I have been following and short ParkerVision (PRKR) since at least 2003. I am not a lawyer, but have been involved in some extensive patent litigation over my own patents.
I have previously made my opinions and views on ParkerVision public and, so far, every substantial prediction I have made about ParkerVision has eventually been proven to be true, including statements made in a 2007 Barron's article about ParkerVision. I have good reasons to believe that the allegations made in the MAXTAK v. ParkerVision lawsuit, namely that ParkerVision consistently and deliberately misrepresents its revenue prospects, commercial relationships, and technological prowess, are largely accurate. I have had discussions about ParkerVision with large numbers of people (both engineers and senior executives) in the RF chip industry, and have never found anyone who had positive things to say about the company or its technology. This includes many who have looked at ParkerVision under NDA and who came away feeling strongly that ParkerVision had nothing of value to offer.
In this report, I will briefly discuss ParkerVision's history and analyze the remaining issues in the Qualcomm (QCOM) lawsuit in some detail. Finally, I will point out some significant errors in the work of the only analyst covering ParkerVision - errors which result in numbers that are high by a factor of twenty.
Since going public in 1993, ParkerVision has managed to raise and spend well over $250 million without ever having a profitable year. Over the last seven years, PRKR has only received about $100,000 of cumulative revenue. The majority of the $250 million raised has been absorbed by salaries and bonuses, outside consulting, and patent filings. The latest 10K reveals that Robert Sterne, a board member, has been paid more than $18 million in legal fees from the company, while the top three executives (including the founder, chairman and CEO Jeff Parker) have pulled out more than $25 million in cash compensation (along with a significant amount of stock options).
The complete absence of product sales over the last seven years or of any profits over its entire history shows that ParkerVision has primarily been in the business of selling stock to new investors. Because each investor and analyst eventually discovers that Jeff Parker's promised sales and profits always fail to occur, the company needs to change its story (and its shareholders and analysts) every 3-5 years. The July 2011 filing of a lawsuit against Qualcomm is simply the most recent ParkerVision story used to raise more cash and avoid delisting. Like the previous narratives, it can be quite convincing to those that lack the historical context. Nonetheless, this iteration has allowed PRKR to boost its stock price and fund its cash flow burn of approximately $4-5 million/quarter (2012 10K). For the many reasons cited below, I believe the claims in the Qualcomm lawsuit will eventually be revealed as grossly exaggerated, leaving the current crop of shareholders with same negative outcome as was suffered by previous shareholders.
ParkerVision Chip Sales Potential
Since forming a relationship with Via Telecom in late 2007, Jeff Parker has repeatedly announced that a handset deal for D2P (Direct2Power, i.e. ParkerVision's name for its power amplifier) was imminent, but to date nothing has resulted from this partnership. During ParkerVision's fourth quarter 2012 conference call, Jeff Parker admitted that Via Telecom had cancelled their contract/license with ParkerVision, and that ParkerVision would pay Via Telecom $1.3 million this year to do enabling engineering for a handset with ParkerVision technology. Technology, that according to Jeff Parker, was supposedly finished and working years ago. In reality, this is nothing more than a payment to maintain a relationship with Via Telecom.
I do not believe there is even a small possibility of D2P producing revenues in 2013 or 2014, just as there were no revenues in past years. ParkerVision currently has no customers or even a shippable D2P or D2D (Direct2Data) chip. I am confident that the D2P technology (at least as described in the ParkerVision patents) does not and will not result in high-efficiency power amplifiers, and thus the projections in the Apr 3, 2013 Ladenburg Thalmann report are as believable as the previous ParkerVision analyst reports from 2007-2009. There are a number of other serious errors in the Ladenburg Thalmann report, as I discuss later.
In 2011, with the stock well below $1, a few months of cash on hand, and the ability to sell a revenue story to investors floundering, ParkerVision decided to make one more play to keep the company alive. On July 20, 2011 ParkerVision filed a lawsuit against Qualcomm alleging patent infringement. In this lawsuit, ParkerVision claims ownership of wide class of "passive mixers" used in radio receivers. While many outstanding issues remain in the case, I will focus on validity, laches, indirect infringement, injunctions and willful infringement. For those interested in reading the legal filings, an archive of the PACER filings may be found at ParkerVision v. Qualcomm. Before we get into the legal portion of this case, let's first look at what ParkerVision has claimed to invent.
Active and Passive mixers - What does PRKR claim to have invented?
