Beijing Meishe Network Technology Co., Ltd. v. TikTok Inc.

Opinion

BEIJING MEISHE NETWORK Case No. 23-cv-06012-SI TECHNOLOGY CO., LTD., Plaintiff, ORDER RE: DAUBERT MOTIONS v. Re: Dkt. Nos. 625, 627, 629, 632, 634, 636, TIKTOK INC., et al., 637, 640, 642 Defendants. Plaintiff filed six Daubert motions seeking to exclude expert testimony and defendants filed three such motions. The Court held a hearing on these motions on August 29, 2025. The Court’s determination as to each motion is detailed below. Federal Rule of Evidence 702 permits the introduction of expert testimony only if “(a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert’s opinion reflects a reliable application of the principles and methods to the facts of the case.” Fed. R. Evid. 702. “To qualify as an expert, a witness must have ‘knowledge, skill, experience, training, or education’ relevant to such evidence or fact in issue.” U.S. v. Chang, 207 F.3d 1169, 1172 (9th Cir. 2000) (citing Fed. R. Evid. 702). Expert testimony is admissible “only if it is both relevant and reliable.” Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 141 (1999). The proponent of the expert testimony has the burden of proving the proposed expert 1996). On December 1, 2023, Rule 702(d) was amended to “clarify and emphasize that expert testimony may not be admitted unless the proponent demonstrates to the court that it is more likely than not that the proffered testimony meets the admissibility requirements set forth in the rule.” Fed. R. Evid. 702 (Adv. Comm. Note, 2023). However, Rule 702 “should be applied with a ‘liberal thrust’ favoring admission.” Messick v. Novartis Pharm. Corp., 747 F.3d 1193, 1196 (9th Cir. 2014) (citing Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 588 (1993)). “Shaky but admissible evidence is to be attacked by cross examination, contrary evidence, and attention to the burden of proof, not exclusion.” Primiano v. Cook, 598 F.3d 558, 564 (9th Cir. 2010). Rule 702 requires that the trial court act as a “gatekeeper” by “making a preliminary determination that the expert’s testimony is reliable.” Elsayed Mukhtar v. Cal. State Univ., Hayward, 299 F.3d 1053, 1063 (9th Cir. 2002) (citations omitted); see Daubert, 509 U.S. at 597. The decision whether to admit or exclude expert testimony lies within the trial court’s discretion. See General Elec. Co. v. Joiner, 522 U.S. 136, 141-42 (1997); United States v. Calderon-Segura, 512 F.3d 1104, 1109 (9th Cir. 2008). The Ninth Circuit has acknowledged that Daubert “may be harder to apply when the expert testimony is ‘experience-based’ rather than ‘science-based,’” but an examination of reliability may be more important in considering “experience-based” opinion. United States v. Valencia-Lopez, 971 F.3d 891, 898 (9th Cir. 2020). Additionally, Rule 702 “makes no relevant distinction between ‘scientific’ knowledge and ‘technical’ or ‘other specialized’ knowledge.” Kumho Tire, 526 U.S. at 147. At the outset, the Court notes that all the challenged experts bring extensive qualifications to their role in this litigation. With one exception, the parties do not challenge the other side’s experts on a lack of qualifications or experience, but rather on reliance on insufficient factual bases or improper methodologies. I. Defendants’ Daubert Motions A. Yanfang Wang Defendants did not submit a formal Daubert motion to exclude the report of Chinese Judge Yanfang Wang. Nonetheless, in their reply brief in support of their motion for summary judgment, defendants raise for the first time an argument that the report should be excluded because Judge Wang did not sign her report under penalty of perjury. Dkt. No. 659-1. Federal Rule of Civil Procedure 56(c)(4) states that “An affidavit or declaration used to support or oppose a motion [for summary judgment] must be made on personal knowledge, set out facts that would be admissible in evidence, and show that the affiant or declarant is competent to testify on the matters stated.” The concerns of this rule are “functional”—ensuring the witness “is competent to testify to the conclusions and opinions in the report.” Am. Fed’n of Musicians of United States & Canada v. Paramount Pictures Corp., 903 F.3d 968, 977 (9th Cir. 2018). The Court will not exclude Judge Wang’s report on this basis, in particular when defendants gave plaintiff no opportunity to respond to this argument via a Daubert motion and when Judge Wang has provided deposition testimony under oath about her report. See Est. of Sanchez v. Cnty. of Stanislaus, No. 1:18-CV-00977-ADA- BAM, 2023 WL 7612399, at *8 (E.D. Cal. Nov. 14, 2023) (excusing the requirement for a signature under penalty of perjury when expert provided deposition testimony). B. Michael Shamos Plaintiff enlisted Dr. Michael Shamos to conduct a comparison of defendants’ and