claims are construed narrowly, they would be patentable over Worstell, but would not be infringed by Marvell’s chips. If the claims are construed broadly, Marvell’s chips may infringe, but the claims would be anticipated by Worstell. Specifically, Marvell argued that Worstell disclosed all the limitations of the asserted claims, including a “signal-dependent branch metric function.” However, the Federal Circuit decided that a jury could reasonably have found the claims both valid and infringed, despite Worstell, because the claims required a set of signal-dependent branch metric functions, and was not anticipated by a method which only had one such function. Marvell’s chips, like the claims, had a set of signal-dependent branch metric functions.
Marvel also argued that no reasonable jury could have found infringement of claims that required “Viterbi-type” calculations. The Viterbi method requires a “trellis” of related calculations. Marvell argued that its chips performed some, but not all of the calculations in the trellis, and did not infringe. The district court, however, correctly construed “Viterbi-like” to include embodiments that performed some, but not all, of the calculations in the trellis.
Marvel further argued that its detector simulator did not infringe, because it simulated a signal detector and was not actually the detector itself. However, the claims recite a “detector” used for processing samples received from a read head, and thus do not require an actual read head to infringe. Accordingly, the jury could find that the simulator was a “detector” as recited by the claim.
Marvel argued that the district court abused its discretion by not applying laches to bar pre-suit damages. The district court found that CMU’s delay in bringing suit was unreasonable and inexcusable, and that Marvell suffered some evidentiary prejudice, not economic harm. The district court weighed the overall equities, and held that laches should not apply, because Marvell copied CMU’s patents consciously and deliberately for a decade. Marvell argued that Serdarevic v. Advanced Medical Optics required that a plaintiff could not rely on an unclean hands defense to laches unless the defendant’s misconduct was responsible for the plaintiff’s delay in bringing suit. However, Serdarevic involved a dispute over inventorship, and did not apply to infringement suits. Instead, deliberate copying by Marvell was sufficient to defeat a laches defense. Thus, the Federal Circuit determined that the district court did not abuse its discretion by denying Marvell’s laches defense.
Marvel challenged the enhancement of damages for willfulness. While Marvell’s chips blatantly copied the CMU patents, Marvell’s invalidity defenses raised in litigation were not objectively unreasonable. Specifically, while the jury had sufficient evidence from the Worstell reference to find no invalidity of the infringed patents, the disclosure in Worstell was sufficiently unclear that Marvell’s defenses were not objectively unreasonable. Further, the district court was incorrect to hold that the defenses were unreasonable solely because they were developed for litigation, based on a reference unknown to Marvell before litigation commenced. Thus, the award of enhanced damages for willfulness was reversed.
Marvell finally challenged the damages award, alleging that the damages expert should have been disqualified, and that the 50 cent royalty per chip was unreasonable. However, there was no reversible error in either determination. The damages expert, while not holding a Ph.D, had sufficient experience and training to provide expert opinions under the Daubert standard. Further, even though the royalty rate was high, use of the technology embodied in the CMU patents was a “make or break” technology for Marvell, and thus CMU would have had sufficient bargaining power to obtain such a high royalty rate. Thus, the 50 cent per chip royalty was not unreasonable.
However, no reasonable jury could have concluded that chips manufactured and delivered abroad were infringing under the evidence presented. While chips eventually imported into the U.S. are infringing under § 271(f), that subsection does not apply to those delivered abroad. Instead, CMU would need to prove infringement under § 271(a) – that the products were “sold” in the U.S. Here, even though the manufacture and delivery were abroad, actions taken in the U.S. may have constituted a “sale,” such as the negotiation and agreement to a contract in the U.S. However, the factual record has not been sufficiently developed to entitle Marvell to judgment as a matter of law that the sales did not occur in the U.S., and thus a new trial would be necessary on that issue.