CARNEGIE MELLON UNIVERSITY, Plaintiff, v. MARVELL TECHNOLOGY GROUP, LTD. et al., Defendants, 986 F. Supp. 2d 574

OPINION NORA BARRY FISCHER, District Judge. I. INTRODUCTION This is a patent infringement case brought by Plaintiff, Carnegie Mellon University (“CMU”), against Defendants Marvell Technology Group, Ltd. and Marvell Semiconductor, Inc. (collectively “Marvell”), alleging that Marvell has infringed two of its patents, U.S. Patent Nos. 6,201,-839 (the “'839 Patent”) and 6,438,180 (the “'180 Patent”) (collectively, the “CMU Patents”). CMU contends that Marvell’s infringement was willful. (Docket No. 461). Marvell counters that the CMU Patents are invalid. (Docket No. 465).

This matter was tried before a jury for four weeks, with jury selection starting on November 26, 2012. (Docket No. 760). A number of motions for Judgment as a Matter of Law (“JMOL”) were made before the verdict was rendered. (Docket Nos. 699; 701; 703; 731; 738; 740; 742; 747). The Court denied these motions on the record on December 21, 2012. (Docket No. 759). The case was then presented to the jury. After deliberations, the jury entered a verdict on December 26, 2012 in favor of CMU on infringement, validity, and willfulness, awarding damages in the amount of $1,169,140,271.00 (Docket No. 762).

Following the trial, the Court entertained post-trial motions, wherein the parties: (1) renewed their earlier JMOL contentions; (2) moved for a new trial on several grounds; (3) argued the equitable defense of laches; and (4) requested a permanent injunction, post-judgment royalties, supplemental damages, interest, enhanced damages, as well as attorney fees. (Docket Nos. 786-811). These matters have been completely briefed (Docket Nos. 823-829; 832-837; 849-855; 857-863), and the Court heard argument on same from May 1 through May 2, 2013. (Docket No. 873).

The Court writes now to explain its reasoning for denying the pre-verdict motions for JMOL, and to rule on the renewed JMOLs, the Motions for New Trial, and Motion for a Remittitur. II. FACTUAL BACKGROUND A. Technology in Suit The patents-in-suit are generally directed to the method of sequence detection in high density magnetic recording sequence detectors. See '839 Patent col. 16 11. 20-23. 1. Hard Disk Drive Data Recordings Hard disk drives (“HDD”) contain a platter or disk that holds data on concentric tracks. (Docket No. 673 at 154). The device bears a visual resemblance to the classic record player. (Id.).

Just as a record player has a needle attached to the tip of the arm, an HDD has a “read head” that reads and writes data onto these tracks. (Id.). Each track is made up of a track width, and this track width is broken into millions of bit regions. (Id.). The track is made of magnetic material. (Id.). The bit regions are magnetized to store data in the form of “zeros” and “ones.” (Id,.). As the track moves underneath the read head, the read head picks up the fields emanated from these magnetic regions on the track and turns the fields into read back signal samples. (Id. at 155).

However, the read back signal samples are not exactly equal to what is actually written on the disk. (Id. at 156). For instance, the read back signal may read “0.3” when a “zero” was written on the track. (Id.). These discrepancies occurring during the read back process are referred to as “noise.” (Id.). 2. Viterbi-Like Detector and the Trellis Concept A Viterbi-like read channel detector found in the HDD takes the read back signal samples and determines the sequence of symbols written on the disk using a trellis. (Id. at 157-158). This process is called “sequence detection.” (Id. at 158). A trellis

section is used to represent a string of bits sitting on a medium. (Id.). There are four potential sequences of two bits, called states: 01,11, 00, 10; and they can be connected by branches. (Id. at 162-163). A trellis is used to represent a string of these bits; for example, a three-bit string of 011, would be represented by a “01” connected by a branch to “11.” (Id.). One trellis

section includes all possible bit sequences. (Id.). In this instance, a single trellis

section of Oil, 010, 111, 110, 001, 000, 101, 100 is represented as follows: (Docket No. 771 at Ex. C at 14). A trellis can then be created to represent a sequence of any length. For example, a six bit sequence is represented as follows: (Id. at 18). Through this trellis, one can trace a path that is equivalent to a specific sequence of symbols. (Id.). For example 100101, is shown below: (Id. at 17). The detector determines the “best path” through the trellis, meaning the best or most likely written sequence on the disk, using branch metric values. (Docket No. 673 at 169).

The read back signal samples are taken by the detector to compute the branch metric. (Id. at 170). The path with the lowest branch metric values becomes the detected sequence. (Id.). Thus, the detector calculates the path with the lowest cumulative branch metric value to determine the detected sequence of zeros and ones written on the disk. (Id. at 172). 3. Noise Bit regions are not homogeneous. (Id. at 175). Rather, they are made up of small tiles or magnetic grains that create regions of magnetization that do not fall within straight bit regions on the track. (Id.).

As the bit regions become narrower in high density recording and more bits are packed onto a smaller area, there will be fewer grains per bit region. (Id. at 176). With fewer grains, islands of grains may develop in which the detector cannot accurately read the data. (Id. at 176-177). This is shown below in a diagram in which green represents “zeros,” and blue represents “ones.” (Docket No. 771 at Ex. C at 24). So, as the density of the recording increases, the amount of noise or uncertainty in the signal also increases. (Docket No. 673 at 179).

As seen below, the amount of noise is also affected by the specific sequence of bits written on the track. (Docket No. 771 at Ex. C at 25). This is correlated signal-dependent noise, because the noise signals from one boundary to the other move together, either attracting or moving away from each other. (Docket No. 673 at 179). Noise was previously assumed to be white, or flat, at all time instances and in all branches. (Id. at 183-184). Using this noise assumption in determining disk signals worked in the low density environment of the 1970s and 1980s. (Id. at 184).

A Viterbi-like detector computed Euclidean branch metric values based on the assumption that the noise was white. (Id.). Next, the industry used another assumption, that of correlated noise, where the noise had structure but the structure was the same regardless of the symbol sequence (i.e., written symbols). (Id. at 186). The current assumption is that of correlated signal-dependent noise. (Id. at 193). This is media noise in the read back signal, whose noise structure is attributable to a specific sequence of symbols. (Id.). Below is a comparison of the three forms: (Docket No. 771 at Ex. C). 4.

The CMU Patents With the last model of signal-dependent noise, the detected sequence is obtained by maximizing the likelihood function. (Docket No. 673 at 206-207). The CMU Patents start by showing that such a likelihood function is dependent on all the read back signals and all written symbols from the entire disk. (Docket No. 673 at 206-207). This is expressed as: '839 Patent Eq. 1. As there are billions of symbols on the disk, the likelihood function is broken up into smaller per sample functions. (Docket No. 673 at 208).

The CMU Patents derived a function based on the observed signal samples; postulated a sequence of written symbols; then applied certain mathematical manipulations to turn the function into a quotient of a likelihood function, as seen below. (Id. at 214-215). '889 Patent Eq. 4-6. The resulting function can be used to create different embodiments, as disclosed in the CMU Patents. (Docket No. 673 at 220). One embodiment is called the correlation matrices embodiment, expressed in Equation 13 of the '839 Patent: i '839 Patent Eq. 13; (Docket No. 673 at 221).

Another form of embodiment is the Finite Impulse Response (“FIR”) embodiment. The FIR filter coefficients operate on a plurality of signal samples and are different for each specific sequence of written symbols. (Docket No. 673 at 225-226). When applied to the FIR embodiment, the coefficients account for the structure of signal dependent noise attributable to that specific sequence. (Id. at 227). Once the FIR has been applied to all of the sequences to account for the noise of a specific sequence, a Viterbi-like detector can work on the result. (Id. at 228).

As the recording density increases, such detectors become better compared to signal insensitive detectors. (Id. at 234). There is little benefit to increasing the amount of data on a disk if it cannot be accurately read. (Id. at 90). The patents’ solution constitutes the “optimal” detector, such that when the media noise is the dominant factor there is no better solution. (Id. at 70-71; Docket No. 677 at 170). Media noise became increasingly significant as the industry moved from longitudinal to perpendicular recording around 2005. (Docket No. 678 at 53-53,114, 226).

Media noise has become the main limiting factor in accurately reading bits from the disk, with 90% of the noise in read channels coming from media noise. (Docket No. 673 at 54). Consequently, academic institutions and private industry undertook research to address this media noise problem at both a theoretical level and product implementation level. (Id. at 40-41, 141; Docket No. 707 at 233). B.

CMU/DSSC Background Carnegie Mellon University is a leading research university located in Pittsburgh, Pennsylvania, with highly ranked engineering, information technology, and computer science programs. (Docket No. 671 at 187). (Id.). Dr. Jared Cohon had been the President of the University since 1997. (Id.). In response to the storage industry’s decreasing presence in the United States, CMU created the Magnetics Technology Center (MTC) in 1983 with the support of industrial funding. (Docket No. 682 at 27).

In collaboration with the National Science Foundation, the MTC became the Data Storage Systems Center (“DSSC”) in 1990. (Id.). The DSSC is an interdisciplinary center at CMU, funding long-term research and development through federal grants, university investments, and corporate sponsorship. (Docket No. 682). Since 1983, business partners have become “associate members” of the DSSC, by paying a yearly $250,000 fee, sponsoring faculty chair positions, hiring students and making other investments in the University. (Docket No. 682 at 41-42).

