AMENDED FINDINGS OF FACT AND CONCLUSIONS OF LAW
JED S. RAKOFF, U.S.D.J.The present dispute arises out of the ill-fated decision by defendants Barnes and Noble, Inc., BarnesandNoble.com LLC, and Nook Media LLC (collectively, “B <& N”) to enter the market for electronic readers (“eReaders”). In 2009, B <& N opened its electronic “bookstore” and launched a device for reading electronic content (“eContent”) known as the “Nook.” The Nook ultimately proved less than profitable, but not before plaintiff Adrea LLC (“Adrea”) notified defendants that several features on the Nook allegedly infringed Adrea’s patented intellectual property. After negotiations between the parties broke down, Adrea filed suit in 2013 accusing B & N of infringing three U.S. patents: U.S. Patent No. 7,620,703 (“the ’703 Patent”); U.S. Patent No. 7,299,501 (“the ’501 Patent”); and U.S. Patent No. 7,298,851 (“the ’851 Patent”). The ‘following year, a jury found B & N liable for infringing the ’703 and ’501 Patents, but determined that such infringement was not willful. The jury awarded Adrea $1.33 million in damages.
Subsequently, however, on July 23, 2015, the Court granted B & N’s post-trial motion, holding as a matter of law that the ’501 Patent was invalid. Accordingly, the Court ordered a limited new trial to determine damages on. the ’703 Patent. Then, ten days before trial, the Supreme Court in Halo Electronics, Inc. v. Pulse Electronics, Inc., — U.S. -, 136 S.Ct. 1923, 195 L.Ed.2d 278 (2016) reduced the standard for finding willful infringement, prompting plaintiff to also move for a new determination on willfulness. The parties agreed that their second trial, now expanded to include both these issues, should be a bench trial, and, accordingly, the Court conducted a three-day trial.
On the basis of the Court’s assessment of the evidence presented at that trial, includihg its assessment of the witnesses’ demeanor and credibility, the Court now awards damages of $266,832.82 for what the Court finds was B & N’s non-willful infringement of Adrea’s ’703 Patent. This Opinion and Order sets forth the Court’s findings of fact and conclusions of law, and directs the parties to supply the information necessary for the entry of final judgment.
B & N launched the Nook in or around November 2009. 2016 Trial Transcript (“Trial Tr.”) (Yurkerwich) at 39:3-8. *306Among other features, the device allowed users to purchase eContent from B & N without using a web browser. Adrea, LLC v. Barnes & Noble, Inc., No. 13 CIV. 4137(JSR), 2016 WL 859685, at *2 (S.D.N.Y. Feb. 24, 2016). Instead, users clicked a “Shop” button on the devices, which took them directly to B & N’s electronic bookstore. Id. Users then could make purchases without needing to type in the URL, BarnesandNoble.com. See id.
Around the same time, however, on November 17, 2009, the Patent Trademark Office issued the ’703 Patent, which had been previously pending, to Yevgeniy Eugene Shteyn. JTX-003 at 2. In relevant part, the patent allows users of a consumer device to retrieve content relating to the usage of the device from the Internet at the press of a button, without having to browse the Internet or search for the desired content. ADREA, LLC v. Barnes & Noble, Inc., No. 13 CIV. 4137 JSR, 2013 WL 6334489, at *1 (S.D.N.Y. Dec. 2, 2013). Shteyn subsequently assigned all rights and interests in the ’703 Patent to Konink-lijke Philips Electronics N.V. (“Philips”). 2016 Pretrial Consent Order, Dkt. No. 217 at 5. In August 2010, in turn, Philips and three other companies formed the joint venture that is Adrea: Philips; Intertrust Technologies Corporation (“Intertrust”); Discovery Communications, Inc. (“Discovery”); and Sony Corporation of America (“Sony”). Trial Tr. (Yurkerwich) at 46:5-25. The companies purchased equal shares in the joint venture by contributing equal value—Discovery, Sony, and Philips assigned certain patents, while intertrust contributed services. Id.
