Google LLC v. Sonos, Inc.

Opinion

NORTHERN DISTRICT OF CALIFORNIA

Plaintiff, No. C 20-06754 WHA

v.

GOOGLE LLC, ORDER RE MOTIONS TO STRIKE PORTIONS OF EXPERT REPORTS Defendant.

In this patent infringement action, both sides move to strike portions of each other’s expert reports. For the following reasons, patent owner’s motion is GRANTED IN PART and DENIED IN PART, and alleged infringer’s motion is DENIED. Sonos, Inc. accuses Google LLC of infringing U.S. Patent Nos. 10,779,033; 10,848,885; and 10,469,966. These patents generally concern multi-room “smart” speaker technology. Specifically, the ’033 patent covers technology related to transferring playback between devices, i.e., “casting,” whereas the ’885 and ’966 patents cover technology related to managing groups of smart speakers. In the lead-up to trial, both parties have filed motions to strike portions of each other’s expert reports (Dkt. Nos. 464, 469), as well as new motions for summary judgment (Dkt. Nos. undersigned granted at the hearing after Google withdrew its opposition (Dkt. No. 557). This order considers only the motions to strike. 1. SONOS’S MOTION TO STRIKE. According to Sonos, Google improperly introduced new invalidity theories, non- infringement theories, and non-infringing alternatives in its expert reports. By way of background, Sonos explains that Google served invalidity contentions pursuant to Patent Local Rule 3-3 on December 6, 2021, which it urged Google to amend to no avail in March 2022 (Sonos Br. 3; Moss Exh. G). Separately, Sonos served interrogatories asking Google to describe its non-infringing alternatives and non-infringement positions on August 7, 2021 (Moss Exh. H at 16, 19–20). In the last week of fact discovery, Google served supplemental responses to these interrogatories, describing its non-infringing alternatives and non-infringement positions on November 21, 2022, and November 29, 2022, respectively (Moss Exhs. I, J). Then, on the last day of fact discovery, November 30, 2022, Google served the opening expert reports of Dr. Samrat Bhattacharjee and Dr. Dan Schonfeld (Moss Exhs. A, D). Google served the experts’ rebuttal and reply reports on January 13, 2023, and January 23, 2023, respectively (Moss Exhs. B, C, E, F). Dr. Bhattacharjee’s reports address invalidity and non-infringement of the ’033 patent, whereas Dr. Schonfeld’s reports address invalidity and non-infringement of the ’885 and ’966 patents. Sonos argues that “despite [its] discovery obligations and procedural rules designed to promote disclosure and efficiency, Google waited until serving its expert reports to disclose a broad array of new invalidity and non-infringement theories,” thereby “hid[ing] the ball until fact discovery had closed” (Sonos Br. 2). Google purportedly failed to amend its invalidity contentions, to disclose its new positions in its interrogatory responses, and to supplement its disclosures as required under Federal Rule of Civil Procedure 26(e)(1). For that reason, Sonos asserts that Google’s new theories are untimely and improper, and that they should be struck from Dr. Bhattacharjee and Dr. Schonfeld’s reports.