DSSC Members at one point or another have included IBM, Seagate, 3M, Hitachi, and Western Digital. (Id. at 43-44). Dr. Mark Kryder testified at trial regarding the history and nature of the DSSC. (Id.). Dr. Kryder was the co-founder and director of the DSSC until 1998, when he left to join Seagate where he eventually became the Chief Technology Officer (“CTO”). (Id. at 25). Upon retiring from Seagate, he returned to CMU and the DSSC. (Id.). According to Dr. Kryder, associate members of the DSSC received disclosures of inventions created in the DSSC and enjoyed a royalty-free license to same. (Id. at 43-44).

The patents-in-suit were invented within the DSSC. (Id. at 43-44). As such, the parties agree that the DSSC members at the relevant time had a royalty-free license to the patents. (Id.). However, Marvell was never a DSSC Member. (Id. at 42). C. Development and Acquisition of CMU Patents From 1995 to 1998, Professor Jose Moura of the Department of Electrical and Computer Engineering and doctoral student Aleksandar Kavcic worked to develop a method for addressing problems in high density and media noise problems related to new generation sequence detectors. (Docket No. 673 at 42).

On March 10. 1997, they submitted an invention disclosure form to CMU’s technology transfer office regarding same. (PI. Ex. 156). The provisional patent application was filed in May 1997, with the final patent application being filed on April 3, 1998. (PI. Ex. 1). This patent which was granted on March 13, 2001 is referred to as the '839 Patent. (Id.). On March 1, 1999, they filed for a second patent as a continuation in part of the first. (PL Ex. 2). It was granted on August 20, 2002, and it is referred to as the '180 Patent. (Id.).

In 1998 and 2000, the pair published papers: “Correlation-Sensitive Adaptive Sequence Detection” in IEEE Transactions on Magnetics and “The Viterbi Algorithm and Markov Noise Memory” in IEEE Transactions on Information Theory, describing their work. (Pl. Ex. 169,183). In 1998, Aleksandar Kavcic received his Ph.D. and left CMU to join the faculty at Harvard University. (Docket No. 673 at 149). He is currently a professor of electrical engineering at the University of Hawaii. (Docket No. 673 at 149). Dr. Moura remains a professor of electrical engineering at CMU. (Id. at 37).

Pursuant to CMU’s policy, half of any proceeds that CMU realizes on these patents, including from this lawsuit, are split between the inventors, Dr. Kavcic and Dr. Moura. (Docket No. 671 at 194-195). D. CMU’s Marketing of the CMU Patents CMU’s Center for Technology Transfer and Enterprise Creation, currently headed by Robert Wooldridge, is tasked with managing the University’s intellectual property. (Docket No. 682 at 96).

In August 2003, Carl Mahler, Esq., a subordinate of Robert Wooldridge, sent fourteen letters to several companies, including Marvell, Toshiba, Western Digital, Fujitsu, Samsung, Hitachi, and Maxtor, asking if they would be interested in licensing the '180 and '839 Patents. (PI. Ex. 422, 431; Def. Exs. 225; 226; 227; 229; 230; 231; 232; 233; 234; 1573). Not all of these companies make read channel or system-on-a-chip (“SOC”) products. (Docket No. 682 at 149-153). Two companies contacted CMU declining to license the technology; the rest never responded. (Id.).

CMU entered into a subscription agreement with its long-time corporate partner Intel in September 2004. (Def. Ex. 255). For a yearly administration fee of $75,000, Intel would have the option to license patents from a rotating pool. (Docket No. 682 at 185). If the inventors approved the licensing, Intel could license a patent for a one-time fee of $200,000. (Def. Ex. 255; Docket No. 682 at 183). In early 2005, it was proposed that the '180 Patent join the pool and in an email regarding same, th.e inventors also expressed interest in having said patent be part of the pool. (Def. Ex. 263).

Regardless of the inventors’ recommendation, the Patent was not licensed by Intel. (Docket No. 682 at 100). E. Marvell and Pertinent Staff A leading fabless semiconductor company, Marvell was founded in 1995 by Dr. Sehat Sutardja, along with his wife, Weili Dai, and brother, Dr. Pantas Sutarjda. (Docket No. 707 at 35). Defendants, Marvell Semiconductor Inc., a California corporation, and Marvell Technology Group, Ltd., a Bermuda corporation, both have their primary place of business in Santa Clara, California. (Id. at 29-34).

Marvell designs and develops a wide variety of integrated circuit devices, including read channel and SOC devices, used in storage products such as hard disk drives. (Id.). Marvell researches, designs, and develops its read channel and SOC products, including the accused products in this case, in Santa Clara. (Id.). The company has grown from seven employees to nearly 7,000 employees, and is now a publicly traded company. (Id. at 53). Dr. Sehat Sutardja is the current President and CEO; Ms. Dai is the Vice President and General Manager of Communications and Consumer Business; and Dr.

Pantas Sutardja is the Director, Vice President, CTO, and Chief R & D Officer. See (Docket No. 707); see also Marvell Company-Global Semiconductor Leadership, http://investor.marvell.com/phoenix.zhtml? c=120802&p=irol-govmanage (last visited September 20, 2013). Together these three individuals own 19% of Marvell. (Docket No. 707 at 146). Dr. Alan Armstrong is the Vice President of Marketing, Storage Business Group and was the company’s Rule 30(b)(6) deposition witness regarding same. (Docket No. 707 at 31). Bill Brennan is the former Vice President of Sales, Storage Business Group. (Docket No. 707 at 31). Mr.

Michael O’Dell is the worldwide director of field application engineering at Marvell and worked for Mr. Brennan in the early 2000s. (Docket No. 726 at 233). Relevant engineering employees include Dr. Zining Wu, Gregory Burd, and Toai Doan. Dr. Wu joined Marvell in 1999 after receiving his Ph.D. in Electrical Engineering from Stanford University. (Docket No. 707 at 217-219). Mr. Burd joined Marvell in the same year. (Docket No. 726 at 129). Mr. Doan, was a manager and principal engineer of signal processing and later Vice President of read channel development. (Docket No. 761 at Jt. Ex. D-l at 1). Mr.

Doan left Marvell in October 2009. (Id.). Last, Dr. Nersi Nazari was Mr. Burd’s acting manager in the early 2000s. (PI. Exs. 280; 283; 285). Of note, he was also Dr. Kavcic’s contact within Marvell. (Def. Ex. 1023). In fact, Dr. Kavcic emailed him in early March 1998, inquiring about Marvell’s work on detectors, sending him a link to his recent publications, resume and work, and seeking information on job openings at Marvell. (Def. Ex. 1023). F. Marvell’s Development of Read Channel Products Both Dr. Wu and Mr.

Burd worked to develop new technologies for digital signal processing and read channel technologies. (Docket No. 707 at 221). As discussed, a read channel is situated between a drive’s controller and the analog recording head, providing an interface so that digital data can be read from the disk. (Id. at 30). As data is packed more tightly onto the disk, errors arise from adjacent data tracks. (Id. at 96). The extent to which the error can be corrected limits how much data can be stored on the disk. (Docket No. 707 at 231).

Hence, the team at Marvell worked to increase the signal-to-noise ratio (“SNR”) in its read channel chips, addressing media noise and other sources of noise, such as asymmetric noise, baseline wander, and thermal noise. (Id. at 230-234). One of their earlier projects from around 1999 to 2001 was implementing iterative coding, a different method of improving SNR on chips. (Docket No. 678 at 119). This form of coding was also the basis of Dr. Wu’s Ph.D. thesis at Stanford. (Docket No. 707 at 255). However, iterative coding was not initially successful for Marvell. (Docket No. 687 at 119-124). In fact, Mr.

Doan called these chips a “lost cause” and Mr. Brennan said many referred to them as “coffee warmer” chips because they used so much power. (Id.). G. MNP/NLD Chip Development Because iterative coding was not initially successful, the team at Marvell continued to work on other potential solutions to the noise problem. (Docket No. 687 at 119-124). In 2001, Mr. Burd read the papers by Dr. Kavcic and Dr. Moura explaining their invented method and he began working on developing a “solution” for Marvell based on same. (Docket No. 726 at 137). CMU asserts that Marvell “copied” Dr.

Kavcic’s method as described in said articles and claimed in the '180 and '839 Patents. (Docket No. 677 at 54-55). In response, Marvell maintained throughout trial that Mr. Burd had developed his own sub-optimal “solution,” using Dr. Kavcic’s scheme only as a launching pad, as later expressed in Marvell’s U.S. Patent Number 6,931,585, which listed the '180 Patent as prior art. (Docket No. 726 at 125-135). The method Mr.

Burd developed, originally named KavcicPP, was renamed “MNP” in January 2003 and later incorporated into Marvell’s EMNP and NLD technology, all of which are used on read channel chips and SOC chips (collectively, the “Accused Chips”). (PI. Exs. 368; 823). In connection with same, simulators were developed by Marvell engineers to mimic chips, so that Marvell engineers could run and test the chip systems before producing the chips in silicon. (Docket No. 707 at 113-114).