Adrea was responsible for licensing its patent portfolio, id. at 85:22-86:9, which it did once pursuant to a settlement of a patent dispute originally between Discovery and Amazon (the “Amazon Agreement”). See id. (Barnes) at 296:7-14; id. (Yurkerwich) at 261:4-6. Prior to Adrea’s formation, Discovery sued Amazon alleging that Amazon’s eReader, the “Kindle,” infringed Discovery’s ’501 and ’851 Patents. 2016 Trial Tr. (Yurkerwich) at 52:9-11. Amazon cross-claimed alleging that Discovery had infringed six of its own patents. Id. at 52:12-14.
Adrea took over the litigation in August 2010 after receiving Discovery’s patents. 2014 Trial Tr. (Shamoon) at 94:19-95:10, 135:10-15, 135:19-23. Although Philips assigned the ’703 Patent around the same time, Adrea did not add the patent to the litigation. 2016 Trial Tr. (Yurkerwich) at 217:16-218:12; id. (Barnes) at 322:7-14, 357:17-358:1.
In November 2011, Adrea, Amazon, and Discovery signed the Amazon Agreement settling the suit. JTX-032 at 1-2. The agreement licensed and cross-licensed several portfolios of patents, including:
a. A license from Adrea to Amazon for Adrea’s Discovery patents;
b. A cross-license from Amazon to Discovery for the six patents in Amazon’s cross-claims; and
c. An option for Amazon to license Adrea’s Sony and Philips portfolios (including the ’703 Patent) for $2,5 million. Amazon had up to two years to exercise the option, during which time Adrea agreed not to sue for alleged infringement. Adrea, however, could retroactively sue if Amazon did not exercise the option.
JTX-032 §§ 3-6. Pursuant to the agreement, Amazon paid $10 million upfront to settle the case in general, and later paid another $2.5 million to exercise the option.1 *3072016 Trial Tr. (Barnes) at 300:1-16, 301:22-302:9, 302:15-303:6; 307:22-308:14; JTX-037.
The following year, in March 2012, Adrea sent B & N “sample claim illustrations” for the patents it believed B & N had infringed. Ederer Decl. Ex. A, ECF No. 245, at 1. The illustrations listed six patents including the ’501, ’851, and ’703 Patents. Id. Several months later, in May 2012, Adrea offered B & N a license to Adrea’s patent portfolio for $0.50 per Nook, to which B & N never responded. 2016 Trial Tr. (Yurkerwich) at 45:1-5, 123:5-6, 256:13-15, 339:22-340:9; id (Barnes) at 386:9-17.
On June 14, 2013, as previously noted, Adrea filed a complaint in this Court accusing B & N of patent infringement. On October 27, 2014, following trial, a jury found B & N liable for infringing the ’703 and ’501 Patents, but determined that such infringement was not willful. The jury awarded $1.33 million in damages. On July 23, 2015, the Court held that the ’501 Patent was invalid and ordered a limited new trial to determine damages. The Court held a bench trial on June 23-24, 2016 and July 15, 2016, during which experts David Yurkerwich and Ned Barnes testified on behalf of Adrea and B & N, respectively.2 At the conclusion, of trial, the parties submitted supplemental briefing on whether the Court should award enhanced damages pursuant to the Supreme Court’s decision in Halo, 136 S.Ct. at 1923.
Since the jury at the first trial already determined that B & N had infringed the ’703 Patent, the first question the Court has to decide on the basis of the evidence presented at the second bench trial is what is a reasonable royalty rate for the ’703 Patent. “The reasonable royalty theory of damages .,. seeks to compensate the patentee ... for its lost opportunity to obtain a reasonable royalty that the infringer would have been willing to pay if it had been barred from infringing.” AstraZeneca AB v. Apotex Corp., 782 F.3d 1324, 1334 (Fed. Cir. 2015) (citing Lucent Techs., Inc. v. Gateway, Inc., 580 F.3d 1301, 1325 (Fed. Cir. 2009)). “One approach to calculating a reasonable royalty ... posits a hypothetical negotiation between a willing licensor and a willing licensee to ascertain the royalty upon which the parties would have agreed had they successfully negotiated an agreement just before infringement began.” Carnegie Mellon Univ. v. Marvell Tech. Grp., Ltd., 807 F.3d 1283, 1303-4 (Fed. Cir. 2015). Adrea, as the present patent holder, bears the burden of proving the amount of damages by a preponderance of the credible evidence. See In re Electro-Mech. Indus., 359 Fed.Appx. 160, 165 (Fed. Cir. 2009).
