Giuseppe Pampena v. Elon R.Musk — Entry #590: ORDER on Post-Trial Motions
Case: Giuseppe Pampena v. Elon R.Musk cand · 3:22-cv-05937
filed October 10, 2022
What this document is
Docket entry #590 · filed July 06, 2026
ORDER on Post-Trial Motions. Signed by Judge Charles R. Breyer on July 6, 2026. (crblc2, COURT STAFF) (Filed on 7/6/2026) (Entered: 07/06/2026)
Who is involved
- Alex Spiro
- Brian Belgrave
- Elon Musk
- Giuseppe Pampena
- John C. Coates
- John Garrett
- Mohammed Samara
- Nancy Price
- Steve Garrett
- Teacher Retirement System of Texas
- The Twitter Investor Group
- X CORP.
- Yoel Roth
Why we have it
We follow this case because it names a company we track, although that company is not a party:
- Marathon Digital: its name “Marathon Digital Holdings, Inc.” appears in a filing in this case.
…Bitcoin to raise money, firms like Marathon Digital Holdings Inc. are selling Bitcoin call options…
A free copy from the RECAP archive of federal court filings (mirrored at the Internet Archive), retrieved November 02, 2026. Federal court filings are public records.
Document text
38 page(s), 143,681 characters, converted from the PDF's text layer · plain text.
Full text
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 1 of 38
1
2
3
4
5 IN THE UNITED STATES DISTRICT COURT
6 FOR THE NORTHERN DISTRICT OF CALIFORNIA
7
8 GIUSEPPE PAMPENA, et al., Case No. 22-cv-05937-CRB
9 Plaintiffs,
ORDER ON POST-TRIAL
10 v. MOTIONS
11 ELON MUSK,
12 Defendant.
United States District Court
13 Buyer’s remorse is not an exception to the securities laws. These laws, in their
14 essence, are about trust. Trust that the country’s financial markets are fair, honest, and
15 transparent. The market always picks winners and losers, but it is vital that those
Northern District of California
16 outcomes are free of manipulation. The Securities Exchange Act of 1934 seeks to protect
17 public investors—big and small—by ensuring a level playing field. By insisting that
18 public statements which could affect the market be truthful, Congress sought to guarantee
19 that all members of the investing public have equal access to information and, most
20 importantly, that the information be accurate. A speaker’s motive is not dispositive; what
21 matters is whether the statements were made with knowledge of their falsity or with a
22 deliberately reckless disregard for their accuracy. Even if the speaker has a change of
23 heart or a momentary regret about a transaction, such qualms do not justify lying to the
24 investing public.
25 This is a case about a breach of trust by Elon Musk. In April 2022, Musk sought to
26 buy Twitter, Inc., and take it private. He offered to buy the company at a significant
27 premium, $54.20 per share or about $44 billion total. Then, in order to sweeten the deal,
28 Musk offered to waive due diligence, making the agreement “seller-friendly.” Twitter
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 2 of 38
1 accepted the offer. The deal was publicly announced on April 25, 2022.
2 But Musk started having second thoughts. The purchase price was high and the
3 Tesla stock that he planned to use for financing was taking a hit, meaning he might have to
4 sell even more of it to compensate. So on May 13, despite waiving due diligence, Musk
5 tweeted that the deal was temporarily on hold until he received more details on how
6 Twitter calculated the level of bots/spam on its platform. Despite Musk’s public
7 statement, the deal was not on hold and Musk had been aware of bot issues on Twitter
8 since before he made his tender offer. The Twitter share price dropped immediately and
9 significantly. On May 17, Musk made a similar tweet, asserting that the number of bots on
10 Twitter was much higher than Twitter was representing and that the deal could not move
11 forward until he received more information. The Twitter share price remained low until
12 October 4, when Musk announced that he would go through with the deal. On that day,
United States District Court
13 the price spiked.
14 Plaintiffs sued. They represent all sellers of Twitter stock between May 13 and
15 October 4. They brought claims based on Musk’s two tweets, a separate statement made
Northern District of California
16 on May 16, and a purported months-long scheme by Musk to artificially depress Twitter’s
17 share price. After a two-week trial and more than three days of deliberations, a jury
18 unanimously found that Musk’s two tweets were false, and that the investing public was
19 damaged.
20 Musk now seeks to set aside this verdict and moves for a judgment as a matter of
21 law or, alternatively, for a new trial. JMOL (dkt. 570). He also asks this Court to decertify
22 the class based on the general nature of the verdict. Mot. to Decertify (dkt. 571). Plaintiffs
23 bring post-trial motions of their own. They seek an award of prejudgment interest and
24 approval of their class notice of the verdict and proposed claims administration process.
25 MPI (dkt. 565); MFA (dkt. 567). As explained below, the Court GRANTS Musk’s
26 motion for a judgment as a matter of law as to his May 17 tweet and DENIES it in all
27 other respects. The Court also DENIES his motion to decertify the class. The Court
28 GRANTS Plaintiffs’ motion for prejudgment interest and motion for approval of their
2
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 3 of 38
1 proposed class notice and claims administration procedure.
2 I. MOTION FOR JUDGMENT AS A MATTER OF LAW OR, IN THE
ALTERNATIVE, A NEW TRIAL
3
Musk argues that Plaintiffs, as a matter of law, failed to prove (1) loss causation, (2)
4
material falsity, (3) scienter, (4) reliance, and (5) damages. JMOL at 2–4. He also asserts
5
that he is entitled to a new trial, in the alternative. The Court GRANTS the motion only as
6
to loss causation for Musk’s May 17 tweet. The Court addresses each argument in turn.
7
A. Legal Standard
8
A motion for judgment as a matter of law is appropriate where “a reasonable jury
9
would not have a legally sufficient evidentiary basis to find for the [nonmoving] party on
10
that issue.” Fed. R. Civ. P. 50(a)(1). A court can grant a judgment as a matter of law “if
11
the evidence, construed in the light most favorable to the nonmoving party, permits only
12
United States District Court
one reasonable conclusion, and that conclusion is contrary to the jury’s verdict.” Pavao v.
13
Pagay, 307 F.3d 915 918 (9th Cir. 2002). A court must uphold a jury’s verdict if “it is
14
supported by substantial evidence that is adequate to support the jury’s findings, even if
15
contrary findings are also possible.” Escriba v. Foster Poultry Farms, Inc., 743 F. 3d 1236,
Northern District of California
16
1242 (9th Cir. 2014). A court must review the evidence as a whole and “disregard
17
evidence favorable to the moving party that the jury is not required to believe,” and the
18
court “may not substitute its view of the evidence for that of the jury.” Johnson v. Paradise
19
Valley Unified Sch. Dist., 251 F.3d 1222, 1227 (9th Cir. 2001) (internal quotations
20
omitted).
21
Under Rule 59(a)(1)(A) of the Federal Rules of Civil Procedure, a trial court may
22
grant a new trial after a jury trial “for any reason for which a new trial has heretofore been
23
granted in an action at law in federal court.” Specifically, a court may grant a new trial
24
“only if the verdict is contrary to the clear weight of the evidence, is based upon false or
25
perjurious evidence, or to prevent a miscarriage of justice.” Shimko v. Guenther, 505 F.3d
26
987, 993 (9th Cir. 2007) (internal quotations omitted). Such a motion may be granted on
27
insufficiency of evidence grounds “only if the verdict is against the great weight of the
28
3
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 4 of 38
1 evidence, or it is quite clear that the jury has reached a seriously erroneous result.” Incalza
2 v. Fendi N. Am., Inc., 479 F.3d 1005, 1013 (9th Cir. 2007) (internal quotation marks
3 omitted).
4 B. Loss Causation
5 Musk challenges the sufficiency of the evidence that his May 13 and 17 tweets
6 caused the class to suffer a loss in the market. JMOL at 14. For the May 13 tweet, Musk
7 asserts that Plaintiffs failed to prove (1) proximate causation, (2) that the market perceived
8 his statements as true, (3) that there was a corrective disclosure, and (4) that the
9 misrepresentation caused loss after his July 8 termination letter. Id. at 15–19. Plaintiffs
10 respond that there was sufficient evidence of proximate causation and corrective
11 disclosure. JMOL Opp’n at 8–10. Musk also argues that there was no market reaction to
12 the May 17 tweet, so Plaintiffs had no evidence of loss. JMOL at 15. Plaintiffs counter
United States District Court
13 that the May 17 tweet helped maintain the price deflation caused by the May 13 tweet.
14 JMOL Opp’n at 7–8. The Court agrees with Musk as to the May 17 tweet.
15 1. May 13 Tweet
Northern District of California
16 a. Proximate Causation
17 “Loss causation is shorthand for the requirement that investors must demonstrate
18 that the defendant’s deceptive conduct caused their claimed economic loss.” Lloyd v.