"Mixers" are the part of a radio receiver that down-converts the radio frequency signal to a lower-frequency signal for additional processing. All receivers have at least one mixer. Some mixers convert the RF directly to baseband (direct conversion), while others convert first to an intermediate frequency (IF), and eventually to baseband in a multistage process.
Over the years, in press announcements, conference calls and white papers, ParkerVision has claimed that D2D didn't use or need a mixer to directly convert to baseband. In their lawsuit against Qualcomm, ParkerVision is now claiming to have invented a broad subset of a type of mixer the industry calls a "passive mixer" (as compared to the primary alternative, an "active mixer"). However, passive mixers have been around for at least 50 years. They are typically implemented with either diodes or semiconductor switches. ParkerVision claims-in-suit (of the six patents-in-suit) essentially claim any switch-based passive mixer that open a switch for "non-negligible" time periods and which store "non-negligible" transferred energy. This basically includes any non-diode-based passive mixer (which is also called a "commutating" or "switching" mixer in the technical literature.) Since these types of mixers have been the subject of papers and patents since the 1940's, there is a lot of (invalidating) prior art which I will discuss later in this article.
Validity, Infringement & the Markman Ruling
The recent increase in ParkerVision's stock price is primarily the result of receiving a favorable Markman ruling. A Markman ruling defines the meaning of the terms used in claim-in-suit, which is needed to determine both infringement as well as validity. ParkerVision needed to win most of the Markman terms in order to avoid summary judgment against them of non-infringement. While the very broad Markman ruling helps with infringement, it hurts validity, because a broad Markman greatly expands the prior art available to Qualcomm in their defense. The tradeoff was a necessary one for ParkerVision, in that the burden of proof is higher (clear and convincing evidence) for invalidity than for infringement.
A primary issue now between the parties is validity of the claims-in-suit. That is, is there invalidating prior art? A patent claim may be invalid for several reason, the most basic of which is called "anticipation", which means that every element of a patent claim is contained in a single prior art reference. A prior art reference is a paper or patent that was available publicly before the invention date of the patent-in-suit. I was able to find over 40 very good invalidating prior art references with a few days of searching. While I can't put copies of these papers online due to copyright laws, most are available from Google. If I can find 40 clear invalidating references with just a few days of searching, I am confident that Qualcomm has found many more. None of these references were submitted to the patent office during the prosecution of the ParkerVision patents-in-suit. One other reference (Tayloe98) was submitted for the three of the patents-in-suit, but not for the other three. It turns out that Qualcomm has implemented Tayloe98 pretty much exactly in at least some of the accused chips. There is an earlier reference ("vanGrass90") which is actually broader than Tayloe98 and was not disclosed to the patent office in any application-in-suit.
By request, I can make available a document which describes how a single reference (Shahani97) invalidates a large number of claims-in-suit. I will leave the full explanation of how every claim-in-suit is invalidated by prior art to Qualcomm's expert witness. The expert witness reports were due in March. Along with the judge, we don't see any of these reports until additional motions about the reports are submitted to the court. Many commentators, including Ladenburg Thalmann, have stated that Qualcomm must not have any prior art since they haven't filed it with the judge. This shows a lack of understanding of how the legal process works. Qualcomm's prior art references will be the subject of many filings sometime this summer.
Laches applies when a patent holder knew-or-should-have-known about infringement, but did nothing. In this case, laches will be a major issue. In 2002, Qualcomm made an announcement about a direct conversion receiver (ZeroIF2002), about which Jeff Parker made a comment in a public conference call to the effect that Qualcomm would have to pay royalties. This clearly exhibits that ParkerVision was aware of the potential infringement quite early. This awareness made it ParkerVision's legal obligation to investigate such potential infringement and notify Qualcomm. However, the first notification by ParkerVision of potential infringement was the filing of this lawsuit in July 2011. The normal result of a finding of laches is that any pre-filing infringement is not subject to damages - thus damages would only be awarded for infringement after the date of the lawsuit. There is a presumption of laches after six years, so that it would be ParkerVision's obligation to disprove laches. I believe that laches will apply, but we will see briefings about this issue before the trial date, and my opinion might change with additional facts.
Estimated probability of at least one ParkerVision claim found valid and infringed
Given the sheer amount of invalidating prior art, I think the odds of Qualcomm succeeding in invalidating all of the claims-in-suit is quite high. Given a reasonably technical jury, I would estimate this at 95+%, but it is exceedingly unlikely that a Jacksonville jury will have even a single technically trained person. Because of these jury issues, I put the odds of Qualcomm's success at invalidating all claims above 70%.