plaintiff’s source codes in an attempt to prove defendants’ alleged copying. Defendants challenge Dr. Shamos’ report and offered testimony on several grounds. Dkt. No. 638-1. First, defendants seek to limit Dr. Shamos’ testimony to the versions of defendants’ TikTok app that he specifically reviewed. During discovery, defendants made available for review ninety- eight versions of their code. Dr. Shamos claims that eighty-nine of the ninety-eight versions infringed plaintiff’s intellectual property. Dkt. No. 638-2 (Shamos Report) ¶ 92. Dr. Shamos’ report focuses on just three versions of TikTok code, which he considered “representative” examples. Id. has not provided methodology or explanation as to how he chose his representative samples. Plaintiff responds that Dr. Shamos does in fact give context as to why these three versions were selected, citing paragraphs eighty-two through ninety-two of the report. The Court will allow Dr. Shamos to testify to his opinion as to copying in versions beyond the three “representative” samples if plaintiff can lay a sufficient foundation during examination for such testimony. Second, defendants seek to exclude comments from Dr. Shamos that offer improper legal opinions. The Court agrees that Dr. Shamos may not call TikTok code a “derivative work” or label the two codes “substantially similar” as these are issues for the jury to decide. But Dr. Shamos may describe the similarities he has perceived and testify, in his opinion, how defendants’ code has been copied from plaintiff’s code. Defendants also seek to exclude Dr. Shamos’ testimony regarding the economic value of plaintiff’s trade secrets and the reasonable measures taken by plaintiff to protect the trade secrets. The Court holds that Dr. Shamos may not introduce evidence into the record about the protection measures plaintiff has taken, but he may comment about the adequacy of any measures that have been established in the record by fact witnesses if plaintiff can lay a foundation for his expertise in that arena. Further, Dr. Shamos’ assertion that plaintiff’s code “clearly has economic value” is too conclusory to be offered in that form. See Shamos Report ¶ 104. Third, defendants object to Dr. Shamos’ characterization of the trade secrets at issue in this case, arguing that he treated subcomponents of one trade secret as separate trade secrets. Plaintiff responds that Dr. Shamos’ report addresses the trade secrets in a manner consistent with plaintiff’s representations in discovery. Plaintiff disclaims any attempt to expand its claims at this stage of the case. Plaintiff’s response is sufficient to resolve this issue. Fourth, the parties rehash a dispute about source code comparisons that went through several iterations during discovery. Defendants repeatedly asked plaintiff to identify which parts of defendants’ code allegedly infringed on which parts of plaintiff’s code, leading to multiple versions of a comparison document known during the discovery process as “Exhibit D.” Ultimately, Exhibit D provided a high-level comparison, matching line groupings from defendants’ code with line groupings from plaintiff’s code. The exhibit did not include any specific text within those codes. The Special Master denied this sanction, allowing plaintiff to rely on the specific comparisons in Dr. Shamos’ report. Dkt. No. 576 at 34-35. However, the Special Master also then declared that the report “shall serve as the finite ‘universe’ of all alleged specific, line-by-line copying claimed by Plaintiff” and “Plaintiff may not later attempt to argue that Dr. Shamos’ report plus groupings of lines in Exhibit D (or plus some other document) somehow collectively assert all alleged instances of copying in this case.” Id. The Court noted this ruling in a subsequent order, stating, “The Special Master further clarified that plaintiff could not rely on source code included in Exhibit D but not the expert report.” See Dkt. No. 620 at 5 n.3. The parties are now at odds again because Dr. Shamos included the last version of Exhibit D as Exhibit D to his expert report. See Dkt. No. 663-7. Plaintiff argues that the Special Master did not intend to exclude Dr. Shamos from relying on any exhibits to his own report, including Exhibit D. Dkt. No. 663-1 at 12. The Special Master seemed to be aware that Dr. Shamos had “subsumed” Exhibit D into his report. See Dkt. No. 638-7 at 19. The Court will not prohibit Dr. Shamos from referencing Exhibit D to his expert report, but he may not use Exhibit D as a basis for comparing code. Dr. Shamos conceded in his deposition that Exhibit D is not a code comparison document. See Dkt. No. 638-4 at 433 (“[Exhibit D] doesn’t actually compare code. What it does is it lists files of code that contain infringing material.”). Dr. Shamos may speak to the