Reformat Cross Appeal sur replyCase: 20-2011 Document Filed: 05/28/2021
Nos. 20-2011, - 2094 IN THE
United States Court of Appeals
FOR THE FEDERAL CIRCUIT
APPLE INC.,
Plaintiff-Appellant,
V. WI-LAN INC.,
Defendant-Cross-Appellant.
On Appeal from the United States District Court for the Southern District of California
No. 3:14-cv-02235-DMS-BLM
Hon. Dana M. Sabraw
NON-CONFIDENTIAL CROSS-APPEAL REPLY BRIEF FOR DEFENDANT-CROSS-APPELLANT WI-LAN INC.
INTRODUCTION
Apple admits that the district court’s decision to overturn the jury’s original damages verdict rested on a fundamental error. The district court ordered a new trial based on its view that Dr. Madisetti, Wi-LAN’s expert, did not measure the benefits of “VoLTE associated with improved voice quality ‘during loading,’” where Wi-LAN’s patented technology ensures that time-sensitive voice packets are not delayed by competing data. Instead, the court asserted, Madisetti measured the benefits of VoLTE “in general terms.” Id. Apple now concedes, however, that “Madisetti’s methodology . . . ‘compared VoLTE and Skype call-quality during loading.’ ” And Apple admits that loading (i.e., multitasking) is “where the patented technology matters most.” Those concessions are fatal to the district court’s conclusion that Madisetti lacked “a sufficient factual basis” to testify about the benefits of the patented technology.
Apple quibbles that Madisetti did not pepper every sentence of testimony with the words “during loading” or “during multitasking.” But it cannot deny that Madisetti’s actual measurement of the invention’s benefits—a 2.3-point increase in Mean Opinion Score (“MOS”)—was expressly limited to improved call quality “while multitasking.” Nor can Apple deny that Wi-LAN’s other experts used the same 2.3-point increase under loading conditions to produce their damages estimates. It is thus indisputable that Wi-LAN’s damages evidence was directed to the benefit of the patented technology during loading. The district court erred in ruling otherwise. Rather than seriously defend the district court’s rationale, Apple concocts a new theory. In Apple’s view, it was not enough that Madisetti measured VoLTE’s improved voice quality during multitasking, “where the patented technology matters most.” Instead, Apple insists, Madisetti was required to categorically exclude every hypothetical contribution that (unspecified) other technology might have made. The law has never demanded that. If Apple thought Madisetti failed to account for other (unidentified) technology, it was free to cross-examine him about that. Apple never did.
Apple likewise fails to defend the district court’s ruling that other damages evidence—Wi-LAN’s rate sheets and its license with Samsung—was inadmissible. Apple recognizes that such evidence could not itself justify discarding the jury award; Apple concedes that Wi-LAN’s new-trial cross-appeal “rises and falls solely with the admissibility of Madisetti’s evidence.” Regardless, the rate sheets and Samsung license were plainly relevant to and admissible on the hypothetical negotiation between Apple and Wi-LAN. The original damages award should be reinstated.
Apple’s defense of the ruling insulating Intel-based iPhones from liability— based on an expired term license—fares no better. That agreement expired before the chipsets in the accused phones were sold to Apple. And while the license stated, “for clarity,” that it would survive expiration with respect to past “Licensed Activities that were actually engaged in during the Term License Period,” the agreement nowhere authorized Intel to conduct additional, future activities after the Term License Period. Apple implausibly reads a modest “clarifying” clause as transforming an express term license into a virtual perpetual license.
THE JURY’S ORIGINAL DAMAGES VERDICT SHOULD BE REINSTATED
Wi-LAN’s Direct Valuation Evidence Was Properly Admitted
At the first trial, Wi-LAN supported its proposed $0.85/phone royalty—which the jury ultimately awarded—with evidence showing the value of the call-quality improvements the ’145 patent provides during “multitasking” or “loading” conditions (i.e., when multiple applications are using the phone’s network connection). Dr. Madisetti relied on an independent study comparing VoLTE—which uses “express lanes” for voice packets as claimed in the ’145 patent—to Skype, Apple’s non-infringing alternative, under multitasking conditions. Under those conditions, Madisetti explained, VoLTE offered at least a 2.3-point improvement in MOS over Skype. Wi-LAN’s other experts then calculated that Apple would have paid up to $1.22/phone for a 2.3-point voice-quality improvement during multitasking. The District Court’s Rationale for Granting a New Trial Cannot Be Sustained
After trial, the district court deemed that evidence inadmissible because it mistakenly thought the benefit Madisetti had calculated did not reflect “improved voice quality ‘during loading’”—the benefit provided by the ’145 patent—but instead reflected VoLTE’s improved voice quality generally. Before this Court, Apple does not seriously defend the district court’s Assertion that Madisetti’s evaluation was not directed to “voice quality ‘during loading.’” To the contrary, Apple concedes that “Madisetti’s methodology ... ‘compared VoLTE and Skype call-quality during loading.’”