I spoke with several very experienced patent litigators and they told me that the typical number of anticipating prior art references brought to trial is between zero and one (i.e. usually none, but occasionally a single reference). It is very rare to have this many distinct and straightforward prior art references, none of which were shown to the patent office. Unfortunately, Qualcomm will only be able to show a small fraction of the invalidating references to the jury due to time constraints. Therefore, I believe the odds of a jury finding at least one valid and infringed claim is 30% or less. If all claims are found invalid, then by collateral estoppel, ParkerVision would not be able to sue anyone else on those claims. If any claims are found valid, but not infringed, there remains the possibility of lawsuits against other companies, but no damages from Qualcomm.
Estimated Damages if any claim(s) found valid and infringed
If the jury finds any claims valid and infringed, then the jury will look at damages. A U.S. court can only assess damages for infringing products manufactured or shipped in the United States (the Ladenburg Thalmann analyst gets this wrong.) Two good overviews of this are Signore and Petersen.
Indirect infringement is another "big issue" left in the case. Since Qualcomm manufactures all ICs outside of the United States., and essentially all of those ICs are assembled into devices outside the United States, the number of Qualcomm transceiver chips directly sold or manufactured in the United States by Qualcomm are, to first order, zero. Indirect infringement, if proved, would allow ParkerVision to collect damages on Qualcomm transceiver chips that come into the United States as part of other companies' devices (Samsung, Apple, etc.). As shown in the reference above, ParkerVision has a high burden to prove indirect infringement as they must not only prove that Qualcomm knew they were infringing, but also that Qualcomm believed that they had no viable defenses. While ParkerVision has recently survived a motion to dismiss (where all facts are assumed to be in ParkerVision's favor), ParkerVision still has the burden of proving a very difficult case. If ParkerVision loses indirect infringement before or during the trial, damages (and royalties) from Qualcomm will be trivial even if the patents are found valid and infringed by the jury and ParkerVision would be forced to sue the device makers directly. Even though I believe it likely that Qualcomm will prevail on this issue, I will calculate damages under the assumption that ParkerVision wins on the issue of indirect infringement.
To assess potential damages I need to estimate Qualcomm's sales of transceiver chips within the United States. Since all of Qualcomm's chips are manufactured outside the United States, only U.S. sales are relevant. I will assume that ParkerVision prevails on indirect infringement so that any Qualcomm transceiver chip in any device shipped into the U.S. would be considered for damages. Approximately 300-450 million relevant "devices" are shipped per year in the United States (handsets, notebooks and pads). Qualcomm has a 50% market share of baseband chips, but a lower market share of transceivers (only transceivers contain the accused mixer, mixers are used in baseband chips). I estimate Qualcomm's transceiver market share to be about 30%. Thus the number of Qualcomm transceiver chips under consideration for damages would be in the range of $100-$150 million per year. Given that the average selling price of these transceivers is $2-$6 each - I will use an average of $4. These numbers result in total Qualcomm transceiver sales subject to damages of $400-$600 million per year.
Damages in patent cases can be assessed as lost sales or as a "reasonable royalty". Since ParkerVision has never sold any handset chips (either before or after the alleged infringement) and has no currently active licensees, lost sales are not relevant, so the reasonable royalty (RR) method must be used by the jury. The size of a RR may depend on various factors, but one key issue is whether there are equally-good, non-infringing alternatives. It is easily proven that active mixers (which ParkerVision admits are non-infringing) are just as good both in cost and performance. This can be shown in two ways. The first is that most other transceiver companies are using active mixers with the same cost and performance as Qualcomm. Secondly, I believe Qualcomm will be able to demonstrate a handset at trial with an active mixer replacing the allegedly infringing passive mixer.
For this reason, I believe that a jury is most likely to find a RR of 1% or less is appropriate, although higher numbers are certainly possible. I will use a RR of 5%, which I consider extremely high but remotely possible. For 1% RR, annual damages would then be $6 million, for a 5% RR annual damages would be $30 million. If laches applies then the damages awarded would be for 2.25 years of infringement (filing of lawsuit to end of trial). The math in this scenario is, 2.25*$30= $68 million; or about $14 million for a RR of 1%. Keep in mind, that when you are running your own numbers of a potential award, that ParkerVision's attorneys McKool Smith take about 30-40% of any award as its contingency fee.