creation of this exhibit and information that is readily discernible from the face of the exhibit, but may not offer any additional text-to-text comparisons at trial beyond the comparisons already offered in the other exhibits to his report. As explained above, the motion to exclude testimony from Dr. Shamos is GRANTED in part and DENIED in part. C. Sam Malek Plaintiff’s expert Dr. Sam Malek attempted to provide a technical apportionment of the value that plaintiff’s allegedly stolen intellectual property added to defendants’ applications. Dkt. No. 630-2 (Malek Opening Report). Dr. Malek did not analyze original source code but rather the “features and functionalities” of defendants’ TikTok and CapCut apps, in particular the audio/video “foundational” to defendants’ audio/video editing library. See Shamos Report ¶¶ 75-76. The parties do not dispute that one expert may rely on another expert; indeed, both parties rely on this principle at times. However, defendants challenge Dr. Malek’s approach, arguing that by comparing features and functionalities and user interface screens—not source code—his methodology is too divorced from the actual material at issue in this case. Defendants also claim that Dr. Malek relies on an incorrect assumption that all of the source code in defendants’ audio/video editing library infringes Meishe’s copyrights when Dr. Shamos found that only a small percentage of defendants’ source code was taken from Meishe. Finally, defendants argue that Dr. Malek only analyzed Android code and should not be able to transfer those conclusions to iOS code. All of defendants’ arguments go to the weight of Dr. Malek’s testimony, not its admissibility. Dr. Malek’s approach is rational if one assumes its basis is correct: Dr. Shamos’ conclusion that defendants’ audio/video editing library would not function without Meishe’s code. See Shamos Report ¶¶ 75-76. Defendants may challenge that assumption. Ultimately, it is defendants’ burden to apportion away any profit that is not attributable to infringement. See Oracle Am., Inc. v. Google Inc., No. C 10-03561 WHA, 2016 WL 2342365, at *7 (N.D. Cal. May 3, 2016). Separately, defendants contest the admissibility of Dr. Malek’s reply report regarding whether Meishe’s code could be reverse engineered, arguing that it contains material that should have been offered at the opening report stage. Dr. Malek’s reply report responds to an unchallenged report by defendants’ expert Anthony DeRosa, who attempted to reverse engineer plaintiff’s software. See Dkt. Nos. 666-3 (DeRosa Report), 666-4 (Malek Reply Report). It is undisputed that plaintiff bears the burden of proving that its trade secrets cannot be readily ascertained. Defendants argue that since plaintiff bears the burden on this issue, it may not introduce a new report at the reply stage.1 In opposition, plaintiff contends that Dr. Shamos’ opening report briefly discusses reverse 1 Defendants selectively cite to an exchange with the Court from December 6, 2024, where defendants argue the Court limited reply reports to matters introduced by defendants at rebuttal and where defendants must carry the burden. See Dkt. No. 627-12. Plaintiff contests defendants’ interpretation of the exchange. What is clear is that the Court added a reply date “on the engineering and Dr. Malek’s report is limited to discussing Mr. DeRosa’s reverse engineering experiment. Simply put, plaintiff argues that Dr. Malek’s reply report could not have been offered on opening because it only responds to Mr. DeRosa’s conclusions. The Court holds that Dr. Malek may offer testimony regarding Mr. DeRosa’s reverse engineering project. While this is an issue where plaintiff bears the burden, there can be no dispute that plaintiff could not challenge the approach of Mr. DeRosa before learning about it. To be clear, though, Dr. Malek is limited to a critique of Mr. DeRosa’s efforts—he may not offer any other testimony about whether Meishe’s trade secrets are readily ascertainable. Defendants’ motion to exclude Dr. Malek’s testimony is therefore DENIED. D. Jonathan Putnam Plaintiff’s expert Jonathan Putnam provided a calculation of the unjust enrichment defendants received based on their alleged infringement and misappropriation of plaintiff’s intellectual property. Dkt. Nos. 626-2 (Putnam Opening Report), 626-5 (Putnam Reply Report). Defendants seek to exclude the expert testimony of Jonathan Putnam for two reasons. Dkt. No. 626- 1. First, defendants argue that some data underlying the report, taken from the company Business of Apps, is inherently unreliable. Second, for his evaluation of damages from trade secret misappropriation, defendants argue that Dr. Putnam failed to apportion damages between the trade secrets, and that this failure renders the resulting analysis unreliable or