The hypothetical negotiation here would have taken place in November 2009, when the ’703 Patent issued and the infringement began. See Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d, 1292, 1312 (Fed. Cir. 2011). The parties to the negotiation also would have been B & N and Philips, because Adrea had not yet been formed.
While in the past courts have generally focused on the 15 factors set forth in Georgia-Pacific Corp. v. U.S. Plywood Corp., 318 F.Supp. 1116, 1120 (S.D.N.Y. 1970) to guide their analyses in such instances, this Court, aided by the Federal Circuit Bar Association’s recent Model Patent Jury Instructions and by the particular facts of this case, focuses in particular on two factors: (a) the value that the claimed invention contributes to the accused product; and (b) comparable license agreements such as those covering the use *308of the claimed invention. Here, these factors are one and the same because the Amazon Agreement is both the best indicator of the value of the ’703 Patent and the only comparable license agreement.3
Superficially, one might say that Amazon paid only the $2.5 million required to exercise the option licensing the ’703 Patent. 2016 Trial Tr. (Barnes) at 444:19-22. But the economics of the Amazon Agreement do not support this conclusion. One does not acquire an option for free, and the value of the option is therefore greater than the option’s exercise price. See id. (Yurkerwich) at 284. To this end, Section 3 of the Amazon Agreement states that Amazon agreed to pay $10 million upfront in exchange for “the releases, covenants, licenses, and other rights provided hereunder.” See JTX-032 § 3 (emphasis added). But the contested provision setting the $10 million sum does not specify to which lights it relates.
The Court therefore looks beyond the four corners of the agreement for guidance. Specifically, the Court adopts Adrea’s recommendation to calculate the total consideration paid by Amazon pursuant to the agreement and use external metrics to determine the amount attributable to the ’703 Patent. Tallying this consideration is simple for the cash payments: $10 million and $2.5 million. Adding the value of Amazon’s cross-licenses to Discovery poses a more difficult inquiry. While B & N’s expert, Barnes, did not conduct any valuation of the cross-licenses, Adrea’s expert, Yurkerwich, appraised them at $10 million. See 2016 Trial Tr. (Yurkerwich) at 194:5-17. However, the economics of the Amazon Agreement do not support Yurk-erwich’s estimate.4 Adrea is the owner of the patents licensed to Amazon, not Discovery, and any value given directly to Discovery would come at the expense of Adrea’s joint partners. It is impossible to believe that Discovery’s partners would stand idly by while Discovery pocketed $10 million in value for licenses to the joint venture’s patents.5
The more likely explanation is that the cross-licenses were worth quite little. If the parties valued the patents at only a few hundred thousand dollars, the joint partners may have allowed Discovery to pocket its savings as opposed to squab*309bling over relatively minor distributions.6 Contemporaneous statements by Adrea’s counsel support this conclusion, in particular, when Adrea’s attorneys stated to Amazon during settlement negotiations that Amazon’s potential recovery on its cross-claims was “at most ... a few hundred thousand dollars.”7 DTX-287 at 1. Of course, Adrea’s attorneys likely downplayed the value of Amazon’s patents as a settlement tactic, and so the Court assigns a higher valuation of $500,000.8 Accordingly, the Court finds that Amazon paid $13 million for the licenses in the Amazon Agreement: $10 million in cash up front, $2.5 million in cash to exercise the option, and $500,000 by way of the cross-licenses to Discovery.