2. GOOGLE’S MOTION TO STRIKE. Meanwhile, according to Google, it was Sonos that improperly introduced new validity and infringement theories in its expert reports. Sonos served infringement contentions pursuant to Patent Local Rule 3-1 on October 21, 2021 (Google Br. 4). Google emphasizes that Sonos was granted leave to amend these contentions several times, most recently on January 12, 2023 (Dkt. No. 447). (In some instances, Google opposed amendment, whereas in others, including the most recent instance, Google did not.) Sonos served the opening expert report of Dr. Douglas Schmidt on infringement of the ’033 patent on November 30, 2022 (Kaplan Exh. 2). It served his rebuttal report on validity and his reply report on infringement on January 13, 2023, and January 23, 2023, respectively (Kaplan Exhs. 3, 5). Google argues that Sonos presented new validity and infringement theories in Dr. Schmidt’s rebuttal and reply reports that are based on claim constructions first disclosed in his rebuttal report and that are different than those provided by the Court. Google further argues that Sonos presented new infringement theories in all of Dr. Schmidt’s reports that were not disclosed in Sonos’s infringement contentions, and that Sonos presented new infringement theories in Dr. Schmidt’s reply report that were not disclosed in his opening report (Google Br. 5). Like Sonos, Google asserts that these new theories are therefore untimely and improper, and that they should be struck from Dr. Schmidt’s reports. The motions will be taken up in turn. This order follows full briefing and oral argument. The patent local rules require parties to disclose infringement and invalidity contentions early. See Patent L.R. 3-1, 3-3. As such, they may not use expert reports “to introduce new infringement theories, new infringing instrumentalities, new invalidity theories, or new prior art references not disclosed in the parties’ infringement contentions or invalidity contentions.” See ASUS Comput. Int’l v. Round Rock Rsch., LLC, 2014 WL 1463609, at *1 (N.D. Cal. Apr. 11, 2014) (Judge Nathaniel M. Cousins). But the rules permit amendment of contentions as O2 Micro Int’l Ltd. v. Monolithic Pwr. Sys., Inc., 467 F.3d 1355, 1365–66 (Fed. Cir. 2006). And, they do “not require identification of every evidentiary item of proof.” Oracle Am., Inc. v. Google Inc., 2011 WL 4479305, at *3 (N.D. Cal. Sept. 26, 2011) (emphases omitted). Meanwhile, the Federal Rules of Civil Procedure require parties to supplement interrogatory responses on pain of not being able to use new information moving forward unless the failure to supplement is substantially justified or harmless. See Fed. R. Civ. P 26(e)(1), 37(c)(1). “The threshold question in deciding whether to strike an expert report” or, in this case, select portions of an expert report, is “whether the expert has permissibly specified the application of a disclosed theory or impermissibly substituted a new theory altogether.” Digital Reg of Tex., LLC v. Adobe Sys. Inc., 2014 WL 1653131, at *5 (N.D. Cal. Apr. 24, 2014) (Judge Kandis A. Westmore). When the line between permissible application of a disclosed theory and impermissible substitution of a new theory blurs, the district court “revert[s] to a simple question: will striking the report result in not just a trial, but an overall litigation, that is more fair, or less?” Apple Inc. v. Samsung Elecs. Co., 2012 WL 2499929, at *1 (N.D. Cal. June 27, 2012) (Judge Paul S. Grewal). This order begins with Sonos’s motion to strike and associated arguments.

1. SONOS’S MOTION: NEW INVALIDITY THEORIES. Starting with its first batch of arguments, Sonos avers that Google improperly introduced previously undisclosed invalidity theories in both Dr. Bhattacharjee and Dr. Schonfeld’s expert reports. Sonos seeks to strike material related to: (A) the Tungsten/NexusQ system’s “Magic Playlist” feature; (B) the YouTube Remote system’s provision of “automatic playback”; and (C) certain Bose products. A. TUNGSTEN/NEXUSQ SYSTEM’S “MAGIC PLAYLIST.” First, according to Sonos, Google improperly asserted for the first time in Dr. Bhattacharjee’s opening report that the “Magic Playlist” feature of the Tungsten/NexusQ system anticipated claim 1 of the ’033 patent (Sonos Br. 6–8). Dr. Bhattacharjee’s opening report states that “[i]n addition to allowing users to play user-created playlists, the prior art 1 cloud” (Moss Exh. A § 218). This “allowed users to request that the cloud servers generate 2 and provide a playlist of songs that were related to, for instance, specific tracks or albums, 3 which was called a ‘Magic Playlist’” (Moss Exh. A §] 220). As such, “[t]he Magic Playlist 4 [was] a ‘remote playback queue’” per limitation 1.4 of the ’033 patent, which requires a 5 “computing device [] configured for playback of a remote playback queue provided by a cloud- 6 based computing system associated with a cloud-based media service” (Moss Exh. A J 499; 7 °033 patent 17:39-42). Sonos decries that the Magic Playlist feature appeared nowhere in 8 Google’s invalidity contentions, which only disclosed the Tungsten/NexusQ system as prior art 9 (Sonos Br. 6). 10 But there is no new theory of invalidity here. Admittedly, Google should have been 11 more forthcoming when it stated in its invalidity contentions that “the Tungsten/NexusQ 12 system anticipate[d] the asserted claims of the ’033 patent under at least the interpretation that 13 [Sonos] appears to rely on for its infringement theories” (Google Opp. 5) (internal quotation 14 marks omitted). But Google provided additional information in its invalidity contentions 3 15 sufficient to disclose its theory of invalidity. Note Google’s invalidity claim chart described a 16 how the Tungsten/NexusQ system enabled users to “pull the music directly from the music 3 17 library in the cloud,” with a screenshot showing one user selecting a playlist to make 18 “(alvailable offline”: 19 - 7 □ 21 23 —_ ee 22 ba ae vam DAUM 23 ee 24 _ a ge A aa ole =

06 : ca we nce) Mol OVI Cemkedan) eg | les a: a” a a 27 Item [4] from Invalidity Claim Chart for ’033 Patent. 28 (Moss Exh. K at 9-12).