unhelpful. To make his calculations, Dr. Putnam attempted to connect TikTok’s growth in new accounts with growth in monthly active users, then estimated the value of each new monthly active user. Putnam Opening Report ¶¶ 12-13. Dr. Putnam relied on Dr. Malek’s analysis to isolate how many new users could be attributed to TikTok’s alleged infringement and misappropriation. Id. ¶ 14. Dr. Putnam then applied a calculated damages figure of $6.37 per monthly active user, resulting in an estimate of $1.08 billion in unjust enrichment revenue for TikTok activity in the United States, and find ourselves spending – what is it – larded in papers, debating whether it in fact is in response to an additional $1.34 billion in excess revenue for activity outside of the United States. Id. ¶ 15. For the CapCut software, Dr. Putnam estimates the unjust enrichment to be $6.4 million. Id. ¶ 16. To determine TikTok’s user metrics, Dr. Putnam relied on several sources of data. From defendants, he had access to data on TikTok’s new account numbers from 2016, but the monthly active user data for United States users starts in August 2020, and for foreign users in May 2023. Id. ¶ 77. Since these data do not cover the entire damages period, Dr. Putnam used data from the Social App Report published by the company Business of Apps, which provides quarterly estimates of revenue and monthly active users for major social media companies. Id. ¶ 83; see also 661-4 at 61-62 (Business of Apps Report). That report is not clear on its face whether the quarterly monthly active user figure reflects (a) the average number of monthly users across the three months in the quarter or (b) the sum of monthly active users for the three months. If the average, the number of monthly active users will be higher, and the resulting calculation of revenue per monthly active user will be lower. Dr. Putnam assumed the figure provided was the sum of monthly active users across the quarter, resulting in fewer total monthly active users and a higher revenue-per-user calculation— and thus higher unjust enrichment. Defendants’ rebuttal expert Dr. David Teece challenged this interpretation. Dkt. No. 639-4 ¶¶ 277-280. After reviewing Dr. Teece’s report, Dr. Putnam’s team contacted the Business of Apps to confirm whether the figure provided was the average across the quarter or the sum. Dkt. No. 626-4 (Dep. Putnam) at 39-46. The Business of Apps reported that they simply received the data from the social media companies and that the submitting companies may interpret the data differently. Id. at 41. Facing this uncertainty, Dr. Putnam then conservatively switched from interpreting the figures as sums to averages, reducing his damages estimate by two- thirds in his reply report. Putnam Reply Report ¶ 85 n.70. Defendants argue that “the undisputed reality is that no one knows whether the Business of Apps data is accurate” and that “dressing up his revised damages calculation in a ‘cloak of conservativism’ does not solve the inherent unreliability of the data.” Dkt. No. 626-1 at 4, 6 (citations omitted). As such, they seek to strike Dr. Putnam’s entire report. Id. at 5. Plaintiff emphasizes that Dr. Putnam used the Business of Apps data to fill in gaps in data that he received forecast of revenue per daily active user into a comparable revenue per monthly active user figure, and that resulting figure was lower than Dr. Putnam’s original estimate, but twice as high as his later, more conservative estimate. Putnam Opening Report ¶ 118. The Court understands defendants’ concerns about the Business of Apps data, but the Court believes that these concerns go to the weight, not the admissibility, of Dr. Putnam’s report. Dr. Putnam clearly explained his process for arriving at his final estimates and a jury can evaluate how much weight his final estimates deserve. Defendants next attack Dr. Putnam’s report for failing to apportion damages among the separate trade secrets asserted by plaintiff. Plaintiff’s asserted trade secrets cover (1) the non-public source code for the four asserted versions of Meishe’s software (v1.0, v1.3 rev. 2631, v1.5, and v2.5.4) and (2) “Meishe’s non-public audio and video processing engine framework and workflow,” including sub-modules, as “embodied in Meishe’s non-public source code.” Dkt. No. 626-7 at 7-8. The parties dispute whether the case law requires plaintiff’s expert to apportion damages among individual trade secrets. In one of defendants’ cited cases, the jury found that the defendant misappropriated five of eleven trade secrets and awarded unjust enrichment damages for the misappropriation of only one of those. O2 Micro Int’l Ltd. v. Monolithic Power Sys., Inc., 399 F. Supp. 2d 1064, 1069 (N.D. Cal. 2005), amended on other grounds, 420 F. Supp. 2d 1070 (N.D. Cal. 2006). The court then