Marvell also ran all of its chips against what it considered the optimal benchmark simulator, called KavcicViterbi. (Docket No. 677 at 171-172; Docket No. 761 at Jt. Ex. D at 137-138). The KavcicPP, MNP, EMNP, and NLD type simulators and the KavcicViterbi simulator are collectively the “Accused Simulators” in this litigation. (PI. Exs. 89; 99; 106; 108; 110). The Accused Chips and Accused Simulators are collectively referred to as the “Accused Technology.” H.

Marvell’s Sales of Accused Chips Read channel chips were dominant until around 2004, when the industry transitioned to SOC-type chips. (Docket No. 707 at 309-311). SOC chips integrated several different blocks, including the read channel block, onto one chip to improve speed, power, and cost. (Id. at 226). Today, Marvell has about 800 employees involved in the development of SOC chips. (Id.). CMU alleged infringement by Marvell’s MNP, EMNP-type and NLD type chips, encompassing both read channel and SOC chips. (Docket No. 671).

All of the Accused Chips were custom made to the exact specifications requested by the customer. (Docket No. 678 at 50-142). As noted, the custom designs and sample engineering chips are developed and tested by using them with the Accused Simulators. Sample engineering chips are produced in Asia and sent back to the United States to be tested by both Marvell and its customers. (Docket No. 678 at 105-106; Docket No. 707 at 164). Once the customer places an order, the chips are put into production at Taiwan Semiconductor Manufacturing Company’s (“TSMC”) foundry in Taiwan. (Docket No. 707 at 164).

Marvell field application engineers then assist the hard drive company to install chips into their product and instruct them on how to use the chips. (Docket No. 677 at 178-179). As stipulated by the parties, Marvell sells its chips through a lengthy sales cycle, fin which Marvell must invest significantly in each customer without the assurance of sales. (Docket No. 707 at 32-35). There is first a 3-6 month period of rigorous evaluation and reliability testing by the customer in a stage called “qualification.” (Id.).

This is followed by a 12-18 month development period and then a 3-6 month period before Marvell commences volume production (i.e. until 1 million units are produced). (Id.). Throughout this entire cycle, there is a significant risk the customer will change its mind before the design is selected and the time and expense incurred by Marvell will generate no revenue. (Id.). Since a customer usually uses a selected design for a full generation, the loss of a sales win cannot be remedied until the customer develops a new product or new generation. (Id.).

During this sales cycle, Marvell engineers assist the customer in implementing the Marvell solutions into its product. (Id.). Almost all of this activity, including sales, marketing, evaluation, testing, and development occurs in Santa Clara, California. (Id.). The Accused Simulators are used at various points throughout this sales cycle to formulate product concepts and to design, refine and evaluate chip designs. (Id. at 45). As CMU’s expert witness Dr. Steven McLaughlin testified, the simulators are used for research and development to verify the hardware design for the chip. (Docket No. 677 at 158).

Marvell provides the simulation code to its customers so they can evaluate the functionality and performance of a chip design. (Id.). Marvell’s major customers are Fujitsu, Hitachi/IBM, Maxtor, Samsung, Seagate, Toshiba, and Western Digital. (Docket No. 710 at 243-244). All of these customers go through this process with Marvell at its Santa Clara location. Once the customer is satisfied that the design and testing have met its specifications, the chip designs and engineering samples are sent back to TSMC to begin volume production. (Docket No. 678 at 92).

According to Marvell sales data, between March 6, 2003 and July 28, 2012, Marvell sold 2.34 billion Accused Chips. (Docket No. 686 at 61). The average revenue per accused chip is $4.42, with an average operating profit of $2.16. (Id. at 53-54). As noted, the chips are manufactured by TSMC foundry in Taiwan, and then sent to the customers’ manufacturing sites in Asia to be put into their HDDs. (Docket No. 710 at 360-361). These HDDs are then sold primarily to laptop manufacturers, which incorporate the HDDs into their products at their own factories. (Id.).

A portion of the laptops are then imported back into the United States. (Id. at 164-165). The locations of the chips’ end users are unknown, but CMU presented estimates based on import data calculated by its damages expert that 329,297,799 or 556,812,092 of the 2.34 billion Accused Chips were imported in to the United States. (Id. at 165; Docket No. 770-11 at 7). I. Correspondence Regarding CMU Patents In January 2002, Mr. Burd sent two emails to Mr. Doan, who was then his boss at Marvell, stating that the Kavcic method was patented and assigned to CMU. (PI. Exs. 280; 283).

The following year, Carl Mahler of the CMU Technology Transfer Office sent out fourteen letters to various technology companies, including two addressed to Marvell personnel Dr. Pantas Sutardja and then-General Counsel Matthew Gloss, encouraging these companies to contact CMU if they were interested in licensing the '180 and the '839 Patents. (PI. Exs. 422; 431). There was no known response by Marvell to these letters from Mr. Mahler.

In 2004, Fujitsu, a read channel customer, wrote a letter to Marvell, stating that it had become aware of the '839 Patent and asked for Marvell’s position regarding the relationship between these patents and its own technology. (PI. Ex. 477). There is no known response to this letter. (Docket No. 761 at Ex. C at 531-535). J. Parties’ Evidence as to Infringement At trial, CMU maintained that Marvell makes, uses, offers to sell, or sells chips and uses simulators that infringe Claim 4 of the '839 Patent and Claim 2 of the '180 Patent.

Claim 4 of the '839 Patent provides: [a] method of determining branch metric values for branches of a trellis for a Viterbi-like detector, comprising: selecting a branch metric function for each of the branches at a certain time index from a set of signal-dependent branch metric functions; and applying each of said selected functions to a plurality of signal samples to determine the metric value corresponding to the branch for which the applied branch metric was selected wherein each sample corresponds to a different sampling time instant. '839 Patent col. 1411.10-19.

Claim 2 of the '180 Patent, incorporates Claim 1. Claims 1 and 2 of the '180 Patent state: 1. A method of determining branch metric values in a detector, comprising: receiving a plurality of time variant signal samples, the signal samples having one of signal-dependent noise, correlated noise, and both signal dependent and correlated noise associated therewith; selecting a branch metric function at a certain time index; and applying the selected function to the signal samples to determine the metric values. 2.

The method claim 1, wherein the branch metric function is selected from a set of signal-dependent branch metric functions. '180 Patent col. 15 11. 39-51. CMU argued that Marvell’s MNP and NLD Chips infringed these claims. CMU also asserted that Marvell’s KavcicPP Simulator, MNP Simulator, EMNP Simulator (these three collectively the MNP-Type Simulators), NLD Simulator, and KavcicViterbi Simulator infringed Claim 4 of the '839 Patent and Claim 2 of the '180 Patent. Counsel for the parties prepared a stipulation on the chip technology, affectionately called the “Chip Stip.” (PI. Ex. 823).

The parties agreed that the circuits set forth therein are true and accurate depiction of the circuits within Marvell’s products. (Id.) The stipulation identified which Marvell read channel and SOC models correspond to each of the stipulated circuits. (Id.) CMU submitted the stipulation as evidence of infringement to the jury. (PI. Ex. 823). Marvell’s code for the KavcicPP Simulator (Pl. Ex. 110), MNP Simulator (PI. Ex. 99), EMNP Simulator (Pl. Ex. 89) (these three collectively, the “MNP-Type Simulators”), NLD Simulator (Pl. Ex. 106), and KavcicViterbi Simulator (Pl.

Ex. 108) was also admitted as evidence of infringement. Further, the parties presented competing expert opinions on infringement. For CMU, Dr. Steven McLaughlin opined that Marvell’s chips and simulators infringed CMU’s patents. (Docket No. 677). Dr. McLaughlin testified over two days using a PowerPoint presentation with over 130 slides to help demonstrate his infringement analysis. (Docket No. 677-678). In doing so, Dr. McLaughlin analyzed the Chip Stip, Simulator Code, Marvell’s technical documents, and relevant deposition testimony from Marvell’s engineers in reaching his conclusions.

At trial, he broke down both claims into elements and demonstrated to the jury how the circuitry of the MNP Chips and NLD Chips, in addition to the code of the Accused Simulators, infringed each and every step of the two patents. (Docket No. 771). He was clear that as these are method claims, infringement only occurs when the method is actually run on the chips or simulators. (Id.). Marvell countered with Dr. Richard Blahut, who opined that there was no infringement of the CMU Patents. (Docket No. 711). Dr.

Blahut believes that in the Marvell products, the Viterbi algorithm uses a simple branch metric function that uses the same branch metric function on every branch of the trellis, so there is no selecting step as required in the patents. (Id. at 244). He also opined that there was no selecting function in the Viterbi detector or post processor; hence, there was no applying step as required by the CMU Patents. (Id. at 246). K. Parties’ Evidence as to Invalidity During trial, Marvell maintained that CMU’s Patents were invalid because they were both anticipated and obvious. To this end, Marvell submitted U.S.