Sonos takes issue with the fact that Google did not expressly disclose a “playlist” when it stated that the Tungsten/NexusQ system allowed users to pull music directly from the cloud; after all, “pulling a single media file from a music library in the cloud would not require a ‘playlist’ at all, much less one ‘generated’ by a cloud server” (Sonos Reply Br. 2–3) (emphasis in original). But playlists are clearly pictured in the screenshot above. That they could be made “available offline” reflects that they were stored in the cloud. And, that the text directly above the screenshot referred to “[v]arious forms of streaming media . . . be[ing] requested,” such as “live radio program[s]” from services “includ[ing] YouTube and Pandora,” adequately put Sonos on notice that these playlists could be generated by a cloud server (Moss Exh. K at 11–12). Taken together, that was sufficient disclose Google’s theory of invalidity. This order declines to strike material related to the Magic Playlist feature. B. YOUTUBE REMOTE SYSTEM’S “AUTOMATIC PLAYBACK.” Second, according to Sonos, Google improperly asserted for the first time in Dr. Bhattacharjee’s reply report that the YouTube Remote system anticipated claim 1 of the ’033 patent based on its provision of “automatic playback” of lists of videos (Sonos Br. 7–8). Dr. Bhattacharjee’s reply report explains that “[a]lthough the ‘+’ icon c[ould] be used to add a recommended video to the user’s queue, the list of recommended videos c[ould] also be played back without adding them to the user’s queue” (Moss Exh. C ¶ 79). Dr. Bhattacharjee annotates a screenshot from Google’s invalidity claim chart that “shows the YouTube Remote prior art playing back a list of videos provided by the YouTube cloud servers in the ‘Search’ workspace — in other words, without using the ‘+’ icon to add the videos to the user’s queue” (ibid.). Dr. Bhattacharjee suggests that the YouTube Remote system thereby allowed for playback of a remote playback queue, not just playback of a local playback queue when a user manually added videos using the “+” icon. The annotated screenshot from Dr. Bhattacharjee’s report is reproduced below: 1 F + Es 2 ‘ al i 3 Fg 4 5 — ee ten tesa Pa 6 Par ib tag Lee) 4 ie bette a Li ae ity g + 9 Le] +) 10 md paiproekelin

Annotated Screenshot from Reply Report a 12 of Dr. Bhattacharjee.

a: 14 (ibid.). 15 This order agrees with Google that its invalidity contentions sufficiently disclosed the A 16 underlying theory. Only Dr. Schmidt’s rebuttal report (that Dr. Bhattacharjee’s reply report

= 17 was directed to) limited Google’s theory to playback of a queue of videos a user expressly

18 added to the “Queue” workspace, elsewhere referred to as the “Queue” tab (see Google Opp. 5 19 n.3). Google’s invalidity claim chart provided screenshots of playback of queues of videos 20 from the “Search” and “Recommended” tabs in its discussion of limitation 1.4, like the one 21 above unannotated (Moss Exh. L at 7, 11). Further, any doubt about the disclosure (and 22 feasibility) of “automatic playback” from such tabs appears to be foreclosed upon review of 23 item [4] from that claim chart, a video of the YouTube Remote system that the parties and the 24 undersigned closely scrutinized in evaluation of other motions (Moss Exh. L).! It shows 25 playback of a queue of videos from a “Best of YouTube” tab — not the “Queue” tab. Thus, 26 the theory that the queue of videos in the other tabs could have been remote playback queues 27 28 ' Citing https://www.youtube.com/watch?v=EGdsOslqG2s (last visited April 11, 2023).