granted the defendant’s motion for judgment as a matter of law because the expert testimony relied upon for the damages award did not provide a reasonable basis to calculate the damages for the misappropriation of the single trade secret. Id. at 1077. Similarly, a court determined that an expert report that did not apportion damages across twenty-eight asserted trade secrets was unhelpful, in particular because only fifteen of the trade secrets were at issue in the initial bellwether trial. LivePerson, Inc. v. [24]7.AI, Inc., No. 17-CV-01268-JST, 2018 WL 6257460, at *2 (N.D. Cal. Nov. 30, 2018). Here, the number of trade secrets is much smaller and the asserted trade secrets are so closely related that the Court does not consider it necessary to preemptively require trade secret-by-trade testimony is therefore DENIED.2 II. Plaintiff’s Daubert Motions A. Xiangjun Kong Defendants submitted a report from former Chinese Judge Xiangjun Kong to rebut evidence presented in the report of plaintiff’s expert, Judge Wang. Dkt. No. 610-3. In defendants’ summary, Judge Kong’s report makes the following arguments: (1) the record undermines Judge Wang’s conclusion that Xin Ao Te transferred intellectual property rights through an oral agreement; (2) even assuming the existence of an oral agreement, such agreement would not be valid in China; and (3) the October 2015 copyright transfer agreement does not support the transfer-by-oral- agreement theory. Dkt. No. 667-1 at 2-3. Plaintiff seeks to exclude this report in its entirety because Judge Kong did not review all the evidence and because defendants’ counsel helped draft his report. Neither argument persuades the Court that Judge Kong fails the reliability threshold of Federal Rule of Evidence 702. Rule 702 requires that an expert base his or her testimony on “sufficient facts or data.” Fed. R. Evid. 702(b). Plaintiff contends that Judge Kong failed to base his conclusions on sufficient evidence in two ways. First, Judge Kong did not review software licensing agreements between XAT and Meishe from 2018 and 2024, even though at least one Chinese court opinion referenced such agreements. Dkt. No. 641-7 (Kong Dep. Vol. 2) at 58-62. Judge Kong’s failure to consider these documents does not render the factual basis of his report insufficient. Second, plaintiff takes issue with Judge Kong’s testimony during deposition that, when asked whether he read the Chinese court opinions attached to Judge Wang’s report, he only said that he “read the relevant content.” Id. at 35. This statement does not make Judge Kong’s report unreliable.3

2 While not argued in their Daubert motion, defendants argue in their opposition to summary judgment brief that Dr. Putnam’s report is inadmissible for lack of an accompanying sworn declaration. Dkt. No. 649-1 at 12-13. For the same reasons as discussed above for Judge Wang’s report, the Court will not hold Dr. Putnam’s report inadmissible on this basis. 3 Plaintiff also argues that Judge Kong “should not be permitted to opine in contradiction to Plaintiff also seek exclusion of Judge Kong’s report because defendants’ counsel assisted in preparation of the draft. Counsel may assist in the drafting of an expert report. Fed. R. Civ. P. 26 (Adv. Comm. Note, 1993). In determining whether counsel has helped too much, “[t]he key question is ‘whether counsel’s participation so exceeds the bounds of legitimate assistance as to negate the possibility that the expert actually prepared his own report.’” NetFuel, Inc. v. Cisco Sys. Inc., No. 5:18-CV-02352-EJD, 2020 WL 1274985, at *2 (N.D. Cal. Mar. 17, 2020) (quoting Bekaert Corp. v. City of Dyersburg, 256 F.R.D. 573, 578 (W.D. Tenn. 2009)). In other words, the opinions expressed must be those of the expert. Id. The Court holds that Judge Kong did not improperly rely on the assistance of defendants’ counsel. As Judge Kong testified in his deposition, he provided his opinions to the attorneys, the attorneys wrote a draft for his review, he provided some modifications, and then they formulated a final draft. Dkt. No. 641-7 (Kong Dep. Vol. 2) at 16. Judge Kong thus provided substantive input from beginning to end and, importantly, he testified that the report “contained all of my opinions and viewpoints.” Id. Judge Kong did not relinquish his expert role to the attorneys, and his report shall not be excluded on that basis. Plaintiff’s motion to exclude Judge Kong’s report is DENIED. B. Nigel Jones Defendants retained Nigel Jones to analyze the allegedly copied code and respond to the reports from Dr. Shamos and Dr. Malek. Dkt. No. 631-4 (Jones Report) ¶ 8. Plaintiff asserts that Mr. Jones’ report is unreliable for several reasons. First, plaintiff argues that Mr. Jones inappropriately focused on a sampling of plaintiff’s allegedly copied code, rather than its entirety. Plaintiff suggests that Mr. Jones should have examined the files listed in Exhibit D to Dr. Shamos’ report. Mr. Jones provided an analysis of Exhibit D and found it difficult to interpret on its face. See, e.g., Jones Report ¶¶ 1314, 1376. The