Patent No. 6,282,251 (the “Worstell Patent”) as prior art for purposes of its anticipation defense. (Def. Ex. 187). This patent was filed on March 21, 1995, three years before the CMU Patents were filed. (Id.). As further evidence of invalidity, Marvell also presented a 1992 IEEE

article by Inkyu Lee and John Cioffi titled “Performance Analysis of the Modified Maximum Likelihood Sequence Detector in the Presence of Data-dependent Noise” and a 1992 IEEE Transactions on Magnetics

article by Weining Zeng and Jaekyun Moon titled “Modified Viterbi Algorithm for a Jitter-dominant 1-D2 Channel” (Def. Exs. 37; 38). Again, the parties had dueling expert witnesses appear. Dr. John Proakis testified for Marvell, and opined that the CMU Patents were invalid based on the aforementioned prior art. (Docket No. 726). CMU called Dr. McLaughlin in rebuttal to testify on the subject of invalidity. (Docket No. 736). Dr. McLaughlin countered Dr. Proakis’ testimony and concluded that the CMU Patents were not invalid based upon the two IEEE articles and/or the Worstell Patent. (Id. at 73).

CMU also submitted a critical 1997 email from Glen Worstell, stating in relevant part that he: had reviewed the ‘Correlation Sensitive Adaptive Sequence Detector’ patent proposal (i.e. the proposal of the CMU Patents) ... A couple of years ago I did some work on a Viterbi detector modification to account for noise correlation. This invention is related, but goes beyond my work and is probably more interesting. (PI. Ex. 161). L. Damages Evidence CMU sought money damages from Marvell for infringement, in the form of a $0.50 per chip royalty on all Accused Chips sold by Marvell from March 2003 to the present.

CMU proffered evidence that the Accused Technology was “must have” for Marvell and thus the parties would have agreed to this running royalty at a hypothetical negotiation in March 2003. In support of its position, CMU first called Dr. Christopher Bajorek as an industry expert. (Docket No. 678 at 72-73). Dr. Bajorek opined that Marvell and its customers used the MNP and NLD technologies during the sales cycle; the sales cycle essentially took place in the US; that the MNP and NLD technology had become industry standard; and that the same technology was “must have” for Marvell. (Id.). Dr.

Bajorek testified that Seagate, IBM, HDST, Samsung, Western Digital, and Toshiba use or previously used the patented technology. (Id. at 163-165). Marvell did not counter Dr. Bajorek with a competing expert in his area of expertise. CMU next called Catherine Lawton as its damages expert. (Docket No. 686 at 29). She stated that Marvell sales data showed sales of 2.34 billion Accused Chips between March 6, 2003 and July 28, 2012. (Docket No. 686 at 61).

She then analyzed sales data provided by Marvell to calculate an “excess profits” benchmark of $0.42 per chip and “operating profit premium” benchmark of $0.06 to $0.72 per chip, which she used along with other pertinent facts to arrive at a reasonable royalty of $0.50 per chip. (Docket No. 710 at 170-171). Her analysis is examined in more detail later herein. CMU also submitted supporting evidence in the form of internal Marvell communications and presentations, including Marvell presentations to customers, deposition testimony from Marvell sales and marketing executives such as Mr. Brennan and Dr.

Armstrong, as well as the joint stipulation regarding Marvell’s sales cycle. (Pl. Exs. 220; 240; 244; 297; 331; 333; 651; 938). Marvell rebutted this damages calculation by presenting its own damages expert, Creighton Hoffman. (Docket Nos. 709; 710). Mr. Hoffman based his opinion primarily on the DSSC Agreements (Def. Exs. 17; 39; 40), the Intel offer to license (Def. Ex. 255), and his perception of a lack of marketing and licensing of the patents by CMU or their Inventors. (Docket No. 709).

His ultimate opinion was that a reasonable royalty in this case would be a one-time royalty payment of $250,000.00. (Id. at 242-245). Marvell did not submit any evidence on other licensing agreements or alternative pricing opinions. M. Evidence of Alleged Willfulness CMU argued that Marvell’s infringement had been willful by submitting evidence of Marvell’s internal communication about the patents, including the aforementioned emails from Mr. Burd (PI. Ex. 280, 283), the letters received from CMU (PI. Exs. 422; 431), the letter from Fujitsu (PL Ex. 477), and deposition testimony of Dr. Wu, Mr. Doan, Dr.

Armstrong, and Mr. Burd. See (Docket No. 677 at 53-55; Docket No. 761 at Jt. Ex. C, D). CMU also submitted Dr. McLaughlin’s expert testimony to the extent that he opined that the MNP was “copied” from the CMU Patents. (Docket No. 677 at 82). Marvell presented evidence to show that it had not willfully infringed, relying on internal Marvell correspondence and presentations on the Accused Technologies and proof of Marvell’s own patents, some of which cite the CMU Patents. (Def. Ex. 266). Marvell also offered portions of Dr. McLaughlin’s deposition testimony to disprove copying from Dr. Kavcic and Dr.

Moura, as well as testimony at trial, such as that of Mr. Burd, (Docket No. 726 at 125-126), and Dr. Wu, (Docket No. 707 at 326), who stated they did not- copy the CMU Patents. As Dr. Sehat Sutardja testified, Marvell’s people “are not stealers.” (Docket No. 707 at 92, 326). N. Jury Verdict On December 21, 2012 the jury was charged to decide issues of infringement, validity, damages, and willfulness given all of the evidence before it. The jury deliberated for nearly two days to render its verdict, returning on December 26, 2012. (Docket No. 762).

With respect to infringement, the jury found that CMU had proven by a preponderance of the evidence that Marvell’s MNP-Type chips, MNP-Type simulators, NLD-Type chips, NLD-Type simulators, and Kavcic-Viterbi simulator literally infringe Claim 4 of the '839 Patent and Claim 2 of the '180 Patent. (Id. at Q. 1-10). The jury held that CMU had proven by a preponderance of the evidence that Marvell had induced at least one of its customers or an end user to infringe Claim 4 of the '839 and Claim 2 of the '180 Patent in the United States with both the MNP-Type and NLD-Type Chips. (Id. at Q. 11,13).

It additionally found that CMU had proven by a preponderance of the evidence that Marvell contributed to the infringement of Claim 4 of the '839 and Claim 2 of the '180 Patent in the United States by at least one of its customers or an end user with both the MNP-Type and NLD-Type Chips. (Id. at Q. 12,14). On invalidity, the jury found that Marvell had not proven by clear and convincing evidence that Claim 4 of the '839 and Claim 2 of the '180 were invalid on the grounds that they were anticipated by pri- or art or because they would have been obvious at the time the invention was made. (Id. at Q. 15, 16).

After finding that the claims infringed and are not invalid, the jury awarded $1,169,140,271.00 to CMU for the use of the patented methods. (Id. at Q. 17). Regarding willfulness, the jury found that Marvell had actual knowledge of the '180 and '839 Patents prior to commencement of the lawsuit on March 6, 2009. (Id. at Q. 19, 22). It determined that Marvell did not have an objectively reasonable defense to CMU’s claim of infringement on either the '180 or '839 Patent. (Id. at Q. 20, 23).

Finally, the jury found that once Marvell learned of the '180 and '839 Patent, there was clear and convincing evidence that Marvell actually knew or should have known that its actions would infringe both Claim 2 of the '180 Patent and Claim 4 of the '839 Patent. (Docket No. 21, 24). III. PROCEDURAL HISTORY CMU filed its complaint in this case on March 6, 2009. (Docket No. 1). Since then this case had gone through extensive discovery and motions practice, including a Motion to Transfer (Docket No. 55), Claim Construction (Docket No. 143), and several rounds of

Summary Judgment proceedings. Following the Court’s

Summary Judgment and Daubert rulings, and upon consideration of the parties’ pretrial proffers, the Court convened a two-day hearing on October 17 and October 18, 2012 to address nineteen Motions in limine (five by CMU and fourteen by Marvell). (Docket Nos. 578; 579; 590; 591). The Court issued rulings shortly thereafter. (Docket Nos. 595; 596; 601; 602; 604; 605; 607-614).

On November 9, 2012, the parties filed their, responses to objections to exhibits, responses to objections to deposition designations, joint stipulations, proposed jury instructions, proposed limiting instructions, proposed voir dire, and proposed verdict slips. See (Docket Nos. 615-626; 640-644). The Court held a lengthy two-day Pretrial Conference on November 14 and 15, 2012, during which objections to exhibits were ruled upon and arguments on trial issues were heard. (Docket Nos. 636; 638; 645; 648; 650; 653).

The parties subsequently submitted trial briefs on the issues of law that would need to be addressed by the Court during trial. (Docket Nos. 647; 652). Jury selection proceeded as scheduled on November 26, 2012, and trial began on November 28, 2012. (Docket Nos. 666; 669; 671). The Court heard argument, accepted briefing, and ruled on a number of motions made during trial regarding witnesses, exhibits, and points of law. The parties agreed that the Plaintiff, CMU would not rest its case until the conclusion of testimony by three defense witnesses.

Once CMU rested its case, Marvell moved for judgment as a matter of law on “Non-infringement,” “No Damages,” and “No Willful Infringement.” (Docket Nos. 703; 701; 699). At the end of Marvell’s evidence, CMU moved for “Judgment As a Matter of Law on Marvell’s Invalidity Defenses.” (Docket No. 731). Following CMU’s rebuttal, Marvell filed a Motion for Judgment as a Matter of Law on Invalidity, and renewed its earlier Motions for Judgment as a Matter of Law on Non-Infringement, No Damages, and No Willful Infringement. (Docket Nos. 738; 740; 742; 747).