The next issue is how to allot this $13 million. The experts once again disagree with one another, but the Court finds it reasonable to initially apportion this amount based on which patents have been litigated. Since litigation entails both costs and risks, Discovery and Adrea would likely only assert their most valuable intellectual property. 2016 Trial Tr. (Yurkerwich) at 116:17-24. The Court thus assigns equals hares of the $13 million to the ’501,-’703, and ’851 Patents (25% per patent), and an additional 25% share to the remaining patents in Adrea’s portfolio. Id. at 117:2-5. This last share adds an element of conservatism and accounts for the possibility that Adrea would assert an additional patent in future litigation. Id. at 117:6-12.
The Court is cognizant of the limitations of this allocation. Since Adrea had yet to litigate the ’703 Patent at the time of the Amazon Agreement, Amazon would not have known for certain that the ’703 Patent presented a litigation risk. Amazon, however, is a sophisticated actor that likely would have conducted due diligence on Adrea’s portfolio and likely would have identified the ’703 Patent. Furthermore, while documents prepared by Adrea indicate that several other patents in Adrea’s portfolio are valuable, see id. (Barnes) at 378:20-379:3; 379:18-380:10, these patents have never been asserted in litigation and their value thus appears to be limited.
The Court cannot end its inquiry, however, with which patents have been litigated. Patents in the same suit often have different values based on a range of factors including, for example, the patents’ lifespans. If a patent expires relatively soon, all things being equal it will generate less revenue than a patent with a longer expiration. Along these lines, the lifespan of the ’703 Patent is almost twice as long, *310on average, as that of its sister patents. Id. (Yurkerwich) at 120:16-121:6.
While it is therefore plausible that the ’703 Patent is worth more than its fellow patents, id. at 119:23-120:14, the Court finds that it is only worth slightly more. There is no evidence that the ’703 Patent was expected to generate annual revenue comparable to that of its sister patents, and the Court is not persuaded that the patented technology was all that significant. The patented feature allowed customers to use a “Shop” application to purchase products from a seller. Users, however, could still purchase eContent by simply typing the seller’s URL into a web browser, and neither Adrea nor Yurker-wich explain why customers would not use the latter method.9 For this reason, the Court adjusts the share value of the ’703 Patent from 25% to 39%, which amounts to $5,070,000 or a per unit royalty rate of $0,039.
The last task, with respect to royalties, is to account for the differing circumstances of the hypothetical negotiation— namely, what the experts call “litigation risk.” At the time of the Amazon Agreement, neither Amazon nor Adrea knew whether the patents at issue were valid or actually infringed. Given this unpredictability, Adrea licensed its patents at a discount. The hypothetical negotiation, on the other hand, would have involved no such uncertainty because the parties must assume that the ’703 Patent was valid and infringed. See Aqua Shield v. Inter Pool Cover Team, 774 F.3d 766, 771 (Fed. Cir. 2014). B & N therefore would not have received Amazon’s discount.
The experts disagree on how much more B & N would have paid. Yurkerwich proposes a multiple of 2.38x,10 and relies primarily on an internal business forecast prepared by Intertrust prior to Adrea’s formation (the “Intertrust Pitch”).11 See id. (Yurkerwich) at 227:18-228:12. Intertrust apparently prepared the pitch to convince Sony and Discovery to join the Adrea joint venture, and the document estimated the royalties that Adrea hoped to receive for licensing a portfolio including the Discovery and Sony patents.
The problem is that the Intertrust Pitch does not mention Philips or the ’703 Patent. While the pitch states that additional patents may be added to the Adrea portfolio, there is no evidence that Philips intended to participate or to contribute the ’703 Patent. Id. at 231:3-15. Indeed, if Philips did intend to assign valuable patents to Adrea’s portfolio, it is odd that the company would leave such significant information out of a document designed to persuade potential investors. The Inter-trust Pitch’s estimated royalties therefore shed no light on Amazon’s litigation risk discount.12
*311Fortunately, the correct explanation is often the simplest one—and appears here on slide 18 of the Intertrust Pitch. See PTX-166, at ADREA0141127. The slide states that the pitch’s estimates assume a 33% discount rate, and explains under the heading, “Primary Execution Risks,” that “[pjatents in litigation suffer in court and impact ability to scale licensing programs” and “litigation extends through full term causing patents to age out.” The heading also states that “licensing activity remains primarily restricted to US market” and “eBook market remains small rather than transition to mass market,” meaning that not all of the 33% discount rate is attributable to litigation. If the Intertrust Pitch has any probative value, it is to establish a maximum litigation discount rate of approximately 32%. The Court adopts this percentage and uses a litigation risk multiplier of 1.47x.