The Court is not concerned with whether Judge Kong is right or wrong, but whether a fact-finder might consider his testimony helpful. See Alaska Rent-A-Car, Inc. v. Avis Budget Grp., Inc., 738 Court will not hold Mr. Jones responsible for failing to provide a detailed analysis of the lines listed in Exhibit D when plaintiff’s experts avoided that task as well. And for the sampling that Mr. Jones conducted, he provided an explanation of his methodology. See, e.g., ¶ 704. Defendants explain how Mr. Jones tried to pull sample data in several ways. Dkt. No. 675-1 at 5-6. The Court finds that Mr. Jones’ methods can be understood and evaluated by a jury. Plaintiff’s concerns about selective sampling may be explored through cross-examination. Second, plaintiff takes issue with how Mr. Jones analyzed the protectability of individual code elements, arguing in essence that his approach was too granular. Plaintiff’s expert Dr. Shamos provides specific examples in a reply report detailing how Mr. Jones’ approach may be unhelpful. See, e.g., Dkt. No. 680-13 (Shamos Reply Report) ¶ 24. The parties’ briefing on the issue of copyrightability also addresses this issue at length. For purposes of plaintiff’s Daubert motion, the Court again concludes that Mr. Jones’ approach does not disqualify his testimony. Plaintiff’s concerns go to weight, not admissibility. Third, the parties dispute the proper way to measure the quantitative significance of defendants’ alleged infringement. Mr. Jones’ formula divides the number of copied lines by the number of lines of Meishe’s entire code, including material Meishe did not designate as its own work. Parsing the arguments, the parties appear to agree that the denominator in the equation should be the plaintiff’s “work as a whole.” See Newton v. Diamond, 388 F.3d 1189, 1195 (9th Cir. 2004). The disagreement focuses on whether plaintiff’s copyrighted work includes third-party or open- source code. Defendants point to Meishe’s 2025 description of its asserted copyrights where, for example, plaintiff stated that “v1.0 is copyrighted by Meishe in its entirety.” Dkt. No. 675-6. Plaintiff notes that, in a 2023 response to defendants’ interrogatories, it did not identify third-party or open-source code as part of its asserted works. See Dkt. Nos. 697-5, 697-8. Mr. Jones’ report contains tables that alternatively exclude and include this additional code,4 providing a result for each. Jones Report ¶¶ 1380-81. The parties can further explore this distinction at trial, but the report 4 Defendants note that the additional “missing” code for v1.0 also includes some files Meishe is not inadmissible on this basis. Fourth, plaintiff insists that Mr. Jones brings no expertise to bear on the question of Xie Jing’s access to Meishe’s code. Defendants respond that Mr. Jones’ experience with the program Subversion will help the jury understand the evidence from Subversion program logs. Mr. Jones may testify about how to interpret particular log entries but may not comment on how these entries align with or contradict other evidence in the case, as such testimony falls beyond the province of expert testimony. Fifth, plaintiff argues that Mr. Jones should not be allowed to testify about other software versions plaintiff previously asserted but has since dropped from the lawsuit. Defendants respond that these dropped versions are relevant as to the question of defendants’ willfulness. The Court is not persuaded. Mr. Jones may not offer testimony about versions not at issue because doing so would needlessly confuse the jury. Sixth, a section of Mr. Jones’ report compares Meishe’s asserted source codes with the source code submitted to the Copyright Protection Center of China for what became copyright 2015SR227927, the one registered copyright at issue in the case. Mr. Jones writes that plaintiff has not produced code that matches the code submitted with the registration. Jones Report ¶ 241. Plaintiff argues that Mr. Jones’ opinions improperly challenge the validity of the Chinese registration and are irrelevant. Plaintiff is incorrect. Mr. Jones is not challenging the validity of the Chinese copyright, but attempting to discern which code version corresponds to the copyright. The Court finds that these opinions are not irrelevant since they provide insight into the content of the one registered copyright at issue, even though they have no bearing on Meishe’s asserted unregistered copyrights. Seventh, plaintiff seeks to exclude what it construes as legal conclusions at paragraph 1390 in Mr. Jones’ report. Defendants concede that Mr. Jones cannot offer a legal conclusion. The Court holds that Mr. Jones may not state opinions on the legal questions facing the jury, such as whether the codes are substantially similar or whether defendants are entitled to a fair use defense. Finally, plaintiff argues that Mr. Jones improperly analyzes the third factor in a fair use the copyrighted work as a whole.” 