The Court denied these motions on the record (Docket No. 759 at 52-53), with the parties requesting the Court to explain its rulings in written opinions. (Docket No. 764 at 99). The Court then charged the jury on December 21, 2102, and it returned its verdict on December 26, 2012. (Docket No. 762). As noted, the jury found for CMU on infringement, validity, and willfulness, and awarded damages to CMU in the amount of $1,169,140,271.00. (Id.). The Court entered the parties’ joint proposed form of judgment on January 14, 2013. (Docket No. 769).

Pursuant to the Court’s scheduling order, (Docket No. 763), on February 11, 2013, Marvell filed a Motion for Judgment as a Matter of Law or in the Alternative, Motion for New Trial on Non-Damages Issues, specifically for Non-Infringement, Invalidity, No-Willfulness, and CMU Misconduct (Docket Nos. 805; 806), Motion for Judgment as a Matter of Law, Motion for New Trial And/Or Motion for Remittitur with Respect to Damages, (Docket Nos. 807; 808), and Motion for Judgment on Laches. (Docket Nos. 802-04).

CMU moved for “Permanent Injunction, Post Judgment Royalties, and Supplemental Damages” (Docket Nos. 786; 787), “Prejudgment Interest” (Docket Nos. 788; 789), “A Finding of Willful Infringement and Enhanced Damages” (Docket Nos. 790; 793), and “Attorneys’ Fees Pursuant to 35 U.S.C.

Section 285” (Docket Nos. 794; 810; 811). These matters have been completely briefed (Docket Nos. 823-829; 832-837; 849-855; 857-863), and the Court heard oral argument on same from May 1 to May 2, 2013. (Docket Nos. 872-874). In earlier opinions, the Court had denied, without prejudice, CMU’s Request for Attorneys’ Fees (Docket No. 884), and denied Marvell’s Motion for a New Trial on the Grounds of CMU Misconduct. (Docket No. 900). The Court now turns to the parties’ Motions for JMOL, Motions for a New Trial, and Motion for Remittitur. (Docket Nos. 805; 807). IV.

LEGAL STANDARD A, Judgment as a Matter of Law It is well-established that a motion for judgment as a matter of law “should be granted only if, viewing the evidence in the light most favorable to the non-moving party, there is no question of material fact for the jury and any verdict other than the one directed would be erroneous under the governing law.” Galena v. Leone, 638 F.3d 186, 196 (3d Cir.2011) (quoting Beck v. City of Pittsburgh, 89 F.3d 966, 971 (3d Cir.1996)).

Accordingly, the Court must determine “ ‘whether there are any genuine issues of material fact such that a reasonable jury could return a verdict for [the non-moving party].”’ McGreevy v. Stroup, 413 F.3d 359, 364 (3d Cir.2005) (quoting Debiec v. Cabot Corp., 352 F.3d 117, 128 n. 3 (3d Cir.2003)) (alteration in original); see also Trueman v. City of Upper Chichester, 289 Fed.Appx. 529, 540 (3d Cir.2008) (affirming denial of Rule 50(

a) motion because “the jury could not reasonably have found in [the non-movant’s] favor on his claim against the [movant]”). In ruling on a Rule 50(

a) motion, the Court “must refrain from weighing the evidence, determining the credibility of witnesses, or substituting our own version of the facts for that of the jury.” Eshelman v. Agere Sys., 554 F.3d 426, 433 (3d Cir.2009) (citing Marra v. Philadelphia Housing Auth., 497 F.3d 286, 300 (3d Cir.2007)). “Although judgment as a matter of law should be granted sparingly,” it should be granted where “the record is critically deficient of the minimum quantum of evidence” necessary to support a verdict in favor of the non-moving party. Id. (quoting Gomez v.

Allegheny Health Servs., Inc., 71 F.3d 1079, 1083 (3d Cir.1995)). To that end, “a scintilla of evidence is not enough” to survive a Rule 50 motion at trial. Johnson v. Campbell, 332 F.3d 199, 204 (3d Cir.2003) (citing Goodman v. Pa. Turnpike Comm’n., 293 F.3d 655, 664-65 (3d Cir.2002)). The question is not whether there is literally no evidence supporting the unsuccessful party, but whether there is evidence upon which a reasonable jury could properly find a verdict in favor of the non-moving party. Gomez, 71 F.3d at 1083.

In other words, “a directed verdict is mandated where the facts and the law will reasonably support only one conclusion.” McDermott Int’l, Inc. v. Wilander, 498 U.S. 337, 356, 111 S.Ct. 807, 112 L.Ed.2d 866 (1991) (citation omitted). B. Motion for New Trial A motion for a new trial pursuant to Federal Rule of Civil Procedure 59 can be granted “to all or any of the parties and on all or part of the issues in an action in which there has been a trial by jury.” Fed. R. Civ. P. 59(a).

The Court is also “empowered to order a new trial on its own initiative ‘for any reason that would justify granting one on a party’s motion.’ ” Pryer v. C.O. 3 Slavic, 251 F.3d 448, 453 (3d Cir.2001) (quoting Fed. R. Crv. P. 59(d)). A new trial is most commonly granted in select situations, including: (1) when the jury’s verdict is against the clear weight of the evidence; (2) when new evidence surfaces that would have altered the outcome of the trial; (3) when improper conduct on the part of an attorney or the court unfairly influenced the verdict; or (4) where the jury’s verdict was facially inconsistent.

Davis v. Mountain Farms, Inc., 598 F.Supp.2d 582, 587 (D.Del.2009). The Court’s level of discretion varies, depending on the type of error alleged. Moussa v. Commonwealth of Pennsylvania Dep’t of Pub. Welfare, 289 F.Supp.2d 639, 648 (W.D.Pa.2003) (citing Klein v. Hollings, 992 F.2d 1285, 1289-90 (3d Cir.1993)).

When the motion for a new trial is based on the claim that the verdict is against the clear weight of the evidence, the Court’s discretion is limited — the verdict must be “contrary to the great weight of the evidence; that is, where a miscarriage of justice would result if the verdict were to stand.” Pryer, 251 F.3d at 453. A verdict may not be set aside when there is a plausible or rational basis for the decision. Moussa, 289 F.Supp.2d at 648. The Court must not substitute its own judgment of the facts and assessment of the witnesses’ credibility for the jury’s. Davis, 598 F.Supp.2d at 587.

When the basis for the motion is an alleged error on the part of the court, such as an error in jury instructions or evidentiary rulings, a district court must first determine whether an error was made, i.e., “whether, taken as a whole, the instruction properly apprised the jury of the issues and the applicable law.” Donlin v. Philips Lighting N. Am. Corp., 581 F.3d 73, 78 (3d Cir.2009). If there was an error, the court must then determine “whether that error was so prejudicial that refusal to grant a new trial would be ‘inconsistent with substantial justice.’ ” Bhaya v. Westinghouse Elec.

Corp., 709 F.Supp. 600, 601 (E.D.Pa.1989) (quoting Fed. R. Civ. P. 61). “Generally, a party is not entitled to receive a new trial for objections to evidence that he did not make at or prior to the initial trial, even if they may have been successful.” Ashford v. Bartz, Civ. No. 04-642, 2010 WL 272009, at *4 (M.D.Pa.2010) (citations omitted); see also Kiewit Eastern Co., Inc. v. L & R Constr. Co., Inc., 44 F.3d 1194, 1204 (3d Cir.1995) (“Courts often take a dim view of issues raised for the first time in post-judgment motions. Generally, this is a decision within the sound discretion of the district court.”). V.

DISCUSSION With these standards in mind, the Court discerns substantial evidence upon which a reasonable jury could have found in favor of the non-movants for each of the filed Motions challenging the evidence. A. Infringement CMU had the burden of proving its claims of direct and indirect infringement pursuant to 35 U.S.C. §§ 271(a), (

b) and (c). It alleged that Marvell’s chips and simulators infringe Claim 4 of the '839 Patent and Claim 2 of the '180 Patent. To that end, it produced the opinion testimony of Dr. McLaughlin, who testified over the course of two days. (Docket Nos. 677; 678). Against same, Marvell moved for JMOL on the grounds that CMU had not presented sufficient evidence that a reasonable jury could find infringement by Marvell, or, in the alternative, for a new trial. (Docket Nos. 703; 805). 1.

Legal Standard Direct infringement of a U.S. patent occurs when a party, “without authority makes, uses, offers to sell, or sells any patented invention, within the United States.” 35 U.S.C. § 271(a). Method claims are not infringed simply by the sale of an apparatus that is capable of infringing use. Ormco Corp. v. Align Tech., Inc., 463 F.3d 1299, 1311 (Fed.Cir.2006); Standard Havens Products, Inc. v. Gencor Industries, Inc., 953 F.2d 1360, 1374 (Fed.

Cir.1991). “Because a process is nothing more than the sequence of actions of which it is comprised, the use of a process necessarily involves doing or performing each of the steps recited.” NTP, Inc. v. Research in Motion, Ltd., 418 F.3d 1282, 1318 (Fed. Cir.2005). Thus, direct infringement of a method claim only occurs if each step of the claimed method is actually performed. See Muniauction, Inc. v. Thomson Corp., 532 F.3d 1318, 1328 (Fed.Cir.2008). In this case, the only form of direct infringement at issue is literal infringement.