The district court thought that Madisetti merely evaluated “voice call quality using VOLTE compared to voice call quality using Skype.” Madisetti’s analysis, however, measured voice-quality improvements resulting from VoLTE’s use of the patented invention. The ’145 patent addresses the problem of “voice packets” being “delayed] behind data packets from other applications, resulting in “degraded” quality. The invention creates an “express lane” for voice packets, allowing them to “bypassed” data packets from other applications. VoLTE practices the invention by using two “connections”—one for voice packets, one for other data. Skype, by contrast, lacks an “express lane” for voice packets. See id. Consequently, when phone applications are transferring data, or when the network is “congested,” VoLTE should perform better than Skype, which should suffer degraded call quality from delayed and dropped voice packets. Madisetti evaluated the benefits of VoLTE’s use of the patented express lanes by comparing the performance of VoLTE and Skype under loading conditions. He relied on testing by Signals Research Group, whose testers made calls while running software that performed downloads to simulate multitasking— “synching email, updating applications, etc.” VoLTE fared well under such conditions, with an average MOS of 3.7—thanks to Wi-LAN’s express-lane technology (VoLTE “seemed immune to [the] additional network traffic”).
By contrast, Skype—which lacks the patented express lanes—became virtually unusable while multitasking. Without multitasking, Skype worked well, its quality often indistinguishable from VoLTE. During multitasking, however, “Skype performed very poorly”—usually, “no sound” could be heard. The Skype call’s MOS was typically zero—and never higher than 1.4—during the multitasking download test (data points 1-6 and 9 in the figure below), because Skype “failed to deliver the voice packets.” (Skype During Loading: “background traffic downloading on the smartphone could bring Skype Voice to its knees”.
Comparing VoLTE’s average MOS (3.7) with Skype’s highest MOS (1.4) during the multitasking download test, Madisetti quantified the benefit of the invention as a 2.3-point improvement in MOS “under loaded conditions.” The district court deemed Madisetti’s opinion “without factual basis” because the court did not perceive any “connection” between the patented invention and the testing Madisetti invoked. Yet the Signals report itself attributed VoLTE’s superior voice quality during loading to its support for “giving the voice data packets preferential treatment over best effort data packets,” as claimed by the ’145 patent and pinned Skype’s poor showing on the fact that Skype “treated all packets equally instead of giving preferential treatment to real time voice packets,”
The suggestion that Wi-LAN’s other expert, David Kennedy, understood Madisetti’s benefits analysis to address VoLTE as a whole misconceives Kennedy’s testimony. Kennedy did not take the “‘benefits’” Madisetti measured and “‘further apportion’” them between the invention and other technology. Kennedy used Madisetti’s estimate of a 2.3 point improvement in MOS during loading—along with Professor Prince’s consumer-survey analysis showing customers would pay up to $121.70 extra per phone for that improvement—to determine that the value of the patented invention was about $121/phone. Given Apple’s enormous “negotiating leverage,” Kennedy then reasoned that Apple would have claimed the lion’s share of that value; he thus allocated just 1% of that value to Wi-LAN. That does not further “apportion” between the patented invention and other technology. It reflects the principle that negotiating parties will split the incremental value added by the invention based on (among other things) their relative bargaining power.
Apple’s Attack on Madisetti’s Trial Testimony Fails
Apple concedes that “Madisetti’s methodology . . . ‘compare VoLTE and
Skype call-quality during loading,’ ” “where the patented technology matters most.” But it erroneously urges that his “testimony at trial” departed from that methodology and “conflated the patented invention with VoLTE generally.” Madisetti was clear that he analyzed the “benefits that [the patent] claims pro- vide when used in the iPhones,” by assessing call. While seemingly large at first glance, that figure makes sense: During multi- tasking, Skype, the non-infringing alternative, was effectively unusable. That customers would place a high value on the ability to use apps and download files while making calls is unsurprising. The district court did not contest (nor does Apple contest on appeal) that Prince reasonably measured the consumer value according to standard customer-survey methodology; it objected to Prince’s opinion based solely on his reliance on Madisetti.
He told the jury precisely that at the beginning of his benefits opinion, all while showing the jury a slide declaring that the benefit of the invention was “2.3 points better” voice quality “while multitasking”:
Madisetti tied that 2.3-point improvement during loading to use of the patented invention. He explained that Apple’s iPhone, using VoLTE, infringes because it contains a “VoLTE connection” separate from the “data connection” to give voice packets priority. Skype, by contrast, “doesn’t have an HOV lane,” and “does not use” the invention. Madisetti emphasized that the 2.3-point difference was “under loaded conditions.” He explained how tests were performed on a “loaded network” with “background applications” using bandwidth—the scenario the invention was created to address The Signals report, admitted into evidence also tied VoLTE’s better voice-call quality during loading to its giving “voice data packets preferential treatment” as claimed in the ’145 patent. At the conclusion of that testimony, Madisetti again showed a summary slide explaining that the 2.3-point improvement in MOS score was “while multitasking”.