— and that the YouTube Remote system could have been “operating in a first mode in which the computing device is configured for playback of a remote playback queue provided by a cloud-based computing system associated with a cloud-based media service,” per limitation 1.4 — was adequately disclosed, even if Sonos disagrees with it. This order declines to strike material related to the YouTube Remote system’s provision of automatic playback. Third, Sonos avers that Google improperly relied upon a number of undisclosed Bose products as prior art in Dr. Schonfeld’s reports (Sonos Br. 9–11). According to Sonos, although Google disclosed the primary reference, the Bose Lifestyle 50 System, it did not disclose the “Bose Link communication protocol” and the “Lifestyle SA-2 and SA-3 Stereo Amplifiers” (Sonos Br. 9 (citing Moss Exh. D)). Sonos asserts that the primary reference did not include those products, as demonstrated by the owner’s guide, among other documents (Sonos Br. 9 (citing Moss Exhs. M, N, O, P, Q)). What’s more, according to Sonos, it would be impossible for Dr. Schonfeld to show that these products were part of the primary reference because the primary reference was discontinued before the amplifiers were even released (Sonos Br. 9–10 (citing Moss Exhs. R, S)). As such, “there is little discernable connection between the previously disclosed Bose Lifestyle 50 System, and the newly disclosed communication protocol and amplifiers (other than a common manufacturer)” (Sonos Br. 10). Accordingly, Google’s failure to disclose these products was “inherently prejudicial to Sonos” (Sonos Reply Br. 5). Not so. This order acknowledges that Sonos makes colorable arguments, but Sonos should have made them roughly nine months ago. Google disclosed these Bose products on June 6, 2022, when it served Dr. Schonfeld’s opening report on invalidity for the patent showdown round of summary judgment motions (Kaplan Opp. Exh. 11 ¶¶ 619–722; see, e.g., ¶¶ 653, 677). It served documents relating to these Bose products even earlier, on September 30, 2021 (Kaplan Opp. Decl. ¶ 3). As Google observes, Sonos seems to have sat on its hands waiting to strike expert report language, just like it did during the patent showdown. Again, 3052559, at *5. Google is correct that “[i]f Sonos believed that certain Bose-related theories disclosed in Dr. Schonfeld’s report were missing from Google’s invalidity contentions, Sonos could have raised the issue at any time thereafter,” including during the two depositions of Dr. Schonfeld, one of which was taken after the motion to strike was filed (Google Opp. 9). In any case, as Google observes, Sonos’s technical expert Dr. Kevin Almeroth discussed these Bose products at length in both of his expert reports (Google Br. 10 (citing Kaplan Opp. Exhs. 15 ¶¶ 215–250, 748–964; 16 ¶¶ 348–408, 1249–1557)). Like the order on Sonos’s prior motion to strike, this order will not reward Sonos’s delay tactics, particularly when it is clear that Sonos was not actually prejudiced. See BLK Enterprises, LLC v. Unix Packaging, Inc., 2018 WL 5993839, at *2 (C.D. Cal. Oct. 2, 2018) (Judge Karen L. Stevenson). Because any non-disclosure was harmless, this order declines to strike material related to the Bose products. 2. SONOS’S MOTION: NEW NON-INFRINGEMENT THEORIES. Turning to the next batch of arguments, Sonos contends that Google improperly introduced previously undisclosed non-infringement theories in both Dr. Bhattacharjee and Dr. Schonfeld’s rebuttal reports. Specifically, Sonos enumerates five purportedly new non- infringement theories in Dr. Bhattacharjee’s rebuttal report and one purportedly new non- infringement theory in Dr. Schonfeld’s rebuttal report. This order finds that none are, in fact, new theories. First, Sonos argues that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report “that limitation 1.7 of the ’033 Patent is not infringed because the setPlaylist message sent by the Sender to the MDx server is not identical to the setPlaylist message sent by the MDx server to the Receiver” (Sonos Br. 11). Suffice to say, under limitation 1.7, the “Sender” is a control device and the “Receiver” is a playback device. In response to interrogatory 12, Google expressly stated that: (1) “a YouTube application running on the alleged computing device transmits a ‘setPlaylist’ message ‘to one or more MDx servers,’” (2) “[t]he MDx servers then generate a further setPlaylist