19 CVB Fin. Corp., 811 F.3d 1200, 1209 (9th Cir. 2016) (citation modified). It is “simply a
20 variant of proximate cause” and “the ultimate issue is whether the defendant’s
21 misstatement, as opposed to some other fact, foreseeably caused the plaintiff’s loss.” Id. at
22 1210. “A plaintiff is not required to show ‘that a misrepresentation was the sole reason for
23 the investment’s decline in value’ to establish loss causation. ‘[A]s long as the
24 misrepresentation is one substantial cause of the investment’s decline in value, other
25 contributing forces will not bar recovery under the loss causation requirement.’” Nuveen
26 Mun. High Income Opportunity Fund v. City of Alameda, 730 F.3d 1111, 1119 (9th Cir.
27 2013) (emphasis in original) (quoting In re Daou Sys., 411 F.3d 1006, 1025 (9th Cir.
28 2005)).
4
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 5 of 38
1 “Typically, in securities cases, loss causation is shown by the drop in share price
2 due to a misstatement or omission by the defendant.” See Edenbrook Cap., LLC v.
3 RhythmOne Plc, No. 19-CV-00615-WHO, 2019 WL 1791419, at *8 (N.D. Cal. Apr. 24,
4 2019). Here, Dr. Tabak testified that there was direct evidence that the May 13 tweet
5 caused an immediate, significant drop in Twitter’s share price. Trial Tr. at 1865:12–
6 1866:6 (Tabak). Musk argues that there was no evidence that the false statement—that the
7 deal was temporarily on hold—caused the impact on the share price. JMOL at 15. Not so.
8 Dr. Tabak testified that he evaluated news stories and analyst reports about the tweet and
9 that they were “generally in accord” that Musk’s “declaring the deal temporarily on
10 hold[]” was causing the price decline.1 Trial Tr. at 1866:12–1867:7.
11 b. Market Perception
12 Next, Musk asserts that there was insufficient evidence of loss causation because
United States District Court
13 the market knew the truth that the deal was not actually on hold. JMOL at 16. He points
14 to Dr. Tabak’s testimony that everyone knew that “Twitter did not agree to put the deal on
15 hold” and there were comments “that you can’t put it on hold.” Id. (quoting Trial Tr. at
Northern District of California
16 1908:23–1909:19). But as Plaintiffs note, Twitter’s public statements had nothing to do
17 with Musk and his position on the merger. See JMOL Opp’n at 9. To the public, it could
18 have been entirely possible that Musk was unilaterally holding up the deal. Musk also
19 notes that there was no statistically significant price reaction to his later comments about
20 working on his deal financing, which he argues supports his position that the market knew
21 the deal was not on hold. JMOL at 17. That is certainly evidence that cuts against
22 Plaintiffs. But it does not necessarily mean that the deal generally was not continuing as
23 usual. See Pavao, 307 F.3d at 918 (courts must construe evidence “in the light most
24
1
25 Musk also argues that Dr. Tabak did not distinguish any price impact from the true information
that Musk had concerns about the deal. JMOL at 18. But Dr. Tabak testified that his investigation
26 showed that the literal falsity of Musk’s tweet was what the news was reporting. Moreover, Dr.
Tabak explained that Musk’s heightened concerns could potentially play a part in the market’s
27 interpretation. Trial Tr. at 1905:4–13. But, he continued, Musk “didn’t just imply that he had
heightened concern; he suggested he was taking a certain action by temporarily putting the deal on
28 hold.” Id.; see also Nuveen, 730 F.3d at 1119 (misrepresentations need not be the sole reason for
a decline in share price).
5
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 6 of 38
1 favorable to the nonmoving party”). And it does not render Plaintiffs’ other evidence—as
2 discussed above—about market perception insubstantial. Castro v. Cnty. of Los Angeles,
3 833 F.3d 1060, 1066 (9th Cir. 2016) (courts must uphold verdicts if there is “evidence
4 adequate to support the jury’s conclusion, even if it is also possible to draw a contrary
5 conclusion”).
6 Musk also argues that Dr. Tabak did not disaggregate non-fraud factors that could
7 have impacted Twitter’s price drop, such as the May 15 publication of the SparkToro
8 report about bots on Twitter. JMOL at 18. Additionally, he posits that Dr. Tabak’s
9 decision to “attribute the May 16 stock market drop to the purported fraud” was defective.
10 Id. Neither argument persuades. As to the SparkToro report, there was no evidence that it
11 affected Twitter’s share price. Indeed, Dr. Tabak testified that he did not “see it in any
12 news stories or any analyst reports in that period.” Trial Tr. at 1893:19–24. And as
United States District Court
13 discussed below regarding damages, Dr. Tabak sought to exclude unrelated stock impacts
14 from his analysis. See infra Section I.F.
15 As to Dr. Tabak’s two-day damage calculation window for May 13 and 16, “there
Northern District of California
16 are many cases that find that multi-day event windows are appropriate for event study
17 analysis in securities fraud class actions.” Monroe Cnty. Employees’ Ret. Sys. v. S. Co.,
18 332 F.R.D. 370, 391 (N.D. Ga. 2019) (collecting cases); see Hsu v. Puma Biotechnology,
19 Inc., No. SACV1500865AGJCGX, 2018 WL 4956520, at *4 (C.D. Cal. Oct. 5, 2018) (“no
20 disqualifying problem” with two-day loss causation window); In re Vivendi Universal,
21 S.A. Sec. Litig., 634 F. Supp. 2d 352, 372 (S.D.N.Y. 2009) (three-day window). Dr.
22 Tabak explained that he concluded May 16 was a continuation of the May 13 decline
23 because, after market close, there are news stories, analyst reports, and individuals who do
24 not trade during the day that can affect the next trading day. Trial Tr. at 1896:15–20.
25 Although Musk relies on In re Omnicom Group, Inc. Securities Litigation to argue that
26 news of already-public information is insufficient for loss causation, the news in that case
27 was reporting facts that had been disclosed much earlier. See JMOL at 19 (citing In re
28 Omnicom Group, Inc. Securities Litigation, 597 F.3d 501, 505–06 (2d Cir. 2010) (negative
6
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 7 of 38
1 article describing events transpiring months earlier)). Here, the news Dr. Tabak reviewed
2 was between trading days. See Trial Tr. at 1896:15–20.
3 c. Corrective Disclosure
4 Musk contends that Plaintiffs failed to prove a corrective disclosure theory of loss
5 causation. JMOL at 19. Plaintiffs respond that they do not have to prove a corrective
6 disclosure in a seller case and that they did prove one, in any event. JMOL Opp’n at 9–11.
7 The Court agrees with Plaintiffs.
8 For starters, Musk is incorrect that the law requires Plaintiffs to prove a corrective
9 disclosure for loss causation. See Richardson v. TVIA, Inc., No. C 06 06304 RMW, 2007
10 WL 1129344, at *5 (N.D. Cal. Apr. 16, 2007) (“Other courts have similarly held that a
11 corrective disclosure is not necessary where the subject of the misrepresentation and
12 omissions caused the loss.”). Nevertheless, Musk argues that the Supreme Court’s logic in
United States District Court
13 Dura Pharmaceuticals, Inc. v. Broudo means a seller does not suffer a loss if they sell and
14 rebuy shares before a corrective disclosure. JMOL at 19. But Plaintiffs are right that, in a
15 seller case, loss is realized before a corrective disclosure. See JMOL Opp’n at 10.
Northern District of California
16 Dura merely stands for the proposition that “an inflated purchase price will not
17 itself constitute or proximately cause the relevant economic loss.” 544 U.S. 336, 342
18 (2005). The Supreme Court reasoned that a purchaser of an inflated price could sell shares
19 before the truth comes out, meaning “the misrepresentation will not have led to any loss.”
20 Id. In a seller case, by contrast, a seller suffers a loss at the moment of sale. The seller is
21 forced to sell at an artificially deflated price, immediately realizing their loss. And it
22 would be the misrepresentation that caused the artificial deflation in the first place. If the
23 seller were later to rebuy at the same price, it would be a different transaction entirely and
24 distinguishable from the principle espoused in Dura. See id. (“Shares are normally
25 purchased with an eye toward a later sale.”).
26 Regardless, Plaintiffs provided substantial evidence of a corrective disclosure on
27 October 4, 2022. Dr. Tabak testified that, when Musk filed his statement that he would go
28 through with the Twitter deal, the share price jumped over 20%, resulting in two trading
7
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 8 of 38
1 suspensions for the stock. See Trial Tr. at 1868:3–1869:22. Dr. Tabak concluded that
2 there was nothing other than Musk’s statement that could have accounted for the sharp
3 increase in price. Id. at 1869:23–1870:2 (Tabak) (“The only other news was basically
4 commentary reaction to the SEC filing.”). Musk asserts that Dr. Tabak’s testimony is still
5 not evidence of a corrective disclosure because Dr. Tabak said it was not part of his expert
6 opinion. JMOL Reply at 12. But this mischaracterizes the testimony—especially in the
7 light most favorable to Plaintiffs.
8 Musk’s counsel’s question on this point was very specific. He asked Dr. Tabak to
9 confirm that he was “not claiming that any information that came out was corrective of Mr.