The snippets of testimony Apple invokes do not prove otherwise. Apple urges that Madisetti conflated the invention with VoLTE generally when he testified that “‘you get great quality from the VoLTE that is 2.3 MOS better.’” As Madisetti was making that very statement, however, he was showing the jury the above slide which made clear the 2.3-point improvement was “while multitasking.” Apple’s Assertion that Madisetti “got it wrong in front of the jury,” does not hold water.
He told and showed the jury, time after time, that he was measuring the benefit of the patented invention during multitasking specifically.
Madisetti’s statements that “VoLTE” has better call quality than “Skype,” are also benign. Madisetti occasionally used “VoLTE” and “Skype” as shorthand to refer to the products evaluated in the Signals testing. But the slides he showed the jury while giving that testimony made clear that the testing involved making a call “while emailing and updating apps”—i.e., while multitasking. In all events, observations that the patent offers “‘much higher quality calls’” and “‘improve the iPhone as a whole for voice and cellular data,’ ” are unobjectionable. Users are constantly multitasking. “Many applications automatically connect to the network and generate traffic without any user intervention,” making it “very difficult to prevent” network data use while making a call. Because voice-call packets almost invariably compete with data packets from other applications, nearly all calls benefit from the invention.
Apple’s complaint about Wi-LAN’s closing argument likewise fails. Wi-LAN’s counsel accurately stated that licensing the ’145 patent would have helped Apple “keep up with the voice quality of its competitors” (who had licensed the patent). Because multitasking is ubiquitous, the patented technology improves voice quality nearly all the time. And on the transcript page immediately preceding the one Apple quotes, Wi-LAN’s counsel described the invention as a “feature” of VoLTE that “separates [the] voice packets” to ensure “good sound.” The jury was not left confused when counsel used “VoLTE” as a shorthand just moments after explaining that the invention was simply one VoLTE “feature.” Regardless, the fact remains that Madisetti’s quantification of the patent’s benefits—a 2.3-point improvement—was laser-focused on voice quality during loading. And that 2.3-point improvement was the express basis for the analysis of Wi-LAN’s other experts, who put a dollar value on those benefits in the context of loading. Apple does not explain how the jury could have thought it was awarding the value of VoLTE as a whole when no one even tried to estimate that value.
Apple’s “Isolation” Argument and Newly Invented “Baseline” Theory Cannot Sustain the New-Trial Order
Because it cannot dispute that “Madisetti’s methodology ‘compared VoLTE and Skype call-quality during loading,’” so as to measure the patented invention’s contribution, Apple contends that Madisetti did not fully “isolate” the invention’s benefits. Apple insists that Wi-LAN was required to “prove the absence” of any and all non-infringing features that might have also contributed to improved voice quality during loading. The law does not insist on the “absolute precision” Apple demands. VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d 1308, 1328 (Fed. Cir. 2014). Courts have long recognized that “‘approximation and uncertainty’” are inevitable when it comes to apportionment. Aqua Shield v. Inter Pool Cover Team, 774 F.3d 766, 771 (Fed. Cir. 2014); see Dow- agiac Mfg. Co. v. Minn. Moline Plow Co., 235 U.S. 641, 647 (1915). Accordingly, admissibility under Daubert requires only that expert damages testimony be “reasonable” and “sufficiently tied to the facts of the case.” Summit 6, 802 F.3d at 1296. The rest is for “the jury.” i4i Ltd. P’ship v. Microsoft Corp., 598 F.3d 831, 856 (Fed. Cir. 2010).
That standard was amply satisfied here. Madisetti’s analysis was tied to the contribution of the patented technology—enhanced call quality during loading. Madisetti relied on testing that compared VoLTE, which uses Wi-LAN’s patented “express lanes,” to Skype, which does not. Specifically, he compared VoLTE and Skype voice-call quality while multitasking, where (Apple concedes) “the patented technology matters most.” And the evidence showed that VoLTE’s advantage under those conditions derived from the “key VoLTE feature” of giving “voice data packets preferential treatment over [other] data packets,” as the ’145 patent claims. Madisetti thus tied his analysis to the specific aspect of VoLTE that was claimed: its use of Wi-LAN’s patented technology to create express lanes for voice packets to improve call quality during loading. That amply established the requisite connection between the patent and his estimate. The objections Apple gins up on appeal fail to undermine the reliability of Madisetti’s opinions.