17 U.S.C. § 107(3). Specifically, plaintiff argues that “the portion used” language calls for dividing the copied code by the work as a whole, not the smaller subset of copied copyrightable code by the work as a whole. Defendants cite out-of-circuit authority to the contrary. Dkt. No. 675-1 at 16 (citing Leibovitz v. Paramount Pictures Corp., 137 F.3d 109, 115 (2d Cir. 1998)). The Court considers this a close question but, based on the text of the statute, finds that “the portion used” is not limited to protectable expression. The Court adopts plaintiff’s interpretation. For the reasons discussed above, plaintiff’s motion is GRANTED in part and DENIED in part. C. Kevin Almeroth Defendants enlisted Kevin Almeroth to provide an opinion rebutting Dr. Malek’s apportionment analysis and the analysis of Dr. Shamos regarding trade secret misappropriation. Dkt. No. 628-3 (Almeroth Report) ¶ 29. Plaintiff seeks to exclude certain of Mr. Almeroth’s opinions because plaintiff argues he applied an incorrect legal standard for misappropriation. As with all experts, Mr. Almeroth may not offer testimony that amounts to a legal conclusion. See United States v. Tamman, 782 F.3d 543, 552 (9th Cir. 2015). The Court will permit Mr. Almeroth to testify as to the details of his analysis of any trade secret misappropriation or lack thereof, but he may not opine as to the legal standard nor may he provide a legal conclusion. The Court and the parties will determine the correct standard for the jury to consider through the submission of jury instructions. Accordingly, the Court GRANTS in part and DENIES in part plaintiff’s motion to exclude Mr. Almeroth’s testimony. D. John Kelly Defendants hired Dr. John Kelly to run a “clean room” project “to provide an opinion, based on [his] experience and expertise, on how long and how much it would cost to create a version of a application in a ‘cleanroom’ setting.” Dkt. No. 633-3 (Kelly Report) ¶ 38. The project lasted almost nine months and involved twenty-two Chinese engineers overseen by Dr. Kelly. Id. ¶¶ 113-14. In that time, Dr. Kelly estimated that the team completed more than 80% of the video editing features and could have finished the project in less than two months. Id. ¶¶ 115-116. Defendants’ damages expert, Dr. Teece, considers Dr. Kelly’s estimate of the cost to develop a fully functioning software kit in Dr. Teece’s reasonable royalty analysis. Dkt. No. 639-4 ¶¶ 432-34. A clean room must not be contaminated by any material the project is attempting to replicate and plaintiff argues that Dr. Kelly’s project was tainted and unreliable. Plaintiff’s arguments include the following: (1) Dr. Kelly did not review the Meishe code, so he did not know what to keep out of the room; (2) the version that Dr. Kelly’s team started with could have been contaminated by Meishe code because it was developed by engineers supervised by Xie Jing, the former Meishe employee accused of stealing Meishe’s code; (3) Dr. Kelly did not properly control communications between the clean room team and outside engineers; and (4) the clean room team incorporated ByteDance functionalities developed outside of the clean room. Defendants counter that plaintiff had access to all the clean room materials and communications yet came up with no evidence of actual contamination. Plaintiff may challenge Dr. Kelly’s process at trial, but its concerns go to the weight of his testimony, not its admissibility. Separately, plaintiff argues that Dr. Kelly’s project is not being offered for an appropriate purpose. Plaintiff emphasizes that Dr. Kelly’s report states that his goal was not to fully develop a replacement software, so plaintiff argues that his report cannot be used to establish the cost of a non- infringing alternative. But plaintiff ignores that although Dr. Kelly did not intend to fully develop a replacement software, the explicit goal was to ascertain how much it would cost to do so. By his calculations, Dr. Kelly reached 80% completion, from which he provided an estimate of the cost for 100% completion. Again, plaintiff may attack the credibility of these estimates, but plaintiff’s concerns do not render Dr. Kelly’s report inadmissible. Plaintiff’s motion to exclude Dr. Kelly is therefore DENIED. E. James Pooley Defendants have retained James Pooley as an expert to rebut Dr. Shamos’ testimony that plaintiff took reasonable measures to protect its trade