Marvell Lawsuit: Don't Worry, It Gets Worse

| About: Marvell Technology (MRVL)
This article is now exclusive for PRO subscribers.

An Astounding Lawsuit for Marvell

On December 27th, Marvell (NASDAQ:MRVL) announced a $1.2 billion loss to Carnegie Mellon University in a patent lawsuit. Since October, I have paid for access to the court docket and have read every document available in the docket (some of the documents are available here free of charge). The $1.2 billion award seems astounding, but the patents in question go to the core of Marvell's multi-billion dollar HD controller business. The CMU patents relate to error-correcting codes for magnetic media, and are presumably what allows Marvell to achieve the high gigabyte-per-platter density frequently touted during their conference calls.

In winning the award, CMU successfully argued that not only were the patents infringed, but that the patents were "must have" technology critical to Marvell's success. Particularly damaging was Marvell code and simulations introduced into evidence that were named after the CMU professors who authored the patents. Also damaging was the fact that CMU sent a letter to Marvell's CEO in 2003 offering to license the patents to Marvell. Marvell, however, did not respond to that letter, and it doesn't appear that Marvell ever bothered looking into whether they were infringing the offered CMU patents.

Impacts to Marvell's Bottom Line

CMU asked for, and won, a per-chip royalty on all Marvell HD controller chips sold. CMU's damages expert successfully argued a 50-cent per-chip royalty amount using the well established "Georgia Pacific" method, which uses a theoretical licensing negotiation between Marvell and CMU at the time the patents were first allegedly infringed. In contrast, Marvell's damages expert does not appear to have even offered a "Georgia Pacific" counterargument for a lower per-chip royalty; this is a fatal flaw in Marvell's case. In addition to the $1.2 billion jury award, Marvell will also be required to pay a 50-cent royalty on all infringing HD controller chips sold in the future. The CMU patents currently encompass all of Marvell's HD chips and will not be quickly or easily designed-around. By my calculations, the per-chip royalty will result in roughly a 20% hit to operating margins on their HD controller business (taking operating margins from 50% to 40%), and at least a 50% reduction to the HD controller business net earnings. Since the HD controller business makes up approximately 50% of Marvell's sales, the ongoing impact to earnings will be astounding, and could result in a 25% reduction in net earnings for Marvell as a whole going forward. Investors wishing to verify these numbers should start by taking a look at docket 708.

CMU originally asked for $750 million in back-royalties for the 1.5 billion HD controller chips sold from 2003 to 2010. CMU was denied royalties for all chips sold prior to 2009 due to a patent technicality (CMU was not requiring licensees to stamp the patent number on their devices). The revised amount, $1.2 billion, is for royalties on Marvell chips sold since 2009. The $1.2 billion award amount was a "finding of fact" from the jury and is not likely to be lowered by the judge, since judges only rule on "matters of law" in a jury trial. Because the patent infringement was apparently "willful", the judge could triple the jury's damages since treble damages are considered a "matter of law" and are under the judge's discretion. The $1.2 billion award amounts to approximately $2.25 per share.

Ongoing Royalty Applies to All Sales, Even Foreign

Also of critical importance is that the per-chip royalty applies to all Marvell HD chips sold, whether in the U.S. or abroad. CMU argued another patent technicality, that because Marvell violated the CMU patents during their "sales cycle" here in the U.S., Marvell should pay a per-chip royalty even on chips sold abroad. This is a key issue that Marvell will argue on appeal, and could lower the damages award from $1.2 billion to a more palpable $275 million.

Marvell Hiding Information from Shareholders

Back on October 19th I made a detailed comment on a SA article outlining why I thought Marvell was hiding information from its shareholders:

I used the departure of their CFO as an excuse to take a deep dive into their most recently filed 10Qs. One thing I noticed was that there were some important changes in their 10Qs related to their upcoming patent trial with CMU on November 27th. Search Marvell's two most recently filed 10Q's for "Carnegie Mellon Litigation", and also take a look at

In relation to the CMU litigation, you will notice that in their most recently filed 10Q Marvell removed the language stating that they "cannot determine the likelihood of loss nor estimate a range of possible loss", and buried way down in the footnotes on page 35, in a totally unrelated section, Marvell added the following language "CMU has alleged that it is entitled to damages in the amount of approximately $735 million through March 2010 as a result of the alleged infringement". Although I admire the wizardy of Marvell's management in being able to bury important information deep in the footnotes of their financial statement, I don't see this as particularly shareholder friendly.

Instead of being forthright and disclosing the amount at risk for the CMU lawsuit under the "litigation" section of their August quarterly report, Marvell chose to bury the amount far down in the footnotes together with other "risk factors" such as war and natural disasters. Marvell did eventually come clean by quietly updating the litigation section of their November quarterly report, including the amount at risk (which had ballooned to $1.2 billion). Marvell likely knew the "range of possible loss" for this lawsuit long before their August quarterly report.

Anomalies in Marvell's Accounting

The details of Marvell's share buyback program is another area where Marvell has become less transparent in their quarterly reports and merits further scrutiny. Marvell used to give a monthly breakdown of the shares repurchased in their quarterly reports, but for the past several quarters the monthly breakdown has been missing. On November 4th, I sent a question to Marvell's investor relations department asking about one of several anomalies I had found in Marvell's accounting:


Date: 11/4/2012 12:33AM

Subject: share count discrepancy in 2013 Q1 quarterly report

In the cover page of Marvell's May 10-Q filing it states that the "number of common shares of the registrant outstanding as of May 25, 2012 was 563.4 million shares".

In the 10-Q page 3 it says that there were 580.0 million basic shares as of quarter end on Apr 28, 2012.

On page 19, it states that "Subsequent to the end of the three months ended April 28, 2012 and through May 25, 2012, the Company repurchased an additional 10.0 million common shares"

The numbers don't add up. If at the end if Q1 there were 580.0 million shares, and Marvell repurchased 10.0 million from April 28th through May 25th, there should have been 570.0 million shares outstanding on May 25th. But the filing states that there were only 563.4 million shares outstanding. Where did the other 6.6 million shares go?


Marvell never responded to that e-mail. Although 6.6 million shares may seem an insignificant amount, at the May price of $12 per share it amounts to an $80 million discrepancy in their quarterly report, which is an amount almost equal to the $90 million in net earnings reported for the quarter. These are basic undiluted share counts and cannot be easily explained away by expiring options. Even if these disappearing shares could somehow be attributed to expiring options, there were only 14 million options outstanding at the time of the report, so 6.6 million options expiring would represent a nearly 50% reduction in the options outstanding. The details of Marvell's share buyback and dilution is a subject that deserves an article of its own.

Shareholders Beware

The impacts of the CMU lawsuit alone should give investors serious concern. Given Marvell's reluctance to disclose important information to shareholders, it should now be apparent that Marvell is not run for the benefit of the shareholders, and shareholders should beware.

Disclosure: I am short MRVL. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.