Based on the foregoing analysis, the Court finds that the hypothetical negotiation between Philips and B & N in 2009 would have resulted in a royalty rate of $0,051 per unit.13 Given B & N’s total sales of 5,232,016 eReaders, see 2016 Trial Tr. (Yurkerwich) at 217:6, this amounts to damages of $266,832.82.14
The Court next turns to whether the award of damages should be enhanced based on B & N’s allegedly willful misconduct. Section 284 of the Patent Act permits the Court to “increase the damages up to three times the amount found or assessed.” 35 U.S.C. § 284. For the Court to have jurisdiction to do so, however, the finder of fact must first determine that the infringement was willful. In re Seagate Technology, LLC, 497 F.3d 1360 (2007). Pursuant to Seagate, the Court charged the jury in the first trial that in order to find willful infringement, the “plaintiff must prove, by clear and convincing evidence, that defendants either actually knew they were infringing plaintiffs patents or recklessly disregarded the fact that their actions constituted an unjustifiably high risk of infringement of that patent.” The jury found that Adrea had not met its burden.
The Supreme Court subsequently overturned the Seagate test in Halo, 136 S.Ct. at 1932. The Halo Court held that district courts must have greater discretion in awarding enhanced damages in cases where the defendant’s infringement was egregious, “typified by willful misconduct.” Id. at 1934. The Supreme Court accordingly rejected Seagate’s strict re*312quirement that a patentee prove that the infringer’s defenses were objectively unreasonable, and lowered the evidentiary standard for willfulness from clear and convincing evidence to the preponderance of the evidence,15 Id. The Halo Court did not otherwise modify the requirement that the infringement have been “subjectively willful” or alter the Seagate factors for determining when this is the case. Id.; WesternGeco L.L.C. v. ION Geophysical Corp., 837 F.3d 1358, 1362 (Fed. Cir. 2016). The Court accordingly revisits whether B & N’s infringement was willful under the lower evidentiary standard set forth in Halo, and finds that it was not.
Pursuant to Seagate, the judge or jury considers all relevant facts, including but not limited to: (i) whether or not defendants intentionally copied a product of plaintiff; (ii) whether or not defendants acted in accordance with the standards of commerce for its industry; (iii) whether or not defendants made a good faith effort to avoid infringing the asserted claims; (iv) whether or not there is a reasonable basis to believe that defendants did not infringe or had a reasonable defense to infringement; and (v) whether or not defendants tried to cover up any infringement. See WesternGeco, 837 F.3d at 1362; DePuy Spine, Inc. v. Medtronic Sofamor Danek, Inc., 567 F.3d 1314, 1336 (Fed. Cir. 2009). Adrea’s argument is notable for the factors it omits. There is no evidence that B & N deliberately copied the ’703 Patent; indeed, B & N launched the Nook shortly before the ’703 Patent issued in November 2009. Adrea similarly does not assert that B & N acted in contravention of any relevant “standards of commerce,” or even define what those standards are. Nor does Adrea allege that B & N sought to cover up or conceal its infringement.