secrets. Dkt. No. 635-3 (Pooley Report) ¶ 10. Plaintiff frames Mr. Pooley’s expertise as legal in nature and seeks to exclude his testimony as impermissibly offering legal instructions and conclusions to the jury. Defendants highlight Mr. Pooley’s experience consulting with companies as to how to protect trade secrets and note how other courts have allowed his testimony regarding protective measures. Plaintiff’s reply brief argues that Mr. Pooley does not have specific experience in how to protect trade secret source code. The Court will allow Mr. Pooley to testify based on his expertise as an information security consultant, not an attorney. Mr. Pooley may not offer testimony explaining the law of trade secrets, nor may he offer a conclusion that plaintiff did not take “reasonable measures.” If defendants can establish that Mr. Pooley’s consulting work includes relevant experience with software companies, Mr. Pooley may comment on measures he has advised clients to take and measures he would advise a company like Meishe to take. He may comment on what the evidence reveals about Meishe’s protection measures, and whether that aligns with his understanding of industry standards. He may not comment about the discovery history in this case and he may not simply repeat what defendants’ other experts have told him unless he has independent expertise through which he is evaluating the conclusion of the other experts. To be sure, plaintiff may challenge Mr. Pooley’s specific experience with source code protections and Chinese start-ups upon cross-examination. The Court thus DENIES in part and GRANTS in part plaintiff’s motion. F. David Teece Defendants offer Dr. David Teece as a rebuttal expert to Dr. Putnam’s opinion regarding damages. While Dr. Putnam offers a theory of damages based on unjust enrichment, Dr. Teece’s report criticizes Dr. Putnam’s methodology and focuses on a reasonable royalty analysis as an alternative method for calculating damages. Plaintiff seeks to exclude several sections of Dr. Teece’s report. See Dkt. No. 639-1. knowledge about the methodology behind the econometric calculations in his report. Defendants respond by noting that a deposition “is not a memory test.” Dkt. No. 673-1 at 3. The Court understands plaintiff’s concern to be whether Dr. Teece can adequately explain the tests applied to Dr. Putnam’s analysis in order to place his conclusions in a proper context. The Court will allow Dr. Teece to testify about these tests, but he must be able to adequately explain how the tests work before he can share any outcomes with the jury. Next, plaintiff challenges Dr. Teece’s calculation of incremental costs, which by law can be applied to offset incremental profits earned through unjust enrichment. Plaintiff contends Dr. Teece has inappropriately included costs not associated with the incremental gains in users associated with defendants’ alleged infringement. Plaintiff’s concerns in this regard go to the weight, not admissibility, of Dr. Teece’s testimony. Plaintiff may challenge Dr. Teece’s calculations on cross-examination. Plaintiff also seeks to exclude portions of Dr. Teece’s testimony where he shares his opinion about the reports of Dr. Shamos and Dr. Malek. Dr. Teece admits in his report that he is not a technical expert and that he relied on defendants’ technical experts for the section of his report labeled “Technical Understanding.” Dkt. No. 639-3 ¶ 112. Plaintiff argues Dr. Teece is not relying on others but has offered his own opinions. Dr. Teece gives blanket statements about his reliance on others, but certain portions of his report could be pulled and read as his own improper opinions. At trial, Dr. Teece may not offer his own opinions on technical matters for which he has expressed no expertise, but he may rely on other expert testimony admitted into evidence as long as that reliance is explicitly stated. Plaintiff also challenges Dr. Teece’s report to the extent it relies on material challenged by plaintiff in motions related to plaintiff’s challenges to the Jones and Kelly expert reports. Those challenges are addressed directly in the discussion of those reports above. The Court DENIES in part and GRANTS in part plaintiff’s motion. 1 offered testimony of experts Malek, Putnam, Kong, and Kelly are DENIED. Dkt Nos. 625, 627, 2 634, 642. The motions to exclude the reports and offered testimony of experts Shamos, Jones, 3 Almeroth, Pooley, and Teece are DENIED in part and GRANTED in part, as explained above. Dkt. 4 Nos. 629, 632, 636, 637, 640. 5 The trial remains scheduled to start on October 27, 2025. The pretrial conference is 6 scheduled for October 14, 2025. 7 9 Dated: September 2, 2025 Site WU tee 10 SUSAN ILLSTON 11 United States District Judge a 12

13 14

15 16 € = 17 6 Zz 18 19 20 21 22 23 24 25 26 27 28