10 Musk’s alleged false statement.” Trial Tr. at 1907:11–13 (emphasis added). Dr. Tabak
11 answered that it was “not part of [his] opinion.” Id. at 1907:14. Dr. Tabak did not say that
12 there was no corrective disclosure. He simply said that whether or not the disclosure was,
United States District Court
13 in fact, corrective was not a part of his opinion. See In re Vaxart, Inc. Sec. Litig., No. 20-
14 CV-05949-VC, 2025 WL 1869694, at *2 n.2 (N.D. Cal. July 7, 2025) (noting that what
15 constitutes a corrective disclosure is for the trier of fact).
Northern District of California
16 d. Post-July 8 Loss
17 Musk asserts that his July 8 termination letter was an intervening event and that
18 Plaintiffs did not have evidence of loss causation after that date. JMOL at 19–20. The
19 Court disagrees. As explained further regarding damages, Dr. Tabak calculated the loss
20 attributable to the May 13 tweet by creating a constant percentage of deflation derived
21 from the “merger discount” between Musk’s offered price per share and the actual stock
22 price. Trial Tr. at 1846:13–1847:3 (Tabak); see also infra Section I.F. And he made sure
23 to exclude additional deflation stemming from Musk’s letter from his analysis. See id. at
24 1849:20–1850:9. Nevertheless, Dr. Tabak concluded that there was still deflation
25 attributable to May 13 after July 8, even though it was a smaller percentage of the merger
26 discount. Id. at 1850:4–9. He testified that the deflation only vanished when Musk finally
27 announced he would go through with the deal on October 4. See Trial Tr. at 1868:3–
28 1869:22. Musk raises the fair point that Dr. Tabak did not analyze whether Twitter’s share
8
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 9 of 38
1 price would have fallen to the same level if Musk issued the termination letter but never
2 tweeted on May 13. JMOL at 20 (quoting Trial Tr. at 1918:25–1919:15). But although
3 that undermines Plaintiffs’ position, the argument does not mean the jury was without
4 sufficient evidence of loss causation after July 8. See In re Genius Brands Int’l, Inc. Sec.
5 Litig., 97 F.4th 1171, 1183 (9th Cir. 2024) (misrepresentations need not be the sole reason
6 for price decline).
7 2. May 17 Tweet
8 Musk argues that the lack of a market reaction to the May 17 tweet meant there was
9 no loss causation attributable to it. JMOL at 15. Plaintiffs point out that Dr. Tabak
10 testified that the May 17 tweet maintained the deflation from the May 13 tweet. JMOL
11 Opp’n at 7. The Court agrees with Musk.
12 Dr. Tabak did not provide an expert opinion on price maintenance. In fact, he
United States District Court
13 affirmatively disclaimed such an opinion. To be sure, he did testify that the May 17
14 “maintained the prior deflation.” Trial Tr. at 1881:14–20 (Tabak); id. at 1882:1–6 (“[I]f
15 the question that the jury wants to address is what would have happened if, you know,
Northern District of California
16 having spoken and said something incorrect, Mr. Musk should have corrected it, then the
17 answer should be that these maintained the deflation from May 13th.”). But Dr. Tabak
18 then clarified that he was “not affirmatively saying this is a price maintenance case.” Trial
19 Tr. at 1882:16–20; see also id. at 1882:14–15 (“Q. And you said you’re not offering that
20 opinion, right? A. If that’s what I said, that’s what I said, yes.”).
21 Without an expert opinion on price maintenance, Plaintiffs lack substantial evidence
22 to support a finding of loss causation with respect to the May 17 tweet. See In re REMEC
23 Inc. Sec. Litig., 702 F. Supp. 2d 1202, 1275 (S.D. Cal. 2010) (granting summary judgment
24 on loss causation as plaintiffs lacked expert testimony to prove it). Because Plaintiffs did
25 not have sufficient evidence for a key element of their case, the Court GRANTS Musk’s
26 motion for a judgment as a matter of law for the May 17 tweet.2
27
2
28 As the Court grants Musk’s motion as to the May 17 tweet, it will not address arguments
pertaining to that tweet for the remainder of the order.
9
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 10 of 38
C. Material Falsity
1
Musk asserts that the May 13 tweet was neither false nor material. JMOL at 20.
2
Plaintiffs disagree. See JMOL Opp’n at 12–14. As Musk’s arguments primarily go to his
3
view of the evidence, rather than its sufficiency, the Court denies his motion as to material
4
falsity.
5
1. Falsity
6
Musk argues that the May 13 tweet was not false because it only stated that the deal
7
was temporarily on hold pending details, and the evidence at trial showed that the merger
8
was held up while Musk sought spam information. JMOL at 20. Not so. There was
9
substantial evidence of falsity.
10
In his tweet, Musk stated, “Twitter deal temporarily on hold pending details
11
supporting calculation that spam/fake accounts do indeed represent less than 5 percent of
12
United States District Court
users.” Trial Ex. 97. Despite his public statement, Musk did not instruct his team or
13
lawyers to stop work on the deal. See Trial Tr. at 772:25–773:6 (Musk). Indeed, one of
14
his bankers at Morgan Stanley testified that the tweet surprised her and that Musk never
15
put the deal on hold or directed his team to do so. Trial Tr. at 1602:23–1603:14 (Claassen)
Northern District of California
16
(“We were never directed to put the deal on hold.”). Even a member of the Twitter team
17
testified that Musk had never notified them about a hold and that the deal was not, in fact,
18
on hold—temporarily or otherwise. See id. at 1103:7–17 (Segal) (“Q. Was the Twitter
19
deal on hold? A. No.”); see also id. at 395:12–16 (Gadde) (“Q. Did [Musk] tell Twitter to
20
stop working on the deal? A. Not to my recollection.”). That is sufficient evidence for the
21
jury to find that the tweet was literally not true.
22
Additionally, Plaintiffs assert that the tweet represented to the public that Musk had
23
the right to put the deal on hold, despite Musk waiving due diligence in a seller-friendly
24
agreement. JMOL Opp’n at 13. Musk responds that his statement would be “an opinion
25
not actionable under the securities laws even if incorrect.” JMOL at 21. But opinions can
26
still be actionable if they affirm that the speaker holds the stated belief, they embed
27
statements of fact, and a reasonable investor could understand them to convey facts about
28
10
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 11 of 38
1 the speaker’s basis for the view. Golub v. Gigamon Inc., 994 F.3d 1102, 1106 (9th Cir.
2 2021) (citing Omnicare, Inc. v. Laborers Dist. Council Const. Indus. Pension Fund, 575
3 U.S. 175 (2015)); see also Jury Instructions (dkt. 527) at 20 (instruction on actionable
4 opinions). Here, a jury could rely on evidence that Musk publicly claimed a right that he
5 did not enforce in private to find that his professed opinion was false.
6 2. Materiality
7 Musk also contends that the falsity of the tweet was not material. JMOL at 21. For
8 a misrepresentation to be material, “there must be a substantial likelihood that the
9 disclosure of the omitted fact would have been viewed by the reasonable investor as
10 having significantly altered the ‘total mix’ of information made available.” Basic Inc. v.
11 Levinson, 485 U.S. 224, 232 (1988) (citation modified). There was substantial evidence
12 that the May 13 tweet was material.
United States District Court
13 First off, Musk—a Twitter investor seeking to purchase all outstanding common
14 stock at the time—testified that his tweets about the deal were “material information”
15 about the deal’s status that “the public should know.” Trial Tr. at 667:12–668:5 (Musk).
Northern District of California
16 Brian Belgrave, one of the Plaintiffs, testified that he believed the deal being on hold
17 “created the concern, created the uncertainty that created the fear in the market.” Id. at
18 305:13–22 (Belgrave). The tweet was also quickly followed by a significant reduction in
19 Twitter’s share price. See In re Alphabet, Inc. Sec. Litig., 1 F.4th 687, 703 (9th Cir. 2021)
20 (immediate market reaction supported materiality). And Dr. Tabak testified that this
21 “basically instantaneous” reaction was statistically significant. Trial Tr. at 1825:23–
22 1826:20 (Tabak). That is enough evidence to support the jury’s finding of materiality.
23 D. Scienter
24 Musk contends that there was no evidence of scienter because he acted in good
25 faith, believing his statements were true when he made them. JMOL at 22. He relies
26 primarily on his own testimony to support his position. See id. at 22–23. But “it is rare
27 that perpetrators of a fraud would confess outright.” In re Peoplesoft, Inc., No. C 99-
28 00472 WHA, 2000 WL 1737936, at *3 (N.D. Cal. May 25, 2000). That is why “proof of
11
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 12 of 38
1 scienter in fraud cases is often a matter of inference from circumstantial evidence.”
2 DSAM Glob. Value Fund v. Altris Software, Inc., 288 F.3d 385, 390 (9th Cir. 2002)
3 (internal citation omitted). Courts may consider evidence of “motive” or “personal
4 financial gain” for inferences of scienter. Tellabs, Inc. v. Makor Issues & Rts., Ltd., 551
5 U.S. 308, 325 (2007). The Court concludes that there was substantial evidence to support
6 a finding that Musk knowingly made “an untrue statement with the knowledge that the
7 statement was false or with deliberate reckless disregard for whether the statement was
8 true.” See Jury Instructions at 21.