If any claims are found valid and infringed and survive appeal, what is a plausible estimate of future Qualcomm royalties? Qualcomm can and will switch to equally good non-infringing alternatives (e.g., an active mixer). It would take a couple of years for Qualcomm to switch production over completely (i.e. any royalty would apply to a ramp from 100% of U.S. sales down to 0% over about 2 years, or about 1.5 years effective royalty). There may be other companies using passive mixers in handsets, but most appear, from published papers and patents, to use active (i.e., non-infringing) mixers in their receivers. I feel confident that no one will settle with ParkerVision, and every other company will be quite equally willing to fight the legal battle if necessary.
Therefore, I conclude that the maximum amount that Qualcomm could conceivably pay ParkerVision (past damages plus NPV of all future royalties) would be about $30M/yr * 3.75 yrs = $112M, less 30% for contingency, or about $80 million in total. This is assuming that the patents are found valid and infringed by a jury and that nearly all of the pending legal issues are decided in ParkerVision's favor.
The Ladenburg Thalmann report uses a damages award of $1 billion and a future annual royalty stream starting at $350 million in 2015 and rising to $395 million by 2020, resulting in a royalty NPV of $1.5 billion. I can't find any plausible justification for either number as they are both based on a number of serious errors. For instance, the author uses worldwide sales of Qualcomm's baseband chips to calculate royalties and damages. These are both completely wrong, as damages can only be awarded on U. S. sales, and the allegedly infringing chips are the transceiver chips not the baseband! While Qualcomm does ship some combined baseband/transceiver chips for the very low-end, low-cost market, the bulk of Qualcomm's transceiver sales are standalone - see any Samsung or Apple teardown report. Transceiver chips are far cheaper than baseband ($3-$6 ASP for a transceiver compared to $20-$25 for baseband), and Qualcomm's market share of transceivers is smaller than for baseband. The analyst ignores the near certainty that Qualcomm would switch to a non-infringing alternative, and he neglects the contingency fee. Correcting for these mistakes brings our estimates back in line, i.e. his damages of $1 billion are reduced to $48 million, and the NPV of future royalties is reduced from $1.5 billion to $32 million (both net of estimated 30% contingency fees.)
We do agree on the highest possible royalty of 5% and a roughly 33% probability of ParkerVision winning at trial.
The Qualcomm "Offer"
Much has been said by McKool and Jeff Parker about the Qualcomm "offer" of $600 million plus in 1999. What is most likely is that the "offer" was a non-binding Letter of Intent to enable further technology evaluation by Qualcomm. It certainly is very unclear how the "offer" is directly relevant to any possible damages calculation, since (based on the one document on file with PACER) it was for a royalty based on worldwide sales of a full ParkerVision transceiver design to be transferred to Qualcomm. This is not a direct comparison for damages based on United States sales of a small component (mixer) of a transceiver entirely designed by Qualcomm. At best, divide by 5 (U.S. fraction of world sales) and 2 (some fraction of entire transceiver) which brings the NPV of the "offer" down to about $60 million or so, which fairly close to the 1% number we calculated above projected over a 15 year period.
Willful Infringement, Injunctions, Appeals, and Settlement
The likelihood of willful infringement, which would result in enhanced damages, has been extensively promoted by Jeff Parker and others. It rarely occurs in recent patent cases because the standard to prove willful infringement is very high. As long as Qualcomm has a reasonable belief that the patents are invalid, then willful infringement cannot be found. Given the hundreds of prior art references that Qualcomm is producing, clearly they should have a reasonable belief in invalidity.
With no chip sales and with a history of trying to license the technology, ParkerVision is clearly a non-practicing-entity (NPE). With current case law, NPEs almost never get injunctions as they have no sales to protect and would not be able to step in to satisfy the market during an injunction. For an NPE, the only possible damages are a "reasonable royalty."
If the trial goes against Qualcomm, they have very good grounds for appeal. The Markman ruling itself will certainly be appealed - close to 40% are overturned, and there are clear mistakes made in this Markman ruling. The patent appeals court (U.S. Court of Appeals for the Federal Circuit) does every Markman de novo. No deference is made to the decision of the original judge.
With very little possibility of an injunction and a good likelihood of any adverse judgment against Qualcomm being reversed on appeal, Qualcomm has no reason to settle and every reason to fight what they obviously consider completely invalid patents that never should have issued.
It is hard to understand how ParkerVision's current market cap of over $300 million can be justified by any even slightly plausible outcome of the upcoming trial.
Additional disclosure: I hold both QCOM and RMBS. I was a co-founder of RMBS.