message,” and (3) “it is the MDx server, not the Exh. J at 47–48). Sonos may disagree that the messages sent and received by the MDx server are, in fact, different such that they fail to satisfy limitation 1.7, but this theory of non- infringement was sufficiently disclosed. Second, Sonos argues that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report “that limitation 1.7 of the ’033 Patent is not infringed because a videoID is not used by Receiver [sic] to retrieve at least one media item in the remote playback queue because (a) Bandaid URLs are used to retrieve rather than videoId and (b) a Receiver obtains only chunks of a media item at a time rather than a whole media item” (Sonos Br. 12). No detailed discussion of technical terms is warranted here. As is arguably betrayed by the sentence structure, Dr. Bhattacharjee’s expert report merely provided additional explanation regarding the theory of non-infringement that Google disclosed in various interrogatory responses. Sonos is correct that “disclosing a technical explanation of how a videoId results in retrieval of a media item is not the same as arguing that a videoId is not ‘use[d]’ by the playback device ‘to retrieve’ a media item” (Sonos Reply Br. 8), but Sonos had what it needed to connect the dots and conduct discovery on this theory of non-infringement — which it did, as Google points out (Google Opp. 15 (citing Kaplan Opp. Exh. 8 at 85:3–94:11, 113:15–115:7)). In other words, this theory of non-infringement was sufficiently disclosed. Third, Sonos argues that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report that limitations 1.4 and 1.7 of the ’033 patent are not infringed because, in essence, a computing device is not configured for playback of the remote playback queue prior to the transfer of playback responsibility to a playback device (Sonos Br. 12). Google’s interrogatory language sufficiently introduced its theory of non-infringement here, expressly stating that “when playing back media on the alleged ‘computing device,’ [an] accused YouTube application plays back a local queue stored on the computing device” such that “Sonos has failed to show that [an] accused YouTube application infringes the ‘remote playback queue’ limitations that require playback of the remote playback queue on the computing device” (Moss Exh. J at 60) (emphases added). Sonos’s argument appears to be motivated not by have both a local playback queue and a remote playback queue. To say that Sonos’s contentions did not allow for this possibility, however, is a step too far. This theory of non- infringement was sufficiently disclosed.2 Fourth, Sonos argues that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report “that a Google Hub device is not a ‘computing device’ as claimed by claim 1 of the ’033 Patent” (Sonos Br. 12). But this was directly responsive to an infringement argument made in Dr. Schmidt’s opening report about initiating playback on a Google Hub device and then transferring playback to a second Google Hub device (see Google Br. 9 –11 (citing Kaplan Exh. 2 ¶¶ 178, 247)). According to Google, it could not have disclosed this position earlier because Sonos did not disclose this “new theory of infringement” earlier (Google Opp. 18). For that reason, Google moves to strike Sonos’s infringement contentions related to the Google Hub device, which Sonos opposes (Google Br. 9–11; Sonos Opp. 17–21). As will be discussed below, this order disagrees that Sonos disclosed a new theory of infringement involving the Google Hub device. In any event, because both sides’ experts spoke to the Google Hub device’s alleged infringement, this order declines to strike related responsive material. Fifth, Sonos argues that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report “that a [Google] Hub device running YT Main app does not satisfy limitations 1.5–1.6 of the ’033 Patent because it pauses playback when the Cast icon is selected” (Sonos Br. 12). This order declines to strike such material involving the Google Hub device for the reasons stated directly above. Lastly, Sonos argues that Google stated for the first time in Dr. Schonfeld’s rebuttal report “that the Accused Google Controllers do not ‘cause storage’ of zone scenes and therefore do not infringe asserted independent claims 1 and 9 of the ’966 patent” (Sonos Br. 13) (emphasis omitted). According to Sonos, Google’s interrogatory responses only argued