Insofar as Apple believes Madisetti “could have done more,” that was a basis for cross-examination, not exclusion. See Summit 6, 802 F.3d at 1295-96 (“‘vigorous cross-examination’” is “‘traditional and appropriate means’” for challenging damages testimony); Arctic Cat Inc. v. Bombardier Recreational Prods. Inc., 876 F.3d 1350, 1369-70 (Fed. Cir. 2017) (defendant who found expert’s “comparison problematic” should address that through “‘cross-examination’”).
If Apple believed that features of VoLTE other than Wi-LAN’s patented express lanes contributed to the dramatic call-quality differences measured during loading, it could have presented that theory to the jury. But Apple did not. It never raised that theory on cross-examination, much less identified some “other” responsible feature. Nor did Apple’s expert address the issue. His rebuttal to Madisetti’s benefits opinion spans seven transcript pages and contains no analysis of the Signals study except for a perfunctory dismissal. Apple’s post-trial briefing did not address the issue. This Court should not entertain Apple’s “attempt to raise” that newly minted argument “for the first time on appeal.” Gieg v. DDR, Inc., 407 F.3d 1038, 1046 n.10 (9th Cir. 2005); see Bailey v. Dart Container Corp. of Mich., 292 F.3d 1360, 1362 (Fed. Cir. 2002) (“appellee can present all arguments supported by the record” only insofar as they were “advanced in the trial court”).
Even on appeal, Apple merely speculates that “there could be numerous innovations that impact the call-quality difference between VoLTE and Skype” during multitasking—but never identifies such an invention. Speculation about non-infringing features Madisetti could have overlooked—which Apple never raised below—cannot justify exclusion of Madisetti’s opinion. See Summit 6, 802 F.3d at 1295-96; cf. Prism Techs. LLC v. Sprint Spectrum L.P., 849 F.3d 1360, 1376 (Fed. Cir. 2017) (upholding damages theory based on non-infringing alternative where defendant failed “to introduce argument or evidence” regarding “different non-infringing” technology).
Finally, Apple contends that Madisetti was required to factor out some “baseline call-quality difference” between VoLTE and Skype. Apple again never raised that argument at trial or in post-trial motions. The district court never advanced that theory. Apple cannot raise it for the first time now.
Regardless, the record does not support Apple’s new theory that VoLTE has a “baseline” call-quality advantage absent loading. The Signals report explained that a “typical user wouldn’t notice the difference” between Skype and VoLTE call quality absent loading or multitasking. Apple claims that some Signals tests show a voice-quality advantage for VoLTE in a “baseline scenario” without load. But the tests Apple cites actually involved “heavy to moderate network loading,” “when the network had some of the highest loading [the testers] had observed.” Those tests show the invention at work, not a baseline advantage for VoLTE. Indeed, Signals pointed to VoLTE’s giving “voice data packets preferential treatment”—which VoLTE does using the patented technology—as the reason for its better call quality under “loading” conditions.
Wi-LAN never stated that there was a 0.6-point MOS difference between VoLTE and Skype under “ideal-quality” conditions. Wi-LAN described the test in Figure 9 of the Signals report, which involved making a call while downloading a file to simulate “synching email, updating applications, etc.” While the download was paused, Skype’s score rose to 3.1. But that score did not reflect Skype’s quality under “ideal” conditions. In other tests, Skype scored as high as 3.5, see Appx1023 (Fig. 3), nearly matching (and sometimes exceeding) VoLTE’s performance, see Appx1022 (Fig. 2, showing 3.4 VoLTE score)
Apple overlooks, moreover, that Madisetti’s opinion reflected conservative assumptions that rendered any supposed “baseline” difference immaterial. Madisetti calculated that the invention produced a 2.3-point improvement in MOS during loading by comparing Skype’s highest score during the file download test (1.4) to VoLTE’s average score during that test (3.7). that was exceedingly generous to Skype: In every other measurement during the file download (six out of seven data points), Skype’s score was zero, giving it an average score of 0.2. Id. Madisetti thus used a score for Skype that was 1.2 points higher (1.4 - 0.2) than the average score he reasonably could have used. That wipes out, twice over, Apple’s supposed “baseline” difference of 0.6.
The refinements Madisetti made to his expert report in anticipation of the second trial did not, as Apple urges, suggest that his earlier opinion was unreliable. They merely reflect that, after the district court (erroneously) rejected the opinion he gave at the first trial, an even more conservative approach could be found. Madisetti treated Skype’s MOS during the period when the download was paused as a “baseline.” Signals testing showed that Skype’s MOS improved after pausing the download, from a previous maximum of 1.4 (data point 2) to 3.1 (data point 8). Madisetti therefore treated 3.1 as the ceiling for Skype, compared to VoLTE’s average score of 3.7—an ostensible baseline difference of 0.6. He accordingly reduced his estimate of the patented invention’s marginal benefit during loading by 0.6, from 2.3 (as he testified at the first trial) to 1.7. But that approach was almost too conservative. As explained above, the 3.1 score was not actually Skype’s ceiling. In some tests, Skype scored as high as 3.5 (while VoLTE scored as low as 3.4. Under ideal network conditions, without loading, Skype’s voice quality is essentially indistinguishable from VoLTE’s.
