Patent litigation suffers from a number of issues at present.
- Hedge funds backing non-practicing entities (NPEs) in order to chase a share of billion-dollar judgments.
- Plaintiffs using damages methodologies that have little to no relation to the reality of the patent system in order to obtain those billion dollar judgments.
- NPEs asserting patents that they don’t use, and that they obtained from operating companies who never used the patents either.
- Plaintiffs picking the judge they want in order to avoid having the Patent Office review the validity of their patents.
One recent case ticks all these boxes.
VLSI v. Intel
The scene is the Western District of Texas, Waco Division. The plaintiff is VLSI Technology, a non-practicing entity. The defendant is Intel Corporation, a major U.S. semiconductor and electronics manufacturer and one of the leading semiconductor research companies in the world.
VLSI Technology doesn’t make anything—they never did. They didn’t even have a corporate witness testify at trial. They’re actually a subsidiary of Fortress Investment Group. (Yes, the same Fortress whose subsidiary tried to block the sale of certain COVID-19 tests.) Fortress, a private equity firm owned by Japanese holding company SoftBank, has made something of a business of patent trolling—Fortress-backed entities, especially Uniloc, are some of the most litigious NPEs out there.
There has been a resurgence in this kind of NPE over the past few years. While the AIA and the Alice decision initially curtailed NPE litigation, the prevalence and cost of NPE litigation has increased steadily over the past few years. Much of this increase appears due to former USPTO Director Andrei Iancu’s NHK Spring and Fintiv precedential decisions, which decreased access to the IPR system—especially for defendants sued in the Western District of Texas’s Waco Division.
By filing in the Waco Division, plaintiffs can guarantee that their case will be heard by Judge Albright. And Judge Albright sets a fast trial schedule and has publicly stated that he will not stay his cases pending IPR. As a result, Judge Albright is now overseeing 20 percent of all patent cases filed in the United States.
VLSI (and its lawyers, from Irell & Manella) understood that being in front of Judge Albright would provide a benefit—a significant enough benefit that it dropped a case in front of a judge in Delaware to refile in Texas. And indeed, going to Texas paid off for VLSI here. Judge Albright, Intel filed IPRs against VLSI’s patents. Those IPRs were denied not because they lacked merit but because of Albright’s trial schedule.
The Fintiv rule is a complete betrayal of the intention of IPR—which is probably why it’s currently being challenged in court, and why Director Iancu’s attempt to enshrine it into regulation was shot down by the Trump administration before a rule ever published. And the Intel case shows just how costly those denials can be.
Of course, Irell is also the firm that Director Iancu was at before he went to the Patent Office, a nomination that closely coincided in timing with the opening salvo in VLSI’s campaign against Intel. Irell stands to make tens or even hundreds of millions of dollars from this litigation alone—money that may well not have gone to them if not for the Fintiv rule.
$2.1 Billion On A Theory Lacking Validity
During a trial, each side presents their view of what would be an appropriate level of damages. But experts aren’t supposed to just make up those numbers. They’re required to base them on economic methodologies that are sufficiently robust to meet the Daubert standard—they have to be relevant scientific knowledge that has “a reliable basis in the knowledge and experience of [the expert’s] discipline.”
But the VLSI economic expert relied on a theory known as hedonic regression. Hedonic regression is an economic model more typically used in real estate valuation. And in one previous case, a Federal Circuit judge described that model’s application to patent valuation as “mere guesswork.”
Hedonic regression attempts to estimate the value of a feature—in this case, the allegedly infringed patented technology—as a portion of a total price for a good. This works reasonably well in real estate, but the model has a number of limitations that come up in the context of patents. It relies on purchasers having the option to choose whether a given feature is included or not and on purchasers themselves having knowledge of the value of the feature to them. It makes other assumptions about the market, including on pricing lag and on the ability to separate a feature from other features that are always present with that feature. And where those assumptions aren’t valid, the estimate is going to be flawed.
These sorts of enormous damages verdicts rarely hold up on appeal, often because the damages models used by plaintiffs are flawed. But VLSI can grab headlines while the appeal is pending and try to use them as leverage in extorting licenses from other companies who might not have the resources to fight back.
And when courts fail to exercise their Daubert gatekeeping role and permit juries to see these sorts of flawed damages requests, there’s real harm. Jurors, who are not themselves economic experts, are faced with a choice between two numbers from two competing experts. The experts will both claim the other expert is wrong, and the jurors are unlikely to have the economics background required to evaluate the quality of the economic models being put forward.
From a plaintiff’s perspective, there’s little harm to presenting a ridiculously high number—maybe the jury goes for your number, as they did here. Maybe they treat your number as an “anchor” and split the difference between the plaintiff and defense numbers, in which case plaintiffs still benefit from presenting a high number. And even if the judge throws out your damages report—which has happened to hedonic regression experts before—you can still try to get an alternative damages theory included. There’s little downside to plaintiffs and a whole lot of benefit.
One Case, All The Problems
VLSI v. Intel is just one case. While the amount of damages might be atypical, the rest of the case isn’t. In one recent analysis, 39 out of 41 denials based on Fintiv were on patents related to litigation in the Western or Eastern District of Texas. IPRs relating to cases filed in those two districts were more likely to be denied than instituted. In contrast, 95% of IPRs relating to cases filed anywhere else in the country were instituted.
There’s a reason that more than 85% of cases filed in front of Judge Albright are filed by NPEs—they know that they can escape IPR, the biggest threat to their litigation campaigns, as long as they’re in Waco. Operating companies, whose largest concern is counter-suit, don’t see the same benefit and are less likely to take advantage of this kind of forum selling.
And even the damages in this case are becoming less atypical. Other recent mega-verdicts have exceeded $1 billion, and damages overall appear to be increasing.
And in April, Intel and VLSI will do it all over again—this is only the first in a series of Fortress-backed lawsuits against Intel. The current state of patent litigation doesn’t benefit American companies. In fact, as we’ve seen from this case, it tends to harm them. The only parties that benefit seem to be patent lawyers and hedge fund-backed NPEs.