2 Sonos’s related argument that Dr. Bhattacharjee newly and thereby improperly interpreted remote playback queue in limitation 1.4 to refer to the same remote playback queue in limitation 1.7 is flatly rejected. As Google observes, this is not expert opinion but rather a well-settled this “lack of storage” with respect to dependent claim 3, from which asserted claim 4 depends (ibid.). But as explained by Google, just because it supplemented its interrogatory response to provide additional disclosure regarding non-infringement of claim 3 does not mean that it disclaimed its original response that all of the asserted independent claims did not meet the “causing storage” limitations (Google Opp. 19 (citing Kaplan Opp. Exh. 18 at 39–42)). Google incorporated its previous response by reference into its supplemental responses. As such, Google’s theory of infringement was sufficiently disclosed. 3. SONOS’S MOTION: NEW NON-INFRINGING ALTERNATIVES. Turning to the final batch of arguments, Sonos asserts that Google improperly introduced three previously undisclosed non-infringing alternatives in both Dr. Bhattacharjee and Dr. Schonfeld’s rebuttal reports. First, Sonos contends that Google stated for the first time in Dr. Bhattacharjee’s rebuttal report that a non-infringing alternative existed at the time of infringement “consisting of the ‘receiver device not [] send[ing] the accused getWatchNext and/or getPlayer requests,’ with the ‘receiver device’ instead ‘send[ing] a request to a Onesie agent’” (Sonos Br. 15 (quoting Moss Exh. B ¶ 288)). But this was disclosed in, inter alia, Google’s response to interrogatory 18. In that response, Google explained that the playback device could “send a request to a Onesie agent,” which would be tasked with obtaining media data by making “GetWatchNext” and “GetPlayer” requests, and would then “stream the media data from the Onesie agent to the receiver device” (Kaplan Opp. Exh. 24 at 13)). That Google later incorporated into its response showdown opinions regarding the Onesie agent and the ’615 patent did not serve to change or erase this disclosure regarding the Onesie agent and the ’033 patent (see Google Opp. 21–22). Next, Sonos contends that Google improperly expanded on what was disclosed as “NIA #4” in Google’s interrogatory response and “Alternative #3” in Dr. Bhattacharjee’s rebuttal report: continuing playback at a control device after playback has been transferred to a playback device (Sonos Br. 15). According to Sonos, Alternative #3 contained three additional apply to ‘all of the YouTube applications,’ including YouTube Music” was never disclaimed; rather, Google expressly disclosed that the alternative applied to “Google’s accused products,” which would cover all of the YouTube applications (Sonos Br. 16 (emphasis in original); Kaplan Opp. Exh. 24 at 14). Meanwhile, that “the alternative could be implemented so that all of the YouTube applications default to pausing playback of the media . . . upon transfer” was merely explanation of Google’s existing theory in response to language in Dr. Schmidt’s opening expert report, which stated that an accused product is configured for playback of the remote playback queue even when it is paused and not playing back media (Sonos Br. 16 (emphasis in original); Moss. Exh. B ¶ 280). But this order agrees with Sonos that Google never disclosed this alternative “could also be implemented so that the video and audio would continue to playback on the mobile device after transferring playback” (Sonos Br. 16). Google suggests that its discussion of playback continuing with “only the video alone (and the audio muted)” was “but one example of the disclosed NIA” (Google Opp. 23). Yet there is no language in its interrogatory response — e.g., “e.g.” — that suggests this was but one example (Kaplan Opp. Exh. 24 at 14). As such, Sonos’s motion to strike as to the relevant language in paragraph 280 of Dr. Bhattacharjee’s rebuttal report is GRANTED. Finally, Sonos argues that Google improperly advanced a new non-infringing alternative in Dr. Schonfeld’s rebuttal report, called “No Identification of Groups as Zone Scenes” (Sonos Br. 17 (citing Moss Exh. E ¶¶ 178–81)). Note Dr. Schonfeld himself acknowledged that he did not discuss this non-infringing alternative in his opening report (Moss Exh. E ¶ 178). According to Google, this argument “elevates form over substance” because this non- infringing alternative “follows directly from the Court’s discussion of naming a group” in the prior order on the ’885 patent (Google Opp. 23–24). Even so, it is unclear why Google did not incorporate this non-infringing alternative when it supplemented its interrogatory responses after the patent showdown, and why it did not incorporate this non-infringing alternative in Dr. Schonfeld’s opening report. This was insufficiently disclosed. Google points out that Dr. Almeroth discussed this non-infringing alternative in his reply report (Google Opp. 24 (citing Kaplan Opp. Exh. 23 ¶¶ 268–77)). But this discussion was sparse, and understandably so. Dr. Almeroth expressly “note[d] that Dr. Schonfeld ha[d] not provided sufficient details regarding this new alleged non-infringing alternative, and as a result, [Dr. Almeroth did] not have enough information to fully evaluate whether this alleged non-infringing alternative would be non-infringing, technologically viable, available to Google, or commercially acceptable” (Kaplan Opp. Exh. 23 ¶ 268). It cannot be said that this nondisclosure was substantially justified or harmless. See Fed. R. Civ. P 37(c)(1). Sonos’s motion to strike as to paragraphs 178–81 from Dr. Schonfeld’s rebuttal report is GRANTED. 4. GOOGLE’S MOTION: NEW CLAIM CONSTRUCTION. This order now turns to Google’s motion to strike and associated arguments. First, Google argues that Sonos introduced a new claim construction in Dr. Schmidt’s rebuttal report (Google Br. 5–7). During the patent showdown, the Court construed the term “playback queue” in claim 13 of the ’615 patent as “a list of multimedia content selected for playback.” Google, 2022 WL 3046752, at *5. According to Google, Dr. Schmidt’s rebuttal report “proposed a different claim construction” because it suggested that a playback queue is only a playback queue if it possesses “additional purported requirements” (Google Br. 5–6). Dr. Schmidt’s rebuttal report states, in pertinent part: Thus, according to the Court’s Order, I understand that the claim term “playback queue” refers to a “list of multimedia content selected for playback” with the following characteristics:

• The playback queue is the list of media items that is used for playback; • The playback queue contains the entire list of media items selected for playback; • The playback queue is not being used merely to process the list of media items for playback; • The playback queue is the queue that “runs the show.” (Kaplan Exh. 3 ¶ 107). This order does not read Dr. Schmidt’s language as “propos[ing] a different claim construction” but rather as applying it to rebut contentions in Dr. Bhattacharjee’s reports. opinions of another expert. See MasterObjects, Inc. v. Meta Platforms, Inc., 2022 WL 4856269, at *4 (N.D. Cal. Oct. 3, 2022). Further, this order credits Dr. Schmidt’s deposition testimony that he did not expressly articulate these characteristics in the same way in previous reports because here he was responding to what he perceived as flawed new theories raised by Dr. Bhattacharjee (see Richter Exh. 4 at 143:18-144:14). According to Dr. Schmidt’s rebuttal report, Dr. Bhattacharjee had previously represented that a local playback queue and a remote playback queue were mutually exclusive, and his latest reports took a different position (Richter Exh. 2 ¶¶ 301–05). The bulleted characteristics were therefore included to support Dr. Schmidt’s argument that the relevant prior art can only have one playback queue and — because the prior order indicated that the local playback queue was this playback queue — there was no remote playback queue in the prior art. The merits of this argument will be taken up elsewhere. Google may disagree with Sonos’s application of the Court’s claim construction, but this does not warrant striking it from Dr. Schmidt’s rebuttal report. 5. GOOGLE’S MOTION: NEW INFRINGEMENT THEORIES. Google makes two primary arguments regarding infringement theories that were allegedly first introduced in Sonos’s expert reports. First, Google argues that Sonos’s infringement theories based on the “PlaylistService” and “BigTable” should be struck (Google Br. 7–9). Specifically, Google asserts that Sonos’s infringement contentions did not disclose that the accused remote playback queue is stored on a cloud server. It points to language in Sonos’s infringement contentions that identifies the remote playback queue as “a ‘Watch Next’ queue[] provided by one or more cloud servers” (Google Br. 7 (quoting Kaplan Exh. 1 at 4)) (emphasis added). In other words, according to Google, the Watch Next queue is a “list of media identifiers contained in the Watch Next message that is sent to a YouTube application, not a queue that is stored in the cloud” (Google Reply Br. 6). As such, disclosure of where and how the Watch Next queue is stored in Dr. Schmidt’s opening and reply reports — “by the ‘PlaylistService’ in a distributed storage system referred to as ‘BigTable’” — was purportedly improper because Google was not put on “has long understood that Sonos accuses a playback queue in the cloud” rather than a message providing the media item identifiers from the cloud, and that Google had reasonable notice about the relevance of PlaylistService (Sonos Opp. 9). This order agrees. Sonos points to ample statements and documentation demonstrating Google knew that Sonos was accusing a queue stored in the cloud (Sonos Opp. 9–11). This includes Dr. Bhattacharjee’s opening report, filed the same day as Dr. Schmidt’s opening report (Richter Exh. 7). In his opening report, Dr. Bhattacharjee explained that Sonos “contends that the term ‘remote playback queue’ covers a ‘playback queue’ that is ‘in the cloud’ (e.g., a cloud queue),” and that Sonos’s infringement contentions “identify[] the ‘remote playback queue’ as ‘a “Watch Next” queue’ with ‘recommended videos’ that is provided by ‘one or more cloud servers’” (Richter Exh. 7 ¶ 87)). Moreover, Sonos points to ample statements and documentation demonstrating PlaylistService was sufficiently disclosed (Sonos Opp. 11–13). This includes a reproduced excerpt from a Google document that explains how “[t]he PlaylistDocumentService provides a representation of the queue” and “[i]t calls into the Playlist Service for video IDs in the playlist” (Sonos Opp. 11 (quoting Richter Exhs. 5, 15)). Sonos even provided a source code trace with citations for the code responsible for calls to the PlaylistDocumentService and PlaylistService (Sonos Opp. 11 (citing Richter Exh. 5 at 24)). This is more than adequate. In reply, Google notes that Sonos has provided “shifting (and inconsistent) infringement theories regarding the term ‘remote playback queue’ throughout this litigation” and “these subsequent inconsistent positions do not change Sonos’s contentions” (Google Reply Br. 7–8). But Sonos amended its infringement contentions with leave of the Court and should not be faulted for doing so. True, Sonos did not provide a source code trace for PlaylistService specifically, but “[t]hat a particular document or source code file was not cited in a party’s infringement disclosures does not automatically preclude the party from using that document or