All that underscores why Apple’s made-for-appeal quibbles cannot support the new-trial order. What inferences to draw from the Signals report, or how the invention’s incremental value would be split by the negotiating parties, are classic “battle of the experts” issues. By failing to cross-examine Madisetti on those issues, or have its own expert address them, Apple chose not to engage in battle. That choice is no basis for overturning the jury verdict.
Wi-LAN’s Rate Sheets and the Samsung License Were Admissible
Apple does not contend that the supposed inadmissibility of Wi-LAN’s rate sheets and license with Samsung would themselves justify overturning the original damages award. Apple thus concedes—and Wi-LAN agrees—that the new-trial issue presented by Wi-LAN’s cross-appeal “rises and falls solely with the admissibility of Madisetti’s [benefits] evidence.” Accordingly, if the Court determines that Madisetti’s benefits opinion was properly admitted, it need go no further. The jury’s initial verdict should be reinstated. Regardless, the rate sheets and Samsung license were properly admitted at the first trial.
Rate Sheets Bearing on Wi-LAN’s License Negotiations Were Admissible
The district court erroneously held that Wi-LAN’s rate sheets, which showed $0.65-$1.10/unit licensing offers that Wi-LAN made in real-world negotiations, were inadmissible. This Court has long permitted evidence of patentees’ licensing practices as relevant to the hypothetical negotiation.
Apple nowhere defends the district court’s assertion that the rate sheets were not “relevant.” This Court’s precedent is clear that “proposed licenses may have some value for determining a reasonable royalty.” Whitserve, LLC v. Computer Packages, Inc., 694 F.3d 10, 29-30 (Fed. Cir. 2012).
While the district court’s new-trial order did not reach other arguments Apple raised, those arguments do not warrant a remand, contra Apple. Apple raised and lost those arguments at the retrial, and raises them again here on appeal. See Appx648-650 (arguing that baseband processor must be used as royalty base); Apple.Br.51 n.8 (same); Appx656-657 (objecting to comparable licenses); Apple.Br.56-64 (same); Appx661 (objecting to inclusion of Sprint phones); Apple.Br.44-45 (same). Apple’s remaining argument, that Wi-LAN “skewed” the “damages horizon” with evidence that Licensing Terms to
, Appx660-661, was pure makeweight. That evidence was presented by Wi-LAN’s infringement expert to rebut Apple’s argument that Wi-LAN’s patents were limited to WiMAX, which Apple characterized as having failed in the market. Apple identified no place where Wi-LAN even remotely characterized that evidence as relevant to damages. There was no risk the jury mistakenly thought that evidence related
to damages, much less prejudice that could warrant a new trial.
Licensing Terms
While the district court suggested (as does Apple) that Wi-LAN’s rate sheets were “ ‘outrageous offers’ ” like those in Whitserve, that claim is absurd. Whitserve involved an unaccepted 31% royalty rate offered to justify a 19% royalty. Here, the rate sheets’ $0.65-$1.10/unit range was below the $1.25/unit rate some licensees actually paid.
Apple denies “these rates had been accepted in the real world.” But the report of Apple’s own expert, Lance Gunderson, included a “listing of Wi-LAN license agreements,” that showed Wi-LAN had negotiated rates up to $1.25/unit—well above the royalties in the rate sheets:
Confidential Royalty Payments
Confronted with that at trial, Gunderson claimed he had not “looked at these” specific licenses and was not personally “aware of” licensees paying those amounts. But he offered no explanation why his own report said the opposite— that Wi-LAN, in fact, received those amounts. Apple cannot avoid its own expert’s admission in his report by invoking a convenient bout of trial-induced amnesia.
Apple erroneously asserts that the rate sheets were not apportioned to address that they were portfolio licenses. Wi-LAN’s expert David Kennedy considered various portfolio licenses, including the rate sheets, to produce a range of royalty rates, which he then adjusted to account for the smaller number of patents at issue here. Apple likewise fails to distinguish CSIRO v. Cisco Systems, Inc., 809 F.3d 1295, 1303 (Fed. Cir. 2015). There, as here, rate cards were “presented to prospective licensees,” and the expert adjusted those rates to account for their portfolio nature. And in VirnetX, this Court upheld expert damages testimony that (among other things) relied on the patentees’ established licensing “policy.” VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d 1308, 1330 (Fed. Cir. 2014). While Apple denies VirnetX considered the issue, Apple expressly challenged the expert’s reliance on the “licensing policy” there and this Court affirmed.