Marvell literally infringes if Marvell’s chips and simulators use a method that includes each and every method step in Claim 4 of the '839 Patent or Claim 2 of the '180 Patent. Akamai Techs., Inc. v. Limelight Networks, Inc., 692 F.3d 1301, 1307 (Fed.Cir.2012). If Marvell’s methods as employed by its chips and simulators do not contain one or more method steps in that patent claim, Marvell does not directly infringe that claim. Id. Accordingly, literal infringement must be determined with respect to each patent claim, individually.

There are also two forms of indirect infringement: inducing infringement and contributory infringement. These modes of infringement are governed, respectively, by 35 U.S.C. § 271(

b) and (c). To prove inducement of infringement and contributory infringement, CMU must first prove there is direct infringement. Akamai Techs., 692 F.3d at 1308. Second, “inducement requires that the alleged infringer knowingly induced infringement and possessed specific intent to encourage another’s infringement.” DSU Med. Corp. v. JMS Co., 471 F.3d 1293, 1306 (Fed.Cir.2006) (en banc) (internal quotation marks omitted); see also Global-Tech Appliances, Inc. v. SEB S.A., — U.S. -, 131 S.Ct. 2060, 2068, 179 L.Ed.2d 1167 (2011).

It is enough that the inducer “cause[s], urge[s], encourage [s], or aid [s]” the infringing conduct and that the induced conduct is carried out. Akamai Techs., 692 F.3d at 1308 (internal citations omitted, emphasis added). A defendant must “actively induce” infringement which “require[s] knowledge of the existence of the patent that is infringed” or taking “deliberate actions to avoid confirming a high probability of wrongdoing.” Global-Tech, 131 S.Ct. at 2068-2070.

Thus, induced infringement occurs if Marvell actively induces someone else, such as one of Marvell’s customers, to use a method that is covered by Claim 4 of the '839 Patent or Claim 2 of the '180 Patent. Id. To prevail on a claim for contributory infringement, it must be shown that an infringer sold, offered to sell, or imported into the United States a component of an infringing product “knowing [the component] to be especially made or especially adapted for use in an infringement of such patent, and not a staple

article or commodity of commerce suitable for substantial non infringing use.” 35 U.S.C. § 271(c); see Lucent Techs, v. Gateway, Inc., 580 F.3d 1301, 1320 (Fed. Cir.2009). Thus, in this instance, contributory infringement occurs if Marvell sold or offered for sale a material component of the patented invention that was not a staple

article of commerce, and which Marvell knew was specifically made for use in practicing the claimed methods of either Claim 4 of the '839 Patent or Claim 2 of the '180 Patent. As with induced infringement, a claim for contributory infringement must contain allegations of the requisite knowledge of the patent-in-suit at the time of infringement. Mallinckrodt v. E-Z-Em Inc., 670 F.Supp.2d 349, 355 (D.Del.2009); see also Global-Tech, 131 S.Ct. at 2068. In addition, the patentee bears the burden of proving that the accused products have no substantial non-infringing uses. See Golden Blount, Inc. v. Robert H.

Peterson Co., 438 F.3d 1354, 1363 (Fed.Cir.2006). Before delving into its analysis, the Court notes that expert testimony is not necessary to prove infringement. In a case involving complex technology, however, the Federal Circuit has “repeatedly approved the use of expert testimony to establish infringement” and indeed “where the accused infringer offers expert testimony negating infringement, the patentee cannot satisfy its burden of proof by relying only on testimony from those who are admittedly not experts in the field.” Centricut, LLC v. Esab Grp., Inc., 390 F.3d 1361, 1370 (Fed.Cir.2004).

In reaching its decision, the Court has considered all of the parties’ arguments raised in their briefs and at trial, arguments made at the motion hearing held on May 1 and May 2, 2013, the transcript thereof, and the entire trial record along with the parties’ latest submissions. (Docket Nos. 703; 704; 729; 742; 743; 805; 806; 827; 851; 857; 880; 881). 2. Direct Infringement As the party alleging infringement of the method claims at issue, CMU must demonstrate that Marvell practices every step of the claimed method. See Muniauction, 532 F.3d at 1328. At trial, CMU called Dr.

Steven McLaughlin to provide expert technical testimony about CMU’s patents and whether they are infringed by Marvell’s MNP-type chips, NLD type chips, and related simulators. Dr. McLaughlin analyzed the documents produced by Marvell concerning the MNP, EMNP, NLD, and Simulator Technology and the Chip Stip. (PI. Ex. 823). a.

MNP/EMNP Chips To begin, Marvell asserts that no reasonable jury could find that the accused MNP/EMNP Chips infringe because: (1) they do not select a branch metric function for each of the branches of the trellis at a certain time index; (2) they do not apply each of said branch metric functions to a plurality of signal samples; and (3) the MNP/EMNP module does not determine branch metric values for branches of a trellis. (Docket No. 743). CMU counters that its expert Dr. McLaughlin has demonstrated otherwise through his mapping of the claims onto the Accused Technology. (Docket Nos. 704; 743).

CMU asserted that Marvell’s MNP Chips first infringed Claim 4 of the '839 Patent.

Claim 4 of the '839 Patent provides: [a] method of determining branch metric values for branches of a trellis for a Viterbi-like detector, comprising: selecting a branch metric function for each of the branches at a certain time index from a set of signal-dependent branch metric functions; and applying each of said selected functions to a plurality of signal samples to determine the metric value corresponding to the branch for which the applied branch metric was selected wherein each sample corresponds to a different sampling time instant. '839 Patent col. 14 11.10-19.

In order to show infringement of the '839 Patent, Dr. McLaughlin broke this claim into three “elements.” First, he mapped “a method of determining branch metric values for branches of a trellis for a Viterbi-like detector” onto the MNP technology via the circuits of the Chip Stip by showing how the MNP: (1) is a detector; (2) computes branch metric values for branches of a trellis; and (3) is a Viterbilike detector. (Docket No. 677 at 86-120).

On this point, CMU also proffered supporting internal Marvell documents, such as the 88c7500M Specification from August 2004, which stated the “MNP is an advanced post processing adaptive detector,” and Mr. Burd’s “Media Noise Processor” write up, which stated the “MNP is used to properly take media noise into account during the detection processor.” (PI. Exs. 472; 408) In addition, the “DSP Technical Presentation 5: Data Detection” by Dr. Hongxin Song from 2009 stated “Media noise post-processor is a partial nonlinear detector in data dependent noise channel.” (PI. Ex. 770).

This exhibit also contained circuit drawings of the MNP, trellis models, and detailed the “steps to calculate nonlinear branch metric.” (Id.). Second, Dr. McLaughlin pointed out where the MNP technology contained a method for “selecting a branch metric function for each of the branches at a certain time index from a set of signal-dependent branch metric functions” on the circuits from Exhibit A of the Chip Stip. (Docket No. 677 at 108-117). Third, Dr.

McLaughlin used Exhibit A of the Chip Stip to demonstrate how the MNP used a method “applying each of said selected functions to a plurality of signal samples to determine the metric value corresponding to the branch for which the applied branch metric was selected wherein each sample corresponds to a different sampling time instant.” (Id. at 118-119). He explained that the FIR filter implements the function and applies it to the plurality of signal samples which then computes the metric value, wherein each sample corresponds to a different time instant, referred to as “D.” (Id.).

After finding that each of the elements of Claim 4 of the '839 Patent was used by the MNP technology, Dr. McLaughlin opined that the MNP infringed Claim 4 of the '839 Patent. (Id. at 120). Likewise, CMU claimed Marvell’s MNP technology infringed Claim 2 of the '180 Patents, which incorporates Claim 1. Claims 1 and 2 of the '180 Patent claim: 1.

A method of determining branch metric values in a detector, comprising: receiving a plurality of time variant signal samples, the signal samples having one of signal-dependent noise, correlated noise, and both signal dependent and correlated noise associated therewith; selecting a branch metric function at a certain time index; and applying the selected function to the signal samples to determine the metric values. 2. The method claim 1, wherein the branch metric function is selected from a set of signal-dependent branch metric functions. '180 Patent col. 1511. 39-51. Given same, Dr.

McLaughlin divided these claims into five “elements” and walked the jury through how each was mapped on to the accused MNP circuits. He showed how some of the language of the '180 Patent was the same as the '839 Patent and explained that there was no need to go through the circuit schematics again since the methods were being applied to the same circuit. (Docket No. 677 at 120-125).

He “checked off” that he had already demonstrated that the MNP technology involved: (1) “a method of determining branch metric values in a detector, comprising”; (2) “selecting a branch metric function at a certain time index;” and (3) “applying the selected function to the signal samples to determine the metric values;” and (4) “the method claim 1, wherein the branch metric function is selected from a set of signal-dependent branch metric functions.” (Id.). On the remaining element, Dr.

McLaughlin explained how the MNP technology used a method of “receiving a plurality of time variant signal samples, the signal samples having one of signal-dependent noise, correlated noise, and both signal dependent and correlated noise associated therewith,” which was consistent with the second element of Claim 1 of the '180 Patent. (Docket No. 677 at 122). Based on his knowledge and expertise in the field, as well the Marvell DSP Technical Presentation slides titled “Data Dependent Noise” and “Noise Correlation,” he showed that the MNP technology used this last element. (PI. Ex. 770).

