Intel’s Fed Circ triumph over VLSI: an object lesson in the doctrine of equivalents and insufficient evidence

One of the biggest patent infringement verdicts in the US has been thrown out by the Court of Appeals. WIPR gauges litigators’ reaction to the reversal, including takeaways for defending doctrine of equivalents claims.

The Federal Circuit for the Court of Appeals handed a victory to Intel yesterday, December 4, by overturning the much-quoted £2.18 billion award previously handed to VLSI ( see story).

VLSI had sued Intel in various US courts alleging infringement of several patents covering semiconductor technology, resulting in the headline verdict in 2021 when Intel lost to VLSI in a Texas federal court.

The latest judgment follows a decision by the US Patent Trial and Appeal Board ( PTAB) in June, invalidating the second of two VLSI-owned patents.

WIPR gathered reaction from US patent litigators for their take on yesterday’s bombshell ruling.

Timothy Getzoff, partner, Holland & Hart: “This decision…provides strong support for accused infringers who are defending against a doctrine of equivalents claim.”

“The Federal Circuit decision is notable for a few reasons.

“First, and perhaps most significant, the court ruled as a matter of law that the doctrine of equivalents was not satisfied for the infringement claim as to the ’759 patent. Generally, doctrine of equivalents is a factual issue to be decided by the jury, and in this case the jury found in favour of VLSI that the ’759 patent was infringed under the doctrine of equivalents.

“The Federal Circuit not only ruled that the jury was wrong, but decided the issue in favour of Intel, essentially holding that no reasonable jury could have found in VLSI’s favour. In doing so, the court scrutinised the evidence submitted at trial, especially the expert testimony, and found it to be conclusory and superficial (those are my words—the court used the word ‘insufficient’).

“This decision, while not marking a change in the law, provides strong support for accused infringers who are defending against a doctrine of equivalents claim, as it held the plaintiff to a strict standard for meeting its burden.

“Second, the court also reversed the damages award, although the basis for the reversal was very fact-specific and not likely to provide much guidance for other cases. Many commentators (me included) expected the court to reverse the damages award, as such incredibly high damages awards tend to get more scrutiny on appeal.

“Here, the Federal Circuit reversed by finding a logical error in VSLI’s expert analysis, whereby the expert improperly included infringing and noninfringing capabilities of the Intel chip in coming to his reasonable royalty analysis.

“Because this decision was very specific to the expert’s factual analysis, I do not see this part of the decision having far-ranging implications, except perhaps as another example of how high monetary verdicts are harder to uphold on appeal.

“Although the court reversed the expert’s conclusion because of this error, the court declined to say that the expert’s overall approach and theory were flawed. As a result, on remand for a new trial on damages, it is possible that VSLI’s expert can fix the error identified by the court, but otherwise keep to his overall approach and theory.”

Michael Hawes, firmwide chair, Baker Botts’ Federal Circuit Appellate Practice Group: “The reversal of doctrine of equivalents should be reviewed by any plaintiff…to make sure they offer sufficient evidence.”

“While the Federal Circuit vacated the headline big-number verdict due to its damages and licence rulings, the rulings on infringement may be more important for the practitioner.

“The court upheld Intel’s infringement of the higher-dollar ’373 patent in significant part because the key claim term was not construed, which left Intel facing the ‘only reasonable view of the claim element’ standard.

“On the other hand, the reversal of doctrine of equivalents should be reviewed by any plaintiff relying on that theory to make sure they offer sufficient evidence.”

Ian DiBernardo, partner, Brown Rudnick’s IP Litigation Practice Group: “The issue is [not so much] a lesson for litigants [as] a lesson for due diligence in transactions.”

With respect to affirmance of infringement of the ’373 patent (regarding minimum operating voltage), the court’s decision emphasises the legal standard on appeal and provides a good reminder to litigants: although Intel had competing evidence of non-infringement (here, a comparison done by its expert), VLSI had competing evidence; and ‘[t]he jury reasonably could credit this evidence and reject Intel’s contention….’ (page 12).

“Regarding (‘when’ limitations), there are two noteworthy points on this argument.

“First, the court called out Intel for a tactic commonly used by defendants—namely keeping non-infringement arguments in their back pockets by not raising them during claim construction.

“Here, the court reiterated the law that when a claim term is not construed, the court will defer to the jury’s view unless contrary to the only reasonable view (pages 12-13). There is risk to defendants not asking for a particular construction.

“Second, the court declined to read a causal relationship into the word ‘when’, confirming it to mean simply ‘at that time’ (page 13).

“With respect to reversal of infringement of the ’759 patent, the court’s application of the doctrine of equivalents—and, specifically, the need to present particularised testimony and linking argument of the claimed limitation and alleged equivalent—is not surprising.

“However, by quoting the relevant testimony, the court does seem to want to provide a reminder to claimants and trial courts that expert testimony must answer the question of ‘why’ the two are equivalent (pages 18-19).

“Regarding remand of damages, although the court vacated the damages award, it did acknowledge that patent damages do not require precision, noting that some steps in a sound analysis may involve ‘unavoidable approximation and uncertainty’ (page 21).

“The court’s criticism of VLSI’s expert acknowledged this, noting [that] Dr Annavaram’s methodology was not merely a matter of choosing one reasonable step over another or of estimation in the face of acceptable imprecision of uncertainty.” (page 25).

“Specifically, the court found used data that was not from use of infringing functionality, and data from infringing functionality was available.

“Importantly, notwithstanding VLSI’s damages award being vacated, the court emphasised that VLSI’s error is not fatal to its case. Instead, the court explicitly noted that VLSI should be given ‘an opportunity to provide a corrected damages case’ (page 27).

“Perhaps building on the court’s acknowledgement of ‘unavoidable approximation and uncertainty’ in certain cases, this may be a signal to district courts of the [Court of Appeals’] view to more leniently allow claimants to correct errors in damages methodologies.

“Regarding licence defence, the issue is [not so much] a lesson for litigants [as] a lesson for due diligence in transactions. While hindsight is often 20-20, to those of us that both litigate and handle transactional/licensing matters, the potential implications of ‘affiliate’ language on the definitions of parties and the scope of licences is one of many potential issues to consider.”

Scott Bittman, partner, Crowell & Moring: “This case illustrates the importance of aligning a patent holder’s technical expert with its damages expert.”

“VLSI’s damages model was based on the supposed technical benefits—increased processor speed and power savings—attributable to Intel’s infringement. However, the court determined that VLSI’s technical expert’s calculations were flawed because they compared the performance of the infringing functionality to non-impacted functionality.

“The court noted that these calculations showed a significantly higher power savings over what would have been shown if the correct comparison was performed. As a result, the damages model was deemed unreliable. VLSI will have the opportunity to provide for a corrected damages case in a new trial on the issue of damages.

“This case illustrates the importance of aligning a patent holder’s technical expert with its damages expert to ensure that the technical foundation is solid and can sufficiently support the damages theory being presented.

“VLSI’s damages theory was found reasonable, but the underlying calculations provided by the technical expert were flawed for not being sufficiently tied to the patented features.

“VLSI may still obtain a substantial damages award on remand. Notably, however, VLSI’s patent has been found invalid by the USPTO in an inter partes review final written decision, which can eliminate VLSI’s ability to collect any damages if affirmed.”

This article was updated later on the day of publication to add Scott Bittman's comments.

Did you enjoy reading this story?  Sign up to our free daily newsletters and get stories sent like this straight to your inbox