9 For starters, there was sufficient evidence that Musk had a motive to back out of, or
10 at least renegotiate, the deal. Musk had planned to—and eventually did—use stock from
11 his other company, Tesla, Inc., to buy Twitter. See Trial Tr. at 555:14–556:25 (Birchall).
12 Tesla had “a special place in [Musk’s] heart” but the Tesla share price dropped
United States District Court
13 significantly after he agreed to buy Twitter. See id. at 742:5–7; Trial Ex. 261 (showing
14 decline in Tesla share price). As Plaintiffs note, it could mean Musk may have had to sell
15 more Tesla stock to pay for the deal as it was then-structured. See JMOL Opp’n at 16.
Northern District of California
16 And there was evidence Musk told his biographer that the cost of the deal would require
17 him to “take on a lot of debt” and he believed it would “work out at a much lower
18 price . . . literally half or something.” Trial Tr. at 976:11–23 (Isaacson).
19 There was also evidence that Musk’s publicly-stated reason for his problem with the
20 deal—the number of bots on Twitter—was something he knew he was misrepresenting.
21 As he stated in his May 13 tweet (and others), Musk was publicly arguing that the number
22 of bots on Twitter was much higher than Twitter represented and that he would not close
23 until Twitter gave him answers regarding its methodology. He testified that his impression
24 of the high number of bots came from his user experience on the platform. Trial Tr. at
25 690:24–691:6 (Musk). In fact, he conceded that even before he signed the merger
26 agreement, he thought there were more bots on Twitter than it represented. See id. at
27 693:6–18. Musk testified that, before he signed the deal, he believed spam was “running
28 rampant” and thought he could fix it. Id. at 694:4–17.
12
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 13 of 38
1 But at the same time, Musk said he knew that Twitter estimated its bot percentage
2 by looking at the number of people or accounts that logged into Twitter on a given day, or
3 the monetizable daily active users (“mDAU”). Id. at 711:14–712:8. Musk was aware that
4 mDAU included users who merely lurked on the app and did not leave a publicly
5 noticeable footprint, which he understood included “most of the people in mDAU.” Id. at
6 712:4–11. Nevertheless, Musk repeatedly insisted that there were more bots on Twitter
7 than it claimed merely because bot activity was actually visible, even though mDAU
8 mostly included users without online footprints. See id. at 789:22–25 (“Q. You’re using
9 the wrong denominator, though, aren’t you? A. It’s—I’m using the number that people
10 can experience. Like what is their experience in the platform? Does it seem like only 5
11 percent is bots and spam?”).
12 Accordingly, the evidence at trial supports a jury finding of scienter. A jury could
United States District Court
13 conclude that Musk had a motive to get out of the existing deal and used bots as a pretext
14 to do so. The evidence also showed that Musk was aware of the impact his tweets could
15 have on the stock market. Trial Tr. at 664:13–21. And his conduct behind the scenes,
Northern District of California
16 such as not actually telling Twitter or his team that the deal was on hold, supports a strong
17 inference of knowledge—or at least deliberate recklessness—as to the falsity of his
18 statements.
19 E. Reliance
20 Musk asserts that Plaintiffs did not offer evidence that the fraud-on-the-market
21 presumption applied and that there was also no evidence of individual reliance. JMOL at
22 23–24. Plaintiffs argue that they did not need evidence of individual reliance because the
23 evidence properly established the fraud-on-the-market presumption. JMOL Opp’n at 18.
24 The Court agrees with Plaintiffs that the presumption applied and was not rebutted.
25 The fraud-on-the-market theory recognizes a “rebuttable presumption of reliance on
26 material misrepresentations aired to the general public.” Amgen Inc. v. Connecticut Ret.
27 Plans & Tr. Funds, 568 U.S. 455, 461 (2013). It “rests on the premise that certain well
28 developed markets are efficient processors of public information.” Id. Accordingly, the
13
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 14 of 38
1 market price will reflect “all publicly available information, and hence, any material
2 misrepresentations.” Basic, 485 U.S. at 246. “[A]nyone who purchases the stock at the
3 market price may be considered to have done so in reliance on the misrepresentation.”
4 Halliburton Co. v. Erica P. John Fund, Inc., 573 U.S. 258, 284 (2014). “To invoke the
5 Basic presumption, a plaintiff must prove: (1) that the alleged misrepresentation was
6 publicly known; (2) that it was material; (3) that the stock traded in an efficient market;
7 and (4) that the plaintiff traded the stock between the time the misrepresentation was made
8 and when the truth was revealed.” Goldman Sachs Grp., Inc. v. Arkansas Tchr. Ret. Sys.,
9 594 U.S. 113, 118 (2021). The Court determines that Plaintiffs availed themselves of the
10 fraud-on-the-market presumption because they provided sufficient evidence for each of the
11 Basic factors.
12 For the first factor, it is undisputed that Musk’s tweets were publicly known on
United States District Court
13 Twitter. And for the second factor, as previously discussed, Musk’s tweets were
14 materially false. For the third factor, Dr. Tabak testified that Twitter stock traded on an
15 efficient market. See Trial Tr. at 1809:24–1810:10 (Tabak).3 As to the fourth factor, the
Northern District of California
16 Class Period is limited to class members who sold between Musk’s misrepresentation and
17 when the truth—that he was going to buy Twitter—was revealed. See JMOL Opp’n at 19.
18 Musk asserts that he rebutted the presumption by providing evidence that the
19 market already knew the truth and so his misrepresentations could not have affected the
20 market price. JMOL at 24. A defendant can rebut the presumption with a “showing that
21 severs the link between the alleged misrepresentation and either the price received (or
22 paid) by the plaintiff.” Basic, 485 U.S. at 248. Musk points out that the market knew he
23 had “waived diligence well before May 13,” so his tweet could not have affected Twitter’s
24
3
25 Musk argues that the Plaintiffs’ theory of loss causation did not show an efficient market because
there was no immediate market response to his May 13 tweet. JMOL at 23. “But an immediate
26 response is not required for loss causation.” Miller v. Thane Int’l, Inc., 615 F.3d 1095, 1103 (9th
Cir. 2010) (emphasis in original). Regardless, Musk misconstrues the evidence. Dr. Tabak did
27 testify that there was an immediate decline in the Twitter stock price after the May 13 tweet. Trial
Tr. at 1892:25–1893:2. Musk’s argument against the inclusion of May 16 in Dr. Tabak’s event
28 study confuses the timing of the reaction with its duration. See JMOL at 23 (inclusion of May 16
stock drop “is incompatible with an efficient market”).
14
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 15 of 38
1 stock price. JMOL at 24. But Musk did not carry his burden here “by a preponderance of
2 the evidence.” See Goldman, 594 U.S. at 126. As Plaintiffs note, Musk himself presented
3 evidence to contest what waiving due diligence meant and, therefore, what the market may
4 have known. See, e.g., Trial Tr. at 732:7–9 (Musk) (“Q. You knowingly, voluntarily,
5 strategically waived due diligence in order to make this offer more attractive to Twitter,
6 right? A. On the assumption that the public filings were accurate.”); id. at 1752:5–13
7 (Armstrong) (“Q. And due diligence can be used to gain a better understanding of the
8 disclosures and the financial statements in a company’s annual report, right? A. No, I
9 disagree with that comment.”).
10 Musk’s second rebuttal argument also fails. He contends that the market learned
11 the merger was not on hold at the same time as the May 13 tweet. JMOL at 24. On this
12 point, Musk cites Dr. Tabak’s testimony that people were “immediately saying things and
United States District Court
13 saying that Twitter has not put it on hold, so it would be folded in with May 13th and
14 16th.” See id. (citing Trial Tr. at 1909:10–15 (Tabak)). But Musk is reaching beyond
15 what Dr. Tabak actually said. Dr. Tabak’s testimony was limited to Twitter’s position—
Northern District of California
16 that Twitter did not agree and the market was aware of that. Id. at 1909:16–19 (“Q. Okay.
17 So your testimony is that people knew as of May 13 that Twitter had not agreed to put the
18 deal on hold? A. Yes. They said they hadn’t seen anything from Twitter.”). Knowledge
19 that Twitter did not agree to a hold does not equate to knowledge that the deal was not, in
20 fact, on hold. It says nothing of Musk’s position, which was the opposite of Twitter’s as
21 represented in his tweet.
22 Accordingly, the Court concludes that there was insufficient evidence for Musk to
23 have rebutted the presumption.
24 F. Damages
25 Musk argues that Plaintiffs failed to prove legally sufficient damages because their
26 expert, Dr. Tabak, did not perform the proper analysis for the measure of damages and
27 failed to disaggregate economic loss by market or industry factors. JMOL at 10–14. The
28 Court disagrees.
15
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 16 of 38
1. Measure of Damages
1
Musk contends that Dr. Tabak’s “out-of-pocket” damages model was flawed.