Admission of Samsung’s License Does Not Support a New Trial
The district court erred in ruling that Wi-LAN’s license with Samsung,
Apple’s closest competitor, was inadmissible. The license was highly relevant—it gave Samsung an “adjustment right” that would have “cost Wi-LAN a Royalty Payment” if Wi-LAN granted Apple a cheaper license. It would have made no sense for Wi-LAN to grant Apple a license that would have triggered Samsung’s adjustment right and created a net loss for Wi-LAN. That evidence of Wi-LAN’s “economic circumstances” was clearly relevant to the hypothetical negotiation. See Studiengesellschaft Kohle, m.b.H. v. Dart Indus., Inc., 862 F.2d 1564, 1569 (Fed. Cir. 1998) (“most favored licensee” clause was a “very good reason” for patentee to insist on particular royalty
structure).
While Apple insists that economic evidence is “relevant” only when it directly establishes the invention’s value or is “needed to adjust some other prior license,” it cites no authority imposing that restrictive view.
Economic circumstances evidence was used for those purposes in LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51, 67-68 (Fed. Cir. 2012), and Finjan, Inc. v. Secure Computing Corp., 626 F.3d 1197, 1211 (Fed. Cir. 2010), but neither case held those were the only permissible uses. And Georgia-Pacific analysis includes factors that contradict Apple’s narrow view, such as the parties’ “commercial relationship” and the infringing product’s “popularity.” Georgia-Pacific Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116, 1120 (S.D.N.Y. 1970); see Vectura Ltd. v. Glaxosmithkline LLC, 981 F.3d 1030, 1042 (Fed. Cir. 2020) (considering “leverage in the hypothetical negotiation”). Georgia-Pacific considers what “amount would have been acceptable by a prudent patentee.” Here, a “prudent patentee” would not have agreed to suffer a net loss by granting Apple a cheaper license than Samsung.
Samsung’s license, moreover, was admitted to help “adjust . . . prior license”—which Apple concedes is a permissible purpose. As Wi-LAN explained, its prior licenses with Doro, Vertu, and Unnecto did not trigger Samsung’s adjustment right—only a license to Apple would. The Samsung license thus helped the jury understand why Wi-LAN would not necessarily agree to license Apple on the same terms as those prior licensees, and to adjust the prior licenses accordingly.
Apple’s argument that the Samsung license did not help value the patented technology is beside the point. Wi-LAN did not seek the
license’s admission for that purpose; rather, Kennedy considered its “adjustment right” as one “factor” affecting the terms Wi-LAN would agree to in a hypothetical negotiation with Apple. While Apple complains that Samsung was a portfolio license, Kennedy accounted
for that. And given that the Royalty Wi-LAN received from Samsung was markedly less than the proposed $145.1 million royalty
supported by Wi-LAN’s other evidence—including Wi-LAN’s comparable licenses, rate sheets, and direct-valuation analysis, Apple can hardly argue it improperly “skewed the damages horizon for the jury.”
THE DISTRICT COURT MISCONSTRUED THE INTEL LICENSE
Apple does not defend the district court’s actual rationale for holding that Intel-based iPhones are licensed and therefore non-infringing. Apple concedes that the issue turns on the meaning of “‘Licensed Activities that were actually engaged in during the Term License Period.’” But the district court never construed that phrase. Its analysis omitted the critical language with an ellipsis. “The clause at issue specifically states that ‘the licenses granted . . . shall survive the expiration of the Term License Period.’ ” (emphasis and alterations in original)). Apple’s efforts to backfill the district court’s ruling are unsuccessful. Everyone agrees that Intel’s “Term License” for the asserted patents expired in January 2017. Everyone agrees the Term License survived expiration only “with respect to Licensed Activities that were actually engaged in during the Term License Period.” And everyone agrees the chipsets in the accused Intel-based iPhones were actually sold to Apple after the Term License Period. Yet Apple implausibly maintains that the accused iPhones are licensed because post-expiration chipset sales should be deemed to have been “actually engaged in during the Term License Period.” That Wonderland logic cannot be sustained.
Intel Chipsets Sold After January 2017 Are Not Licensed
The license agreement is clear that Intel received only a “Term License” to the asserted patents—a license “for the Term License Period” that expired January 21, 2017. After January 21, 2017, Intel was no longer authorized to make, use, or sell Wi-LAN’s invention. Chipsets sold to Apple after that date thus are not licensed.
The agreement’s “survival clause” does not undo that arrangement. It serves only a modest role—to state, “for clarity,” that activities that “were actually engaged in during the Term License Period” remain licensed even after the license period expires. As to those past activities—but only those past activities—the licenses for the asserted patents “survive” the term’s expiration. The survival clause does not purport to do more, or otherwise disturb the parties’ agreement that Intel could practice the patents only “for the Term License Period.” The clause merely purports to “clarify”—not change—the provision preceding it.