He also noted that there was no requirement in the '180 Patent that the detector be a Viterbi-like detector. (Docket No. 677 at 121). In light of his prior conclusion that each of the elements of Claim 2 of the '180 Patent, through which the elements of Claim 1 are incorporated, was used in the MNP technology, he concluded that the MNP infringed Claim 2 of the '180 Patent. (Id. at 123). b.

NLD Chips Marvell next argues that no reasonable jury could find that Accused NLD Chips infringe because, again: (1) the chips do not select a branch metric function for each of the branches of the trellis at a certain time index; (2) the chips do not apply each of said branch metric functions to a plurality of signal samples; and (3) the NLDs do not determine branch metric values for branches of a trellis. (Docket No. 743). As with the MNP and EMNP chips, Dr.

McLaughlin started with the '839 Patent and mapped the first element, “a method of determining branch metric values for branches of a trellis for a Viterbi-like detector” onto the NLD circuits from Exhibit D of the Chip Stip. (Docket No. 677 at 140-142). Dr. McLaughlin broke this down further, and showed how the NLD: (1) is a detector; (2) computes branch metric values for branches of a trellis; and (3) is a Viterbi-like detector. (Id. at 140-143). Dr.

McLaughlin based his conclusions on his analysis, using his personal knowledge and expertise in this area, and he also explicitly relied on Marvell documents, such as the presentation titled “Nonlinear Viterbi Detector Application Note — C8830R1.0” by Dr. Hongxin Song. (PI. Ex. 596). In light of the Chip Stip, Dr. McLaughlin opined that the NLD technology contained a method for “selecting a branch metric function for each of the branches at a certain time index from a set of signal-dependent branch metric functions.” (Docket No. 677 at 142-144).

He demonstrated this on the circuit drawing and referenced supporting testimony from Marvell engineer Mr. Burd. (Id. at 145-146). Last, for the '839 Patent, Dr. McLaughlin relied on Exhibit D of the Chip Stip to conclude that the NLD used a method “applying each of said selected functions to a plurality of signal samples to determine the metric value corresponding to the branch for which the applied branch metric was selected wherein each sample corresponds to a different sampling time instant.” (Docket No. 677 at 146-149).

He highlighted the location on the circuit diagrams of the plurality of signal samples, the application of the branch metric function, and the branch metric value. (Id.). CMU also offered deposition testimony from Mr. Burd, who stated that “[a]nd so in fact noise whitening filter is a parameter of the branch metric function,” as well as Dr. Song’s Application Note, to further support Dr. McLaughlin’s conclusion. (Id. at 149; PI. Ex. 596). Given that Dr.

McLaughlin demonstrated how each of the elements of Claim 4 of the '839 Patent was found to be used in the NLD technology, he opined that the NLD infringed Claim 4 of the '839 Patent. (Docket No. 677 at 149-150). Moving on to Claim 2 of the '180 Patent, Dr.

McLaughlin stated that his analysis showed that the NLD technology involved: (1) a method of determining branch metric values in a detector; (2) a detector selecting branch metric functions; (3) application of the selected function; (4) wherein the branch metric function is selected from the set of signal dependent branch functions; and finally (5) that it received a plurality of time variant signal samples, those having one of signal dependent noise and correlated noise. (Docket No. 677 at 154-156). He did not walk through the circuits again as he had already demonstrated his analysis in those respects.

Because Dr. McLaughlin gave expert opinion testimony that each of the elements of Claim 2 of the '180 Patent, through which the elements of Claim 1 are incorporated, used the NLD technology, he concluded that the NLD infringed Claim 2 of the '180 Patent. (Id.). c. Simulators Marvell also contends that the Accused Simulators do not infringe as a matter of law. (Docket No. 743 at 5-7). It alleges that the '839 Patent is directed to a method of determining branch metric values for branches of a trellis for a Viterbi-like “detector.” (Id. at 5) (emphasis in original).

Similarly, it claims that the '180 Patent is directed to a “method of determining branch metric values in a detector(Id. at 5-6) (emphasis in original). Because Marvell’s simulators are not detectors and do not process actual signal samples, they are not implicated by either patent. (Id. at 6-7). In support, Marvell cites to Harris Corp. v. Ericsson Inc. for the proposition that “as a matter of law, running a simulator program does not constitute actually performing the claimed methods in a detector for processing signal samples.” 417 F.3d 1241, 1256 (Fed.Cir.2005).

The Harris case involved a method of using a communication system which could “be directly infringed only by one who uses the system, not by one who makes or sells the components of the system.” Id. at 1256. In Harris, the Federal Circuit focused on the fact that the plaintiff had submitted only a flow chart describing a “simulation program” and had not shown that the “claimed method is actually carried out, rather than simulated, when Ericsson runs this program.” Id. (emphasis added). Accordingly, the plaintiff had failed to present evidence of the method being carried out by the program. Id.

Nevertheless, the Federal Circuit did not create a bright line rule about simulation programs. As described herein, CMU has presented enough evidence to persuade the trier of fact that the claimed method in this case is actually carried out on the challenged simulators. Upon examination of the Accused Simulators, Dr. McLaughlin opined at trial “that Marvell’s simulators infringed the asserted claims.” (Docket No. 677 at 83).

In reaching this conclusion, he studied the code of five simulators provided by Marvell in discovery: (1) the KavcicPP Simulator; (2) the MNP Simulator; (3) the EMNP Simulator; (4) the NLD Simulator; and (5) the KaveicViterbi Simulator. (Id. at 156-166). The first four simulators correspond to particular chips and the last is used as a benchmark. (Id. at 169). Dr.

McLaughlin described the simulators’ use by Marvell: (1) to research and develop the chips; (2) to verify the chips’ hardware design; and (3) to provide simulation code to customers so that they may, in turn, evaluate the performance and functionality of Marvell’s chips. (Id. at 158). This testimony was supported by the deposition testimony of the Marvell corporate designee on this technology, Greg Burd, in which he stated “C Code which is used in our simulation.

So we provide a version of the same C code to the designers to be used to serve as a golden source of test vectors to verify the design against.” (Id. at 169-170). Dr. McLaughlin analyzed the simulator code (PI. Ex. 106), and he compared lines of code to the circuits of the chip. (Docket No. 677 at 156-166). He demonstrated how the simulators mimic the chips and how the different elements of the claims were covered by the code. (Id.). Further, Dr. McLaughlin determined that there was a one-to-one match between what was in the computer code and what was being calculated in the circuitry. (Id.). Dr.

McLaughlin showed this for each of the first four simulators. (Id.). Based on his conclusion that the chips infringed each and every claim, he determined the simulators also infringed. (Id.). For the KaveicViterbi Simulator, Dr. McLaughlin similarly analyzed the code in question. (Id. at 166-173). In doing so, Dr. McLaughlin looked at testimony by Marvell on how it uses this particular simulator. For example, Mr.

Doan, then a Marvell engineer, stated that the KavcicViterbi Simulator was used as the benchmark and that they “continuously run Kavcic algorithm to benchmark any subsequent algorithm we develop at Marvell.” (Id. at 171-172; Docket No. 761 at Jt. Ex. D at 137-138). Some of his analysis was also based on the testimony of Marvell engineer Mr. Burd, who stated in relevant part: The way I do my research which might be different from other people, I first try to understand what’s available out there. So maybe look at some ideas which people came up with before me....

To make sure that I do, in fact, understand what Professor Kavcic is trying to do and at the same time just to see kind of what’s out there. Right? And then I can use this code for benchmarking later. Right? For performance benchmarking later. So this was a launching pad for our research. (Docket No. 677 at 170-173). This (Plaintiff Ex. 93) is KavcicViterbi.cpp class, written by engineers in Marvell, and I do believe it contains the implementation, as understood by our architecture team of the IP which is taught in Professor Kavcic’s papers, and consequently in his patent. (Docket No. 677 at 170-171; PL Ex. 93).

Although Marvell insists that its simulators do not process actual signal samples, Dr. McLaughlin maintained that the simulators do process both synthetic as well as real samples from a hard drive or hard drive disk. (Docket No. 677 at 174-178). To this end, he referred to two Marvell documents to support his conclusion, the first being an “Analysis and Design of Yiterbi Detector” which showed test results for a Toshiba wave form.

The second document he used was an email from Marvell’s Mike Madden to Hitachi, showing test results for wave form captured from the spin stand that were re-sampled, scaled, and fed into Marvell’s simulators. (Pl. Exs. 527; 341). In sum, Dr. McLaughlin opined that: (1) the computer code directly mimics the chip; (2) the chip infringes each and every element of the claims of the CMU patent; and (3) the computer code and the simulator also infringe each and every element of the claims.

While Marvell vigorously disagreed with all these opinions, both parties had an opportunity to make their opposing arguments on the nature of simulators to the jury. See, e.g., (Docket No. 759 at 64-65) (“And [CMU] say yeah, simulated data; simulations, simulator. They know full well that when you sit down at a computer and you put in code and you simulate a formula, that’s not the detector. You’re not infringing anybody’s work when you do that. Everybody does it. They did it.”).