2
JMOL at 11. Out-of-pocket loss is “the difference between the value of what the plaintiff
3
gave up and the value of what the plaintiff received.” Ambassador Hotel Co. v. Wei-
4
Chuan Inv., 189 F.3d 1017, 1030 (9th Cir. 1999). The Court’s jury instructions describe
5
damages as “the difference between the price at which a stock sold and the price at which
6
the stock would have sold had the statement or fraudulent acts reflected the true state of
7
affairs.” Jury Instructions at 27.
8
At the start of the Class Period, the true state of affairs was that there was an
9
ongoing deal to buy Twitter. See Trial Tr. at 608:1–24 (Birchall) (deal work was pushing
10
ahead as usual). Consequently, on May 13, Musk had three options. He could either say
11
nothing, comment on the true state of affairs that the deal was proceeding as usual, or
12
United States District Court
make a misrepresentation. As the jury concluded, Musk made an affirmative
13
misrepresentation.
14
According to Musk, Dr. Tabak only opined on the difference between the actual
15
stock price and the “but-for price” had Musk said nothing, instead of what the price would
Northern District of California
16
have been had his statements reflected the truth. JMOL at 11. Essentially, Musk contends
17
that saying nothing is different from affirmatively speaking the truth. Id. at 12. But this
18
mischaracterizes Dr. Tabak’s testimony. Dr. Tabak repeatedly testified that, for his
19
analysis, saying nothing was “just taking away the false stuff.” Trial Tr. at 1901:3–4
20
(Tabak). He clarified that “Saying nothing is essentially not presenting that
21
misrepresentation.” Id. at 1901:11–14 (emphasis added).
22
None of Dr. Tabak’s testimony is incompatible with the standard in the jury
23
instructions. Plaintiffs pursued the theory that Musk misrepresented the literal state of the
24
Twitter deal.4 See FAC (dkt. 31). For example, Musk’s May 13 tweet represented that the
25
26 4
At the motion hearing, Musk argued that the true state of affairs was the existence of a dispute
27 between him and Twitter or that he was getting cold feet for the deal. See June 18, 2026 Hr’g Tr.
(dkt. 587) at 31:11–25. Perhaps. But there are many possibilities for what the true state of affairs
28 was at the time—and Plaintiffs’ theory is one of them. Ultimately, it is a factual question for the
jury, and the Court concludes there was substantial evidence to support Plaintiffs’ position.
16
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 17 of 38
1 Twitter deal was “temporarily on hold.” Trial Ex. 97. Evidence at trial showed that the
2 deal was never on hold. See, e.g., Trial Tr. 1603:12–14 (Claassen) (“Q. Isn’t it a fact that
3 he didn’t put the deal on hold? A. Correct. We were never directed to put the deal on
4 hold, to my knowledge.”). In other words, Musk’s statement was the opposite of the true
5 state of affairs. Had he said nothing, thereby omitting the false representation, it would
6 reflect the true state of affairs just as if he had merely said that the deal was proceeding as
7 usual—the then-existing status quo. Accordingly, Dr. Tabak’s damages model could still
8 measure the stock price as if Musk’s statement reflected the true state of affairs or, in other
9 words, “the difference between the price [Plaintiffs received] for a security and the actual
10 value of that security at the time.” Cf. Chassin Holdings Corp. v. Formula VC Ltd., No.
11 15-CV-02294-EMC, 2017 WL 66873, at *13 (N.D. Cal. Jan. 6, 2017) (emphasis added)
12 (discussing out-of-pocket damages in the context of a buyer case).
United States District Court
13 2. Disaggregation
14 Musk also argues that the jury did not have “sufficient evidence to disaggregate and
15 isolate economic loss” caused by his misrepresentations “from price drops caused by
Northern District of California
16 general market or industry factors during the Class Period.” JMOL at 12. He contends
17 that Dr. Tabak failed to disaggregate market and industry effects and included impacts on
18 share price unrelated to fraud in the damages calculations. Id. at 13. Plaintiffs respond
19 that Dr. Tabak’s damage calculation methodology—using event studies and a constant
20 percentage measure of deflation—is entirely appropriate and adequately disaggregates
21 damages. JMOL Opp’n (dkt. 573) at 6. The Court agrees with Plaintiffs.
22 It is well-settled that “a damage calculation requires courts to adjust for any share
23 price gains that are unrelated to the fraud.” Luna v. Marvell Tech. Grp., Ltd., No. C 15-
24 05447 WHA, 2017 WL 4865559, at *3 (N.D. Cal. Oct. 27, 2017). There are many
25 methods to disaggregate unrelated factors. As Plaintiffs point out, Dr. Tabak relied on two
26 of them: constant percentage and event studies. See JMOL Opp’n at 6–7.
27 The constant percentage method determines the amount of artificial inflation or
28 deflation on each date of a Class Period. See Homyk v. ChemoCentryx, Inc., No. 21-CV-
17
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 18 of 38
1 03343-JST, 2025 WL 1547625, at *16 (N.D. Cal. May 30, 2025). The method uses a
2 “ribbon,” or the daily level of artificial inflation or deflation caused by fraud, based on the
3 “constant percentage” a share price was calculated to have been inflated or deflated by
4 throughout the period. See id. The method can reliably disaggregate out-of-pocket
5 damages from market and industry factors. Id. (“[C]ourts regularly find that the constant
6 percentage [] methodology . . . is reliable and appropriate.”); see also Police Ret. Sys. of
7 St. Louis v. Granite Constr. Inc., No. C 19-04744 WHA, 2021 WL 229310, at *7 (N.D.
8 Cal. Jan. 21, 2021) (measure of inflation per share over the class period could be
9 accomplished via “constant percentage inflation”). And Dr. Tabak utilized this method to
10 calculate a constant percentage of deflation over the Class Period, with reductions in the
11 percentage to exclude certain non-fraud factors that significantly depressed the share price.
12 See Trial Ex. 342; see also Trial Tr. at 1849:17–1853:17 (Tabak) (explaining deflation
United States District Court
13 over the Class Period).
14 While Musk still criticizes constant percentage methodology, he concedes that it “is
15 appropriate . . . when based on an event study that removes non-fraud-related factors.”
Northern District of California
16 JMOL Reply (dkt. 578) at 8. Dr. Tabak did conduct an event study. See Trial Tr. at
17 1843:12–16. But because only two events were excluded, Musk argues that Dr. Tabak’s
18 analysis insufficiently removes “unrelated market movements.” JMOL Reply at 8. That is
19 not correct.
20 An event study considers “whether and to what extent any non-fraud related
21 information (i.e. ‘confounding information’) contributed to the observed price movement”
22 during a class period. In re FibroGen Sec. Litig., No. 21-CV-02623-EMC, 2024 WL
23 1064665, at *16 (N.D. Cal. Mar. 11, 2024). It is “widely accepted for calculating damages
24 of a class of stockholders.” SEB Inv. Mgmt. AB v. Symantec Corp., 335 F.R.D. 276, 288
25 (N.D. Cal. 2020). Event studies can be performed using a regression model, which is a
26 statistical technique for measuring the ability of one or more variables to explain another
27 variable of interest. See id. The model can be used to observe the typical relationship
28 between the stock price and market factors. Id. Dr. Tabak conducted a regression analysis
18
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 19 of 38
1 for his event study to determine whether there were any statistically significant price
2 decreases. Trial Tr. at 1843:7–1850:22. He only found two non-fraud related events that
3 were statistically significant: Musk’s termination letter on July 8, 2022 and Twitter
4 whistleblower allegations on August 23, 2022. Id. at 1845:2–20. And he excluded them
5 both from the damages calculation. Id. at 1850:4–16. Musk offers no evidence to suggest
6 that any other events ought to have also been excluded.5
7 Accordingly, the Court concludes that there was substantial evidence from which
8 the jury could disaggregate market and industry factors from Plaintiffs’ damages.
9 G. New Trial
10 Despite prevailing on two claims, Musk asks the Court for a new trial because he
11 asserts that (1) the verdict was contrary to the weight of the evidence, (2) there was juror
12 bias, (3) there was taint from Plaintiffs’ unsuccessful theory of scheme liability, and (4)
United States District Court
13 there was cumulative error from trial.6 JMOL at 25–30. Plaintiffs disagree. JMOL Opp’n
14 at 21–30. The Court takes each argument in turn.
15 1. Contrary to the Weight of the Evidence
Northern District of California
16 Musk reincorporates his arguments from his motion for a judgment as a matter of
17 law to argue that, even if substantial evidence supports the verdict, the verdict is contrary
18 to the clear weight of the evidence. JMOL at 25. The Court disagrees. As discussed
19 regarding Musk’s motion for a judgment as a matter of law, the Court does not find that
20 the verdict is contrary to the clear weight of the evidence—far from it. To be sure, there is
21 “substantial evidence that goes both ways” on some of the issues, but that is not enough to
22 warrant a new trial. See Union Oil Co. of California v. Terrible Herbst, Inc., 331 F.3d 735,
23 743 (9th Cir. 2003) (reversing grant of new trial where the trial court “believed that the
24 jury had failed to appreciate certain facts”). It is not the Court’s “place to substitute [its]
25
5
26 Musk did not call a rebuttal expert—or any expert for that matter. Dr. Tabak’s testimony was the
only expert evidence before the jury on damages.