Instead, Adrea points out that B & N did not attempt to design around or avoid infringement of the ’703 patent. While true, this is consistent with B & N’s defense that the Nook did not infringe the ’703 Patent.16 Adrea, in turn, has not shown that B & N lacked a reasonable basis for its position. While Adrea argues that there is “no evidence” that B & N conducted a meaningful investigation relating to the ’703 patent, it is not B & N’s burden to show such evidence. See WBIP, LLC v. Kohler Co., 829 F.3d 1317, 1339 (Fed. Cir. 2016). Similarly, while it is true that the Court rejected most of B & N’s proposed claim constructions during the parties’ “Markman” hearing, see ADREA, 2013 WL 6334489, at *4, this has no bearing on whether B & N subjectively believed its defenses. See Halo, 136 S.Ct. at 1933 (“The subjective willfulness of a patent infringer, intentional or knowing, may warrant enhanced damages, without regard to whether his infringement was objectively reckless.”). This is particularly the case in the context of claim construction, which “frequently poses difficult questions over which reasonable minds may disagree.” Haemonetics Corp. v. Baxter Healthcare Corp., 607 F.3d 776, 783 (Fed. Cir. 2010). Adrea presents no evidence that B & N was aware of the prob*313lems with its claim constructions, that B & N did not believe in the validity of its proposed constructions, or that B & N did not hold this view prior to the onset of litigation. Adrea therefore has not shown by a preponderance of the evidence that B & N lacked good faith, that B & N’s infringement was subjectively willful, and that the Court has jurisdiction to award enhanced damages.
For the foregoing reasons, the Court finds that B & N is liable to Adrea in the amount of $266,832.82. The parties are directed to submit to the Court by January 9, 2017 three-page single-spaced letters giving their competing calculations of the prejudgment interest, if any, to be added to this amount so that final judgment can be entered. The Clerk of the Court is directed to close docket numbers 218 and 221.
SO ORDERED.
. Amazon waited until October 24, 2013 to exercise the option. Curiously, this was six months after Adrea filed suit against B & N.
. Yurkerwich is a certified public accountant and certified valuation analyst. 2016 Trial Tr. (Yurkerwich) at 35:5-15. Barnes, is a certified public accountant and. a certified fraud examiner. Id, (Barnes) at 289:23-290:2.
. The Court notes that while it focuses on these two factors, it has taken account of all the Georgia-Pacific factors, none of which, however, changes the analysis here given,
. Yurkerwich's methodological approach is also suspect. Yurkerwich arrives at this valuation by estimating Discovery’s sales that would have been subject to a royalty payment and applying a 3% royalty rate. 2016 Trial Tr. (Yurkerwich) 66:11-12, 66:17-21, 70:16-71:9, 194:5-13, 198:9-12. Yurkerwich, however, did not base his royalty rate on an analysis of any of the six patents actually cross-licensed by Amazon to Discovery. Id. at 66:22-67:4, 345:8-25. Nor did Yurkerwich base the rate on any patent licensed by Amazon or any licensing fees paid by Discovery. See id. (Barnes) at 345:8-25. Instead, Yurkerwich chose a rate based on six "analogous” patent license agreements (“benchmarks”) found in the public domain. RL (Yurkerwich) at 66:22-67:4, 199:8-200:1, 200:2-7; id. (Barnes) at 345:8-25. Yurkerwich did not review the underlying agreements, but rather relied on descriptions of the patent licenses set forth in securities filings. Id. at 200:2-201:22. Yurkerwich also never spoke with anyone at Adrea about the value of the cross-licenses. Id. (Yurkerwich) at 185:2-5, 191:4-13.
.The situation would be different if Discovery compensated its partners for 75% of the value of the cross-licenses. But there is no evidence that this was the case. Indeed, in an email communication between Adrea's CEO and Amazon dated November 8, 2010, Adrea stated that it would not receive any compensation as a result of Amazon’s cross-license: "Discovery is a minority shareholder in Adrea. We represent Adrea not Discovery .... we ask that you make an offer to settle Adrea vs. Amazon that has a value greater than zero.” DTX-019 at 1.
. Indeed, Amazon likely asserted its cross-claims as a litigation tactic. There is no evidence that Amazon asserted the patents against any entity prior to its litigation with Discovery. There is also no evidence that Amazon had licensed these patents to any other entity. It is therefore likely that the value of these patents was limited.
. Plaintiff’s counsel objected to the admission of this document under Federal Rule of Civil Procedure 408 because Adrea prepared the materials for settlement purposes. Rule 408, however, permits such materials to be used for impeachment purposes, and defense counsel used the communications to impeach Yurkerwich’s testimony that the valuation of the cross-licenses was $10 million. See Playboy Enterprises Int'l, Inc. v. On Line Entm't, Inc., No. CIV.A. CV006618DGT, 2004 WL 626807, at *5 (E.D.N.Y. Mar. 29, 2004).