The determination was purely factual and one which the jury alone would have to decide by weighing the offered evidence and the credibility of witnesses who testified to same. Walker v. Gordon, 46 Fed.Appx. 691, 695 (3d Cir.2002). d. Direct Infringement in Sum CMU presented sufficient evidence, through Dr. McLaughlin as well as supporting Marvell documents, that the MNP, EMNP, and NLD chip technology and the Accused Simulator technology use a method that includes each and every method step of Claim 4 of the '839 Patent and Claim 2 of the '180 Patent. Whether Dr.

McLaughlin’s conclusions were persuasive or otherwise worthy of credence goes to the core of his credibility as a witness, and such “[d]eterminations regarding the weight to be accorded, and the sufficiency of, the evidence relied upon by the proffered expert are within the sole province of the jury.” Walker, 46 Fed.Appx. at 695. Indeed, Marvell took advantage of the opportunity to rebut Dr. McLaughlin’s conclusions by offering opinion evidence through its own non-infringement expert Dr. Blahut that its products did not infringe CMU’s patents. (Docket No. 711 at 204-308).

The jury was free to accept either expert’s opinions or reject them, as the “credibility of the parties’ competing experts is an issue for the jury to resolve, not the Court.” Miller ex rel. Miller v. Evenflo Co., Inc., Civ. No. 09-108, 2011 WL 7037127, at *3 n. 3 (W.D.Pa. Dec. 15, 2011); see also Walker, 46 Fed.Appx. at 695.

Considering the evidence in the light most favorable to the nonmoving party, and giving it the advantage of every fair and reasonable inference in light of the undisputed facts, the Court finds there was adequate evidence upon which a reasonable jury could properly find a verdict in favor of CMU. Therefore, Marvell’s motion for JMOL and motion for a new trial on this issue is denied. The question of whether there was direct infringement by the Accused Chips and Accused Simulators was properly submitted to and decided by the jury. 3. Indirect Infringement a.

Inducement Marvell asserts that it was entitled to a judgment as a matter of law or a new trial on CMU’s claims of indirect infringement for actively inducing infringement. (Docket No. 743, 805). It argues that CMU has not proven: (1) direct infringement; (2) that Marvell had actual knowledge of the patents-in-suit; and (3) that it specifically intended for others to perform acts that directly infringe one or more of the asserted claims. (Id.). First, the Chip Stip lists the models of chips that correspond to the agreed-upon circuits that Dr. McLaughlin analyzed. (PI. Ex. 823).

These chips are sold to Marvell’s customers to be used in hard disk drives, and they are also used by Marvell in research and development phases. (Docket No. 677 at 103-104, 178). As part of his testimony, Dr. McLaughlin analyzed the hardware and firmware settings of Marvell’s customers such as Western Digital, Samsung, and Toshiba. (PI. Exs. 1914; 1915; 1918; 1919). Once again, Dr. McLaughlin put forth expert testimony to prove direct infringement of Marvell’s chips, thereby allowing CMU to advance a theory of induced infringement.

Second, in order to achieve success on induced infringement, CMU was required to show that Marvell had “knowledge of the existence of the patent” or took “deliberate actions to avoid confirming a high probability of wrongdoing.” Global-Tech, 131 S.Ct. at 2068-70. To that end, CMU presented significant amounts of evidence to show Marvell’s knowledge of the patents. Much of this evidence also supported CMU’s claim of willful infringement, which is addressed later herein. CMU proffered the following evidence of Marvell’s knowledge of the patents-in-suit. First, is the email from Dr. Kavcic sent to Dr.

Nazari at Marvell in 1998, providing a link to his publications, resume, and recent work. (Def. Ex. 1023). Second, is the January 3, 2002 email from Greg Burd to Toai Doan and Nersi Nazari, attaching a write up on the KavcicPP and stating “1. Kavcic’s detection scheme is patented (assignee: Carnegie Mellon Univ. 2001).” (PL Ex. 280). Third, there is the January 4, 2002 weekly status report email from Mr. Burd to Mr. Doan, Dr. Nazari, and Ke Han, describing his work and stating “[a]nd of course as I mentioned earlier, Kavcic detector is also patented.” (PI. Ex. 283).

Such evidence indicated that Marvell knew of both the patents and the high likelihood that the Accused Technology infringed, especially given that the very people who designed the Accused Technology, i.e., the engineers, knew of the patents. (Id.). In fact, Marvell hardly argued that it did not know of the CMU Patents; its theme throughout trial was that Dr. Kavcic’s work was its launching point only, and its technology was a “sub-optimal” version of Kavcic’s algorithm. (Docket No. 756).

CMU replied that the MNP/NLD used the Kavcic algorithm, and optimality had no bearing on whether the claims of the Kavcic patented method were infringed. (Id.) In support, CMU proffered one of Dr. Wu’s weekly emails from January 10, 2003 that stated “1. MNP enhancement: Greg and I discussed the approach of using a different noise whitening filter for each branch. It turns out to be the original structure that Kavcic proposed in his paper.” (PL Ex. 366).

Thus, even if, as Marvell insists, it did not know the accused technologies infringed the patents-in-suit, CMU put forth sufficient evidence that a jury could find that Marvell was willfully blind in light of the high probability of infringement. Global-Tech, 131 S.Ct. at 2072. Once presented with the patents, Mr. Doan testified that he never looked at the patents, never directed others to look at them, and never contacted Marvell’s legal department about them. (Docket No. 761 Jt. Ex. D at 125, 130).

Despite same, he reported to Nersi Nazari on his January 14, 2002 status report that they would “continue work on non-linear detector based upon Kavcic’s model.” (Pl. Ex. 285). Additionally, as to Marvell’s knowledge of the patents, CMU’s Technology Transfer representative Carl Mahler sent a letter in August 2003 to Marvell’s CTO Dr. Pantas Sutardja and Matthew Gloss, who was Marvell’s then General Counsel, stating that CMU held patents in the area of correlation-sensitive adaptive sequence detectors, “namely U.S. Patent number 6,201,839 B1 and U.S. Patent number 6,438,180 Bl.” (Pl. Ex. 422, PL Ex. 431). Mr.

Mahler attached the patents to the letter and encouraged Marvell to contact him if they found the patents to be of interest. (Id.). There was no response to these letters. (Docket No. 682 at 150). Similarly, Junya Suwanai of Fujitsu, “a customer for Marvell’s read channel i.e. 5575M, 7500M,” corresponded with Marvell in November 2004, stating that Fujitsu had received a license offer for the CMU Patents-in-suit. (PL Ex. 477).

He wrote that “since it seems that these patents might be related to read channel, we would like to know, by the end of November, your opinion regarding relationship between CMU’s Patents and the above Marvell lead [sic] channel and the specific grounds/reasons for such opinion.” (Id.). No documents were found in relation to this letter, and Marvell’s corporate designee testified that he did not know of any response to this letter. (Docket No. 761 at Jt. Ex. C. at 534-535).

CMU had to prove that Marvell had the specific intent to encourage another’s infringement, which can be demonstrated if Marvell caused, urged, encouraged, or aided the infringing conduct. DSU, 471 F.3d at 1306; Akamai Techs., 692 F.3d at 1308. On this point, CMU showed that Marvell aided its customers’ infringement by producing chips that used the accused methods and instructed its customers to use the chips in infringing modes. (Docket No. 677 at 180-183; Docket No. 678 at 91; PI. Exs. 1913; 1918; 1919).

Specifically, CMU entered into evidence emails, firmware, as well as programming instructions for hardware showing that Marvell directed its customers, including Western Digital, Samsung, and Toshiba, to use the chips in infringing modes. (PI. Exs. 730; 932; 1914; 1915; 1918; 1919). Likewise, Dr. McLaughlin testified about Marvell field application engineers who are deployed to Marvell’s customers to assist them in putting Marvell’s chips into their products and instruct them on how to use the chips. (Docket No. 677 at 178-179).

To that end, Teik Ee Yeo, Western Digital’s corporate designee, testified that the chips it buys from Marvell have the technology enabled, i.e., set to “infringing modes,” and that Western Digital tended to follow the suggestions of Marvell engineers regarding these settings on the chips. (Docket No. 761 at Jt. Ex. B 146). This knowledge of customer use was buttressed by the testimony of CMU industry expert, Dr. Bajorek, who opined that the Accused Technology became industry standard. (Docket No. 678 at 108-112).

Accepting the evidence in the light most favorable to CMU, the nonmoving party, and giving it the advantage of every fair and reasonable inference based on the facts of record, there was sufficient evidence presented at trial to support CMU’s theory that Marvell created technology that employed the accused methods with either actual knowledge or willful blindness to the possibility that its devices infringed the patents-in-suit, and that Marvell sold same to its customers.

CMU produced sufficient evidence that Marvell actively induced at least one of Marvell’s customers to use a method that is covered by Claim 4 of the '839 Patent or Claim 2 of the '180 Patent. Thus, Defendants’ JMOL on this issue was denied, and the question of whether there was induced infringement properly proceeded to the jury. Considering the record as a whole, the jury’s verdict of induced infringement is not against the weight of evidence, and the motion for a new trial on this issue is denied. b.

Contributory Infringement Marvell argues that it was entitled to either a judgment as a matter of law or a new trial on CMU’s claims of contributory infringement for actively contributing to infringement. (Docket Nos. 743; 805). It argues that CMU has not proven: (1) direct infringement; (2) that Marve

[…]