Apple nonetheless demands that the survival clause be construed to do much more than the provision it merely “clarified.” Apple cannot argue that post-expiration sales of chipsets were actually engaged in during the Term License Period. So it argues that such sales were constructively engaged in during the Term License Period: Because Intel sold Apple some chipsets before expiration, Apple insists that Intel could sell it other chipsets after expiration. But that is not what the survival clause says. Apple’s position requires revising virtually every term of the clause.
Were: While the clause covers only activities that were engaged in during the Term License Period, Apple would have it cover activities that will be engaged in after the Term License Period.
Actually engaged in: While the clause covers only activities that were actually engaged in during the Term License Period, Apple would have it cover activities that were at most constructively engaged in during the Term License Period.
During the Term License Period: While the clause covers only activities that were actually engaged in during the Term License Period, Apple would have it cover activities that were actually engaged in after the Term License Period.
Far from honoring the agreement’s “text,” Apple rewrites it.
Apple tries to salvage its position by leaning on the term “Licensed Activities.”
According to Apple, because “Licensed Activities” refers to “categories of acts—‘making, using, selling, offering for sale . . . any and all Licensed Products’”—the survival clause requires only that a given activity “category” was actually engaged in during the Term License Period. If Intel actually engaged in sales of a particular “type of product” during the term, Apple insists, it could engage in additional sales of the “same type” of product after the term’s expiration. But that is not what the survival clause says. It does not refer to “type of product” or “activity categories.” It refers only to particular “Licensed Activities”—particular makings, sellings, etc.—that were actually engaged in during the Term License Period. The stubborn fact remains that sales occurring after the Term License Period were not “actually engaged in during the Term License Period.”
Post-expiration sales thus fall outside the license.
That conclusion flows from the fundamental principle that “each sale” of a patented article constitutes a distinct act of infringement. Catalina Lighting, Inc. v. Lamps Plus, Inc., 295 F.3d 1277, 1291 (Fed. Cir. 2002). Each sale thus requires its own authorization to avoid infringement; a defendant cannot escape liability for a sale of an infringing article simply because the patentee licensed other sales of the same type of product at a different time. Apple nowhere disputes that principle: It admits that “‘each sale must be judged on its own merits’ when deciding whether infringement has occurred.”
Apple insists that principle is irrelevant when determining whether a sale is “licensed.” But whether a sale infringes and whether a sale is licensed are two sides of the same coin—infringement occurs when someone sells the patented invention “without authority.” If “deciding whether infringement has occurred” requires “judging each sale on its own merits”—and Apple concedes that it does—then deciding whether licensing has occurred likewise requires judging each sale on its own merits. Here, each challenged sale of the chipset at issue occurred after the Term License Period and was not “actually engaged in during the Term License Period.”
While each sale requires its own authorization, a license may authorize multiple sales. For example, the license here authorized every sale “actually engaged in during the Term License Period,” without identifying each sale individually. But the license did not authorize, individually or collectively, sales actually engaged in after the term expired.
The license does define “Licensed Activities” by reference to “categories of acts—‘making, using, selling, offering for sale . . . any and all Licensed Products.’ ” But the Patent Act defines infringement by reference to the same categories of acts: An entity infringes where it “makes, uses, offers to sell, or sells any patented invention” without authorization. The identification of categories of acts thus does not override the rule that each sale must be judged on its own merits. To the contrary, it is the source of the rule: “Each sale constitutes an infringement,” this Court has explained, “because ‘whoever without authority makes, uses, offers to sell or sells any patented invention’ is an infringer.” Catalina Lighting, 295 F.3d at 1291 (quoting §271(a)) (emphasis added).
Apple offers no reason to suppose that Wi-LAN and Intel, despite adopting the Patent Act’s language virtually verbatim, sought to give those terms a completely different meaning. Doing so would invite serious disruption. The rule that each act of selling an invention is a distinct act of infringement is a “background legal rule” relevant to contract interpretation. U.S. Airways, Inc. v. McCutchen, 569 U.S. 88, 102 (2013). “Ignoring” such rules “frustrate the contracting parties’ intent and produce perverse consequences.”
Apple’s construction raises more questions than it answers. For example, if the survival clause applies to “categories of sales” for the “same type of product,” how does one define the relevant “categories” or determine whether products sold before and after expiration are the “same type”? The license agreement cannot answer those questions; it never mentions “categories” or “types.” And while Apple urges that the relevant “category” or “type” here is sales of chipsets, it offers no textual foundation for that assertion. The agreement never mentions chipsets or identifies “sale of a particular chipset” type as a relevant category.