.Adrea’s counsel identified on cross-examination a mediation document prepared by Amazon stating, among other things, that the value of the cross-licenses was "in the millions of dollars.” 2016 Trial Tr. 454:3-456:8. The Court, however, admitted this document for the narrow purpose of providing context for a settlement offer to Amazon. Id. at 236:10-237:10. The portion of the mediation document relating to the supposed value of the cross-licenses goes beyond this narrow purpose and is inadmissible hearsay for any other purpose. See id.
. Yurkerwich testified that “some of the documents from Barnes & Noble indicated that this shopping feature was a primary revenue driver.” 2016 Trial Tr. (Yurkerwich) at 132:19-133:1. Yurkerwich, however, did not explain why this was the case, and so the Court is merely left with Yurkerwich’s speculation.
. This number assumes that Yurkerwich is correct that the cross-licenses were worth $10 million valuation. Once adjusted to take into account a more accurate valuation ($500,-000), Yurkerwich’s litigation risk multiple rises to an unlikely 4.12x.
. Yurkerwich took the Intertrust Pitch’s estimate of Adrea's total projected licensing revenue ($67 million) and multiplied it by Amazon’s share of the e-book market (80%) to determine the projected licensing revenue attributable to Amazon ($53.6 million). Yurker-wich then divided $53.6 million by his estimate of what Amazon paid as part of the Amazon Agreement ($22.5 million) to arrive at a litigation risk multiplier of 2.38x.
. Yurkerwich relies on a second document from Adrea to support his proposed multiplier, which the Court finds unpersuasive. About a year before Adrea and Amazon signed the *311Amazon Agreement, Adrea offered to license its portfolio to Amazon and confer a 20% interest in the joint venture if Amazon paid $50 million and transferred certain patents to Adrea. The parties themselves recognize that such settlement negotiations are unreliable, see id. at 182:15-22, and Yurkerwich fails to explain why $50 million was accurate.
. In coming to this number, the Court makes an additional downward adjustment to account for the fact that some of Amazon’s sales of eReaders were to customers outside of the United States. 2016 Trial Tr. (Yurkerwich) at 81:18-23. B & N did not have such foreign sales. Id. The Court thus adjusts the per unit rate paid by Amazon downwards by approximately 11%. Id.
. As a "sanity” check, the Court notes that this per unit rate is similar to that derived from an opening settlement offer made by Adrea to B & N in 2012, in which Adrea proposed a license to its entire patent portfolio for $0.50 per unit. See JTX-031. Since this opening offer was undoubtedly inflated, the Court adjusts it downwards to $0.20 per unit. This produces a per unit rate for the ’703 Patent of $0,102. While this is higher than the rate derived from the Amazon Agreement, it is because B & N in 2012 would have been the second licensee of Philips’s portfolio. In 2009, however, B & N would have been the first licensee, and Philips would have been willing to offer a discount to get its portfolio off the ground. See id. (Barnes) at 296:7-14; id. (Yurkerwich) at 261:4-6, 281:1-25.
. The district court may, of course, still consider this factor when deciding whether to exercise its discretion. Id
. Adrea points to several other criteria set forth in Read Corp. v. Portec, Inc., 970 F.2d 816, 827 (Fed. Cir. 1992), but these factors are not relevant. The Federal Circuit uses the Seagate factors to determine whether the in-fringer’s acts were willful. i4i Ltd. P'ship v. Microsoft Corp., 598 F.3d 831, 858 (Fed. Cir. 2010), aff'd, 564 U.S. 91, 131 S.Ct. 2238, 180 L.Ed.2d 131, (2011). If the court finds willfulness, it then considers the nine factors set forth in Read to determine whether the in-fringer’s behavior was sufficiently egregious to warrant enhanced damages. The Court does not reach the Read factors because Adrea has not shown subjective willfulness under Seagate.