file to support a theory that was timely disclosed.” See Oracle, 2011 WL 4479305, *3 (emphasis in original). Sonos was not required to identify every evidentiary item of proof — or specify that the overarching distributed storage system happened to be called BigTable. The theory that Sonos accused a playback queue stored in the cloud was timely disclosed. Second, as alluded to above, Google argues that Sonos’s infringement theory for Google Hub devices should be struck (Google Br. 9–11). According to Google, Sonos’s Hub-based infringement contentions were limited to a “single playback path” involving a user initiating active playback at the Google Hub device via input at the Google Hub device’s touchscreen display (Google Reply Br. 12; see also Google Br. 10). Meanwhile, Dr. Schmidt’s report introduces two new playback paths, the first involving a user initiating active playback at the Google Hub device via casting from a smartphone to the Google Hub device, and the second involving a user initiating active playback at the Google Hub device using voice (Google Br. 9–10). Sonos responds that Google has interpreted the limitation 1.4 of the ’033 patent to require the computing device to be actively playing back to be satisfied, which is why it spoke to these playback paths at all (Sonos Opp. 17). But in any event, according to Sonos, its contentions did not provide for only one playback path. Again, this order agrees. Nothing in Sonos’s infringement contentions limited Sonos’s theory of infringement to a single “playback path” (Google’s term). And surely it was not a secret to Google that playback on its devices could be initiated via casting and voice commands. Sonos’s contentions cite to one of Google’s own support pages, which explained that “[w]ith YouTube built-in to your Google Nest display,” “you can use voice commands like ‘Play [name of video]’” (Sonos Opp. 18 (citing Richter Exh. 5 at 8)). Here, no new theory of infringement was disclosed. This was simply additional explanation invited by Google’s interpretation of active playback. Accordingly, this order declines to strike material related to “PlaylistService,” “BigTable,” and Google Hub devices. 6. GOOGLE’S MOTION: NEW DOCTRINE OF EQUIVALENTS THEORIES. Finally, Google argues that Dr. Schmidt introduced two new doctrine of equivalents theories in his reply report related to select limitations of the ’033 patent (Google Br. 11–12). According to Google, Dr. Schmidt improperly argues that these limitations are met under the songs or videos in the list] provided by the YouTube cloud infrastructure is the same or different before and after Casting”; and (2) “irrespective of whether the list of one or more media items selected for playback is stored in the same or a different location in the YouTube cloud infrastructure before and after Casting” (Google Br. 11 (quoting Kaplan Exh. 5 ¶¶ 179– 88)). Google asserts that neither of these arguments was included in Sonos’s infringement contentions and there is no justification for Sonos to have waited until Dr. Schmidt’s reply report to present them (Google Br. 11–12). Sonos responds that Dr. Schmidt’s doctrine of equivalents opinions are directly responsive to non-infringement positions improperly advanced for the first time in Dr. Bhattacharjee’s non-infringement report — namely, that Sonos has failed to show that devices are not configured for playback of remote playback queues (Sonos Opp. 21–22). According to Sonos, these are the same arguments that Sonos challenged in its motion to strike. As such, “[i]f the Court strikes these new non-infringement positions from Dr. Bhattacharjee’s report, then Dr. Schmidt need not offer his doctrine of equivalents positions” (Sonos Opp. 22). Because this order did not strike those non- infringement positions, however, it declines to strike this responsive material. For the foregoing reasons, Sonos’s motion to strike is GRANTED IN PART and DENIED IN PART. This order strikes part of paragraph 280 of Dr. Bhattacharjee’s rebuttal expert report and paragraphs 178–81 of Dr. Schonfeld’s rebuttal expert report. Google’s motion to strike is DENIED in its entirety. On these motions alone, counsel submitted 135 pages of briefing, 1,130 pages of unsealed exhibits, and 3,756 pages of conditionally sealed exhibits and accoutrements — all to strike, in the end, fewer than five paragraphs from two expert reports. (Separately, on the summary judgment motions, counsel submitted 164 pages of briefing, 270 pages of unsealed exhibits, and 1,286 pages of conditionally sealed exhibits and accoutrements, plus 322 slides for the hearing.) Alas, much ink was spilled for little purpose. Most “new theories” proved to be 1 disclosed theories. It appears counsel moved to strike not out of prejudice but to secure an 2 advantage for summary judgment and trial, emblematic of the worst of patent litigation. 4 5 Dated: April 12, 2023. Pee 6 [ A = WILLIAM ALSUP 7 UNITED STATES DISTRICT JUDGE 8 9 10 11 12

© 15 16

= 17

Z 18 19 20 21 22 23 24 25 26 27 28