Nor has Apple (which bears “the burden of proving license as a defense”) presented any evidence that all chipsets labeled “ Chipset ” are identical in every respect—or offered any basis for identifying what differences should count when deciding whether products are the “same type.” The background rule that each sale must be judged on its own merits, by contrast, provides clear answers: If a particular sale was “actually engaged in during the Term License Period,” then the license for that sale survives the term’s expiration. If not, then not.
Apple’s position makes a hash of the license agreement’s structure. The provision here, Section 3.2, granted a “Term License” to the asserted patents, while another provision, Section 3.1, granted “Perpetual Licenses” to different patents. As the district court admitted, however, its (and Apple’s) construction would effectively convert a Section 3.2 Term License into “a perpetual license to the patents in suit.” Apple’s quibble that it is not “an unqualified perpetual [license] in the same sense as the license offered in Section 3.1,” is true but irrelevant. The point is that the parties knew how to provide for a “perpetual” license that allows Intel to continue selling some patented products indefinitely. But the parties chose not to provide for a “perpetual” license—of any kind—with respect to the patents asserted here.
Tellingly, while Apple at one point seeks to define the relevant “categories of acts” as, e.g., “ ‘selling . . . any and all Licensed Products,’ ” it does not argue that Intel could sell “any and all” Licensed Products after the license term so long as it sold “any and all” Licensed Products during the license term. Such a construction would be absurd, which is why neither Apple nor the district court advanced it. But it reinforces that the construction Apple does advance lacks any foundation in the license agreement’s text.
Apple’s Remaining Arguments Fail
Wi-LAN explained the survival clause’s modest purpose: It “clarifies” that licensed sales during the Term License Period will still be considered licensed sales, with patent-exhaustive effect, after the Term License Period’s expiration. Apple protests that, “whereas exhaustion doctrine sets out the rights of the licensee’s customers (e.g., Apple), the survival clause clarifies the rights of the licensee (Intel).” Apple overlooks that Intel’s customers can enjoy the benefits of patent exhaustion only if Intel itself is licensed, because exhaustion requires an “authorized sale” of a patented article. Automotive Body Parts Assn v. Ford Global Techs., LLC, 930 F.3d 1314, 1322 (Fed. Cir. 2019). By clarifying that Intel’s pre-expiration sales remain authorized sales after expiration, the survival clause confirms that a necessary precondition for Intel customers to assert exhaustion remains satisfied.
Apple gets no mileage out of the provision stating that the licenses include the right of Intel’s customers to use and sell Licensed Products “worldwide . . . to the same extent that the Patent Rights of Wi-LAN . . . would be deemed to have been exhausted under United States law if such Licensed Products were first sold in the United States.” While Apple describes that provision as “making clear that exhaustion applies to the full extent it can under ‘United States law,’” the provision does not say that. It says that U.S. patent-exhaustion law applies, not that exhaustion applies to the greatest extent that U.S. law would tolerate. When the license agreement was executed in 2011, U.S. law allowed parties to limit patent exhaustion by contract. It thus made sense for the agreement to “clarify” that expiration would not limit the exhaustive effect of earlier sales. Apple never addresses Mallinckrodt. Apple invokes Impression Products, but ignores that Impression Products departed from the law that was in effect when the license was drafted, by abrogating former law under Mallinckrodt.
Apple protests that Section 3.2’s survival clause “cannot add anything meaningful on the exhaustion front” because there is no similar survival clause governing the perpetual licenses granted in Section 3.1. That makes no sense. The survival clause clarifies that pre-expiration activities remain licensed after “expiration of the Term License Period.” Because Section 3.1 grants perpetual licenses (to other patents), there is no need for it to address what happens after “expiration of the Term License Period.”
In the end, what matters is not how the survival clause interacts with exhaustion law. It is whether the clause, by its terms, licenses the Intel chipsets in the accused iPhones. Because Intel’s sales of those chipsets were not “actually engaged in during the Term License Period,” the answer is “no.”
While Wi-LAN offers the only reasonable construction of the survival clause—and certainly the most reasonable—at worst the contract is ambiguous and “‘fairly susceptible’ ” of Wi-LAN’s construction. Consequently, if the Court does not adopt Wi-LAN’s construction as a matter of law, a jury must resolve the ambiguity if there is a new trial.
Apple erroneously asserts that summary judgment must be affirmed “unless the survival clause is ‘susceptible to two equally reasonable, but conflicting, interpretations.’ ” While the two interpretations offered in GMG were found “equally reasonable,” a jury must resolve ambiguity whenever a provision is “ ‘fairly susceptible of different interpretations.’” Regardless, Wi-LAN’s construction is at least as reasonable as Apple’s.
CONCLUSION
The original damages verdict should be reinstated. Any new trial should include Intel-based iPhones.