6
27 Musk also asks that the Court amend the judgment to include only judgment for the class
representatives, as he argues that the verdict did not reflect fraud-on-the-market reliance for class-
28 wide liability. JMOL at 30. As Musk advances the same argument in his motion to decertify the
class, the Court addresses this point in its discussion of that motion.
19
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 20 of 38
1 evaluations for those of the jurors.” Id. at 742–43.
2 2. Juror Bias
3 Musk argues that the entire verdict is a product of bias and must be thrown out.
4 JMOL at 25. He highlights three points that he believes demonstrate bias: (1) a prejudiced
5 jury pool; (2) Plaintiffs’ counsel’s conduct; and (3) the verdict form. Id. The Court takes
6 these one by one.
7 a. Jury Pool
8 For Musk’s first argument, he reaches back to before the jury was even selected.
9 He asserts that the jury pool exhibited significant hostility to him—so much that he could
10 not get a fair trial. See JMOL at 26. The Court disagrees.
11 Musk is correct that that much of the pool did not like him. It is to be expected that
12 Musk is polarizing; he is a public figure and businessman who has recently been involved
United States District Court
13 in American politics at the highest levels. As the Court noted during jury selection,
14 “people feel strongly about Mr. Musk one way or the other.” Trial Tr. at 21:19–22:2 (Jury
15 Voir Dire). That is not to say that Musk was universally hated. Some prospective jurors
Northern District of California
16 liked him. See, e.g., id. at 100:15–16 (prospective juror believing Musk is “a good person
17 and he put a lot of effort to support humanity”). And some were neutral. See, eg., id. at
18 53:25–54:4 (prospective juror with “[n]o views” on Musk). Nevertheless, the Court agrees
19 that there were more negative views than positive ones.
20 But that does not mean that the eventual jury will be similarly biased in its makeup.
21 Courts have significant experience in finding ways to filter out partial jurors. For example,
22 the Court has “broad discretion” to question potential jurors to detect bias. See Mu’Min v.
23 Virginia, 500 U.S. 415, 423–24 (1991). And it exercised that discretion during jury
24 selection to seek out potential bias and evaluate whether potential jurors could be
25 rehabilitated. See, e.g., Trial Tr. at 47:8–48:3 (questioning juror about potential bias); see
26 also United States v. Martinez-Salazar, 146 F.3d 653, 656 (9th Cir. 1998) (courts have
27 leeway to keep jurors “who initially admit[] bias as long as [they] ultimately assert[] an
28 ability to be fair and impartial”), rev’d on other grounds, 528 U.S. 304 (2000).
20
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 21 of 38
1 There is another reason Musk’s objection falls flat: he had the ability to strike
2 jurors. “Challenges for cause are the means by which partial or biased jurors should be
3 eliminated.” United States v. Gonzalez, 214 F.3d 1109, 1111 (9th Cir. 2000). Musk’s
4 counsel had the opportunity to strike potential jurors for cause, even with the Court
5 denying a few challenges. See Trial Tr. at 134:1–147:14; cf. Carpenter v. Chappelle, No.
6 C 98-2444 MMC, 2012 WL 2343615, at *11 (N.D. Cal. June 20, 2012) (“A trial court’s
7 failure to strike even a biased juror for cause will not constitute a Sixth Amendment
8 violation so long as the jury that hears the case is impartial.”). And the Court granted
9 Musk—and Plaintiffs—six peremptory challenges to further eliminate jurors he felt were
10 prejudiced against him. Trial Tr. at 148:12–19. Notably, Musk does not point to any
11 individual juror in his eventual jury that he believes should have been removed. He merely
12 criticizes the pool and attributes bias to the jurors that made the final cut.
United States District Court
13 Consequently, the Court determines that issues with the jury pool did not deprive
14 Musk of a fair trial.
15 b. Counsel Conduct
Northern District of California
16 Musk argues that Plaintiffs’ counsel acted improperly throughout the trial to bias
17 the jury. JMOL at 26. The Court disagrees that counsel’s conduct warrants a new trial.
18 Musk asserts that Plaintiffs’ counsel cultivated an “us-versus-them” theme by
19 asking witnesses about their local ties and emphasizing their own, while at the same time
20 asking Musk about his decision to move himself and his business out of California. JMOL
21 at 26. The Court did not observe such questions rising to the level of attorney misconduct
22 or prejudice. Lawyers regularly attempt to establish camaraderie with witnesses through a
23 shared background. And as Plaintiffs note, the questions did not draw objections from
24 Musk. See JMOL Opp’n at 23. Moreover, while the Court understands Musk’s concern
25 with questions about his leaving the state, Plaintiffs’ counsel is entitled to establish basic
26 background about witnesses and did not go so far that it became troubling. See Apple, Inc.
27 v. Samsung Elecs. Co., No. 11-CV-01846-LHK, 2014 WL 549324, at *13 (N.D. Cal. Feb.
28 7, 2014) (new trials are only appropriate where the misconduct permeates the entire
21
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 22 of 38
1 proceeding).
2 Next, Musk contends that Plaintiffs’ counsel’s statement in closing argument about
3 sending a message with a verdict improperly biased the jury. JMOL at 27. Although the
4 Court agrees that it was an improper statement, it does not warrant a new trial. During
5 closing argument, Plaintiffs’ counsel stated:
And I’m going to tell you now that there are a lot of people in
6 this courtroom—and it’s all about a message, and the message
is: Are we going to allow people like Ms. Price, Mr. Belgrave,
7 and the millions of people in between to not be protected? And
that seal protects them, and that’s why we came in this
8 courtroom.
9 Trial Tr. at 2291:6–12 (Plaintiffs’ Rebuttal). The Court previously instructed Plaintiffs’
10 counsel that he could say a finding of liability would send a message but not that the jury
11 should find a practice wrong because it would send a message. Trial Tr. 2136:11–15
12 (Charging Conference).
United States District Court
13 “[W] here punitive damages are not at issue, urging the jury to ‘send a message’ by
14 its verdict is generally considered an improper appeal to the jurors’ passion and prejudice.”
15 Hammonds v. Yeager, No. EDCV 15-1036 SS, 2017 WL 10560471, at *1 (C.D. Cal. Aug.
Northern District of California
16 9, 2017). The Court concludes that counsel’s statement violated the nature of the Court’s
17 restriction on argument. Counsel essentially asked the jury to find for Plaintiffs because it
18 would send a message. But while counsel’s statement could potentially evoke prejudice, it
19 is not enough to warrant a new trial.
20 For reversal on grounds of attorney misconduct, the flavor of the misconduct “must
21 sufficiently permeate an entire proceeding to provide conviction that the jury was
22 influenced by passion and prejudice in reaching its verdict.” McKinley v. City of Eloy,
23 705 F.2d 1110, 1117 (9th Cir.1983). Counsel’s statement did not rise to this threshold. It
24 did not permeate the proceedings and was “confined to a few seconds of the closing
25 argument.”7 Apple, 2014 WL 549324, at *14. The attorney who made the statement had a
26
27 7
The Court notes that Musk did not object when Plaintiffs’ counsel made the statement. See Kehr
28 v. Smith Barney, Harris Upham & Co., 736 F.2d 1283, 1286 (9th Cir. 1984) (considering lack of
objection to closing argument in affirming denial of new trial motion).
22
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 23 of 38
1 more limited role at trial compared to other counsel for Plaintiffs. And, most importantly,
2 a new trial “is warranted only if counsel’s misconduct affected the verdict.” Mateyko v.
3 Felix, 924 F.2d 824, 828 (9th Cir. 1990). That is not the case here, where the jury found in
4 Musk’s favor for two out of four claims.8
5 c. Jury Verdict
6 Musk also challenges the completed verdict form as proof of prejudice. JMOL at
7 27. Musk points to the table for deflation per share over the Class Period and notes that all
8 the deflation values are in black ink except for one. Verdict (dkt. 538) at 6–10. The only
9 number in blue ink was $4.20. See id. at 8. Musk argues that the number 420 is often
10 negatively associated with him, and that it proves the jury used its verdict to send a
11 message. JMOL at 27–28. Musk’s argument fails for several reasons.
12 For starters, that the jury sided with him on two claims takes the wind out of
United States District Court
13 Musk’s sails. See Chalmers v. City of Los Angeles, 762 F.2d 753, 761 (9th Cir. 1985)
14 (“Without evidence of prejudice we will not grant a new trial.”). It defies common sense
15 that the jury could be so prejudiced against Musk yet absolve him of liability when they
Northern District of California
16 had the chance to send a message. The objective record of the jury’s deliberations further
17 undercuts Musk’s picture of an impassioned jury. The jury did not immediately turn in its
18 verdict after closing arguments. Instead, it deliberated for multiple days before making a
19 decision. And there is evidence that the jury’s deliberations were careful and thoughtful.
20 After all, the jurors submitted six notes about the evidence at trial and the law in their jury
21 instructions, even catching a typographical error in the instructions. See First Set of Jury
22 Notes (dkt. 532); Second Set of Jury Notes (dkt. 537). None of this hints at anti-Musk
23
24 8
The Court also included a jury instruction to remind the jury to base its verdict solely on the
25 evidence it had heard, and the law provided by the Court. Jury Instructions at 2. Musk takes issue
with the instruction, as he believes it prompted the jury about considering media coverage. JMOL
26 at 27. But it is a simple fact that the case would get publicity—there were many people and
reporters in the audience, particularly the days Musk testified. And the Court specifically
27 cautioned the jury they “should not consider [publicity] in determining what [their] verdict should
be.” Jury Instructions at 2. It is “presume[d] that jurors carefully follow the instructions given to
28 them.” Caudle v. Bristow Optical Co., 224 F.3d 1014, 1023 (9th Cir. 2000), as amended on denial
of reh’g (Nov. 2, 2000) (citation modified).
23
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 24 of 38
1 animus.
2 Next, the number 420 is not necessarily indicative of prejudice against Musk. He
3 asserts that 420 is negatively associated with him but does not offer any evidence to
4 support it. To the contrary, 420 is a reference to cannabis/marijuana. See, e.g., United
5 States v. Jobe, 933 F.3d 1074, 1076 (9th Cir. 2019) (noting that 420 refers to marijuana).
6 One need only walk around San Francisco on April 20 to observe how prevalent the
7 celebration can be. What’s more, a joke about 420 played a role in the trial. See Trial Tr.
8 1524:16–19 (Claassen) (“[W]e knew Mr. Musk had joked about 420 quite a bit, and so we
9 threw it out there as 54.20.”). Consequently, it is also possible that a jury in San Francisco
10 recognized and appreciated the joke—just as Musk has.
11 Lastly, the verdict form has many instances where different colors of ink were used.
12 For example, the jury only marked liability for the May 17 tweet in blue ink—the rest of
United States District Court
13 the claims are in black. Verdict at 2–5. The jurors also wrote “Reference Exhibit 342” in
14 blue above the damages section heading. Verdict at 6. The jury used different pens at
15 different times when filling out the verdict form, not only when writing the number 4.20.
Northern District of California
16 Accordingly, the Court rejects Musk’s challenge to the verdict form.
17 3. Scheme Liability
18 Musk asserts that the trial was unfair because Plaintiffs’ claim for scheme liability
19 distracted him from defending the misrepresentation-only case and permitted all types of
20 otherwise prejudicial evidence to be shown to the jury. JMOL at 28. His argument lacks
21 merit.
22 First and foremost, Musk cannot show prejudice from the scheme claim. That is
23 primarily because he prevailed on the claim at trial. Verdict at 5. Musk’s attempt to
24 reframe his purported prejudice as being forced to divide time and resources to defend
25 against two theories falls apart under inspection. See JMOL at 28. He argues that the
26 claim was invalid and that he attempted many times to exclude it from the case—citing
27 eleven of his motions. Id. at n.1. But Musk “cannot just reassert arguments that have
28 already been rejected in hope of a different result.” See Johnson v. U.S. Dep’t of the
24
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 25 of 38
1 Treasury, No. C 11-06684 WHA, 2012 WL 5269620, at *3 (N.D. Cal. Oct. 24, 2012).
2 Litigants “play with fire if they raise the same arguments over and over and fail to
3 acknowledge prior adverse rulings.” U.S. Commodity Futures Trading Comm’n v. Lake
4 Shore Asset Mgmt. Ltd., 540 F. Supp. 2d 994, 1015 (N.D. Ill. 2008). Musk has no excuse
5 to contest that he was on notice of the claim; he should have been prepared to put on a
6 defense at trial. As the Court noted on February 26, 2026 when rejecting Musk’s
7 challenge to the scheme claim, Musk “cannot now reasonably contend that he [was]
8 prejudiced in his ability to adequately prepare a defense against a theory that was part of
9 the complaint, brought up during discovery, permitted during summary judgment, and then
10 permitted again after a pretrial motion in limine.”9 See Order Re Ex Parte Motion (dkt.
11 454) at 6.
12 Second, Musk argues that the Court put no bounds on the scope of the scheme claim
United States District Court
13 and permitted otherwise inadmissible evidence to come in. JMOL at 28. But this is just
14 dissatisfaction with evidentiary rulings. District courts have “broad discretion” in
15 rendering evidentiary rulings. Harper v. City of Los Angeles, 533 F.3d 1010, 1030 (9th
Northern District of California
16 Cir. 2008). Erroneous evidentiary rulings only warrant new trials if the errors “ more
17 probably than not . . . tainted the verdict.” Id. As an initial matter, the Court’s rulings
18 could not have tainted the verdict if Musk was successful on both the scheme claim and
19 one misrepresentation claim. See Verdict. Moreover, there were bounds on the
20 admissibility of evidence at trial: the Federal Rules of Evidence. Musk makes clear that
21 the Court’s rulings were guided by the Rules because he insists that every time he tried “to
22 exclude highly prejudicial evidence wholly irrelevant to Plaintiffs’ narrow Rule 10b-5(b)
23 claims, [he] was met with the same response: It comes in for the scheme claim.” That is
24 entirely a description of how Rule 403, which permits relevant evidence if it is not
25 substantially unfairly prejudicial, works. See Fed. R. Evid. 403. Sometimes, evidence that
26
27 9
Musk’s representation that he could not defend against two theories at the same time is
28 undermined by the fact that the scheme theory is inclusive of the statements in the
misrepresentation theory—the scheme simply covers additional conduct.
25
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 26 of 38
1 may be irrelevant to one claim is relevant to another and is therefore admissible, even if
2 prejudicial.
3 Lastly, Musk contends that the inclusion of the scheme claim precluded him from
4 making certain arguments. JMOL at 29. The only example he provides is his apparent
5 inability to argue that his July 8 termination letter’s impact on Twitter’s stock price meant
6 that losses after that date could not be attributed to the fraud. Id. Musk states that
7 permitting the July 8 letter “as a basis for liability even though it is protected litigation
8 conduct” meant that he could not pursue this argument. Id. That is wrong on two fronts.
9 First, the Court simply found that the letter was not protected litigation conduct. See Trial
10 Tr. at 2077:4–2081:18 (Charging Conference) (finding that the July 8 letter was not
11 protected litigation conduct as a matter of law); see also Jury Instructions at 26
12 (instructions on protected litigation conduct covering the Delaware litigation itself). And
United States District Court
13 second, Musk does not provide a reason for why the scheme claim would bar him from
14 arguing, even in the alternative, that the letter was an intervening event for loss. To the
15 contrary, he did so at trial. See, e.g., Trial Tr. at 1917:8–12 (Tabak) (“Q. You would
Northern District of California
16 agree, though, regardless of whether it was part of a scheme or not part of a scheme, it
17 would be expected that the termination letter that was issued on July 8th would result in a
18 stock price decline, right? A. In this case, yes, I think that’s fair.”).
19 Consequently, the Court once again rejects Musk’s argument regarding scheme
20 liability.
21 4. Cumulative Prejudice
22 Musk repackages his other bases for a new trial into a catch-all argument that the
23 combination of errors requires a new trial. JMOL at 29. The Court rejects this argument,
24 too.
25 “[E]rrors in a civil trial must be considered ‘cumulatively’ and . . . the combined
26 effect of multiple errors ‘may suffice to warrant a new trial even if each error standing
27 alone may not be prejudicial.’” Sidibe v. Sutter Health, 103 F.4th 675, 688 (9th Cir. 2024)
28 (quoting Jerden v. Amstutz, 430 F.3d 1231, 1240–41 (9th Cir. 2005)). But this principle
26
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 27 of 38
1 only applies to cumulative errors by the trial court. See id. (analyzing cumulative error of
2 trial court rulings); Jerden, 430 F.3d at 1240–41 (same). As discussed above, the Court did
3 not find any errors on its part. The only instance of prejudice the Court found was
4 Plaintiffs’ counsel’s statement during closing argument, which the Court explained was
5 insufficient to trigger a new trial. Accordingly, the Court denies Musk’s motion.
6
7 II. MOTION TO DECERTIFY CLASS
8 Musk argues that there is no basis to conclude that Plaintiffs proved the fraud-on-
9 the-market presumption for class-wide reliance. Mot. to Decertify at 6. Because the jury
10 rendered a general verdict, Musk contends there is no way to know if the jury made its
11 decision based on individual reliance or the fraud-on-the-market presumption. Id. at 7. As
12 he believes individual issues abound, Musk asserts that Plaintiffs have failed to establish
United States District Court
13 the requirements of Rule 23 of the Federal Rules of Civil Procedure and so the class must
14 be decertified. Id. at 10–12. Plaintiffs respond that Musk’s motion is procedurally
15 improper and rests on the faulty assumption that the general verdict is ambiguous. Mot. to
Northern District of California
16 Decertify Opp’n (dkt. 574) at 2–9. The Court DENIES the motion.
17 A. Legal Standard
18 Rule 23(c)(1)(C) permits a court to “alter[] or amend[]” an order granting class
19 certification “before final judgment.” Fed. R. Civ. Pro. 23(c)(1)(C). The Ninth Circuit has
20 similarly stated that district courts may modify a class definition as a result of
21 developments during the course of litigation. See Armstrong v. Davis, 275 F.3d 849, 871
22 n.28 (9th Cir. 2001) (recognizing that Rule 23 “provides district courts with broad
23 discretion to determine whether a class should be certified, and to revisit that certification”
24 and that “the district court may redefine the class” (citing Penk v. Oregon State Bd. of
25 Higher Educ., 816 F.2d 458, 467 (9th Cir. 1987))). District courts have a responsibility to
26 continually review “the appropriateness of a certified class in light of developments
27 subsequent to class certification.” Schilling v. TransCor Am., LLC, 2012 WL 4859020, at
28 *1 (N.D. Cal. 2012). Ultimately, whether to decertify a class is within the discretion of the
27
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 28 of 38
1 Court. See Levya v. Medline Indus. Inc., 716 F.3d 510, 513 (9th Cir. 2013).
2 B. Discussion
3 Rule 23(c)(1)(C) only empowers the Court to decertify a class “before final
4 judgment.” Fed. R. Civ. Pro. 23(c)(1)(C). Musk’s motion comes after the Court entered a
5 final judgment on April 3, 2026. Dkt. 557. Moreover, as discussed regarding Musk’s
6 motion for a judgment as a matter of law, the Court concluded that Plaintiffs properly
7 invoked the fraud-on-the-market presumption and Musk failed to present evidence that
8 could have rebutted it. Nevertheless, the Court briefly discusses the merits in light of its
9 responsibility to ensure appropriate certification and Musk raising the issue prior to
10 judgment. See Response to Post-Verdict Procedures (dkt. 549). For the reasons below, the
11 Court concludes that Musk’s basis for the motion is fundamentally flawed.
12 At first blush, when only looking at the verdict form, Musk’s argument has some
United States District Court
13 appeal. The form does not include a special interrogatory where a juror could mark that
14 Musk had rebutted the fraud-on-the-market presumption.10 See Verdict. But Musk’s
15 assertion of ambiguity for class-wide reliance quickly falls apart upon review of the trial
Northern District of California
16 record.
17 That is because there was no evidence of individual reliance during trial. Plaintiffs
18 did not even attempt to introduce such evidence. See Mot. to Decertify Opp’n at 5–6. One
19 of their three class representatives, John Garrett, did not testify, so the jury could not have
20 premised liability on evidence of his individual reliance. And his wife Nancy Price, who
21 was also a class representative, did not testify that Garrett saw Musk’s May 13 tweet,
22 either.11 Further, Plaintiffs’ closing argument made clear that they were not advancing any
23 argument on individual reliance, just fraud-on-the-market for class-wide reliance. Trial Tr.
24 at 2214:7–23 (Plaintiffs’ Closing) (only discussing fraud-on-the-market for reliance); id. at
25
10
26 Cases with successful Rule 10b-5 claims regularly exclude special interrogatories on reliance in
their verdict forms. See, e.g., Verdict Form (dkt. 998) in In re Vivendi Universal, S.A. Securities
27 Litigation, 02 Civ. 5571 (RJH) (HBP); Jury Verdict (dkt. 1611) in Lawrence E. Jaffe Pension Plan
v. Household International, Inc., 1:02-cv-05893.
11
28 Price also admitted on cross-examination that she was not aware of the May 13 tweet “back in
2022.” Trial Tr. at 1365:22–1366:2 (Price).
28
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 29 of 38
1 2177:23–2178:2 (asking the jury “to compensate investors like Mr. Belgrave, Nancy Price,
2 John Garrett, and thousands of other investors”). And Musk’s closing argument only
3 discussed fraud-on-the-market, too. Trial Tr. at 2271:24–2272:13 (Musk Closing)
4 (arguing that Plaintiffs “can’t show reliance on the market”).
5 Faced with the actual trial record, Musk unsurprisingly urges the Court to restrict its
6 focus to the verdict form itself. Mot. to Decertify at 7 (arguing that “the general verdict
7 gives no indication which of these alternative theories of reliance the jury adopted”).
8 Citing Niles v. United States, Musk contends that the Court must turn a blind eye to the
9 record and cannot speculate on which reliance theory was the subject of the verdict. Id. at
10 8. In that case, the Niles court noted that, “where a matter is tried on alternate theories of
11 recovery and a general verdict rendered, appellate courts will not speculate on what
12 particular ground the jury may have found against (the) plaintiff.” 520 F. Supp. 808, 812
United States District Court
13 (N.D. Cal. 1981) (citation modified), aff’d sub nom. Niles By & Through Niles v. United
14 States, 710 F.2d 1391 (9th Cir. 1983).
15 Niles, however, is inapplicable because Musk’s argument fails twice over. As
Northern District of California
16 discussed, the case was not tried on alternative theories; Plaintiffs pursued fraud-on-the-
17 market only from the start. And more importantly, the Court does not have to speculate to
18 understand that the jury relied on the fraud-on-the-market presumption. Indeed, liability
19 could have only been premised on class-wide reliance. As explained, there was no
20 evidence of individual reliance in the record. How else could the jury find reliance where
21 one of the class representatives did not even testify if not for the fraud-on-the-market
22 presumption? Cf. UniRAM Tech., Inc. v. Taiwan Semiconductor Mfg. Co., No. C 04-
23 1268 VRW, 2008 WL 11515597, at *5 (N.D. Cal. Apr. 17, 2008) (“Any doubt in a general
24 verdict in a civil case is resolved in favor of the prevailing party: Absent [special]
25 interrogatories, the law presumes the existence of findings necessary to support the verdict
26 the jury reached.” (internal quotation omitted)).
27
28
29
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 30 of 38
1 Consequently, the Court denies Musk’s motion.12
2
3 III. MOTION FOR PREJUDGMENT INTEREST
4 Plaintiffs ask the Court to award prejudgment interest based on the rate set by 26
5 U.S.C. § 6621(a)(2), compounded daily from October 4, 2022 (the end of the Class Period)
6 to April 3, 2026 (the entry of final judgment). MPI at 10. Musk argues that no
7 prejudgment interest is required and, in the alternative, the Court award simple interest—or
8 at least interest compounded annually—based on the rate set in 28 U.S.C. § 1961. MPI
9 Opp’n (dkt. 576) at 13. The Court GRANTS the motion. The Court sets the rate at the
10 Treasury bill rate pursuant to Section 1961, compounded annually, and running from
11 October 4, 2022 through entry of final judgment.
12 A. Legal Standard
United States District Court
13 Prejudgment interest serves a compensatory function, designed to make the injured
14 party whole. Knapp v. Ernst & Whinney, 90 F.3d 1431, 1441 (9th Cir. 1996). The
15 decision whether to award prejudgment interest is left to the sound discretion of the trial
Northern District of California
16 court, guided by the factors from Osterneck v. Ernst & Whinney. Those factors include:
17 (1) whether prejudgment interest is necessary to compensate the plaintiff fully for his or
18 her injuries; (2) the degree of personal wrongdoing by the defendant; and (3) other
19 fundamental considerations of fairness. Osterneck v. Ernst & Whinney, 489 U.S. 169, 176
20 (1989). Prejudgment interest should not be “speculative” or provide a “windfall” to
21 plaintiffs. Knapp, 90 F.3d at 1442.
22 Courts have broad discretion to determine what rate of interest to apply and when
23 prejudgment interest begins. See Columbia Brick Works, Inc. v. Royal Ins. Co. of Am.,
24
12
25 Musk also argues that “[i]mposing class-wide liability despite the lack of any jury finding” for
the fraud-on-the-market presumption violates his procedural due process rights. Mot. to Decertify
26 at 12. As discussed, the fraud-on-the-market presumption was the only basis that could support
the jury’s verdict of liability, so there are no due process implications for class-wide reliance here.
27 Cf. Finjan, Inc. v. Sophos, Inc., No. 14-CV-01197-WHO, 2016 WL 2988834, at *19 (N.D. Cal.
May 24, 2016) (“[A] general jury verdict can give rise to collateral estoppel only if it is clear that
28 the jury necessarily decided a particular issue in the course of reaching its verdict.” (internal
quotation omitted)).
30
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 31 of 38
1 768 F.2d 1066, 1068 (9th Cir. 1985). In deciding if and how much prejudgment interest
2 should be granted, courts must examine matters encompassed within the merits of the
3 underlying action. Osterneck, 489 U.S. at 176.
4 B. Discussion
5 The Court determines that all three Osterneck factors are applicable and support an
6 award of prejudgment interest.
7 1. Necessity for Full Compensation
8 Contrary to Musk’s insistence, prejudgment interest is needed to fully compensate
9 the class members. As Plaintiffs note, due to Musk’s fraud, members of the class “were
10 deprived of the use and value of their money” for almost four years. See MPI at 3; see also
11 In re Vivendi Universal, S.A. Sec. Litig., 284 F.R.D. 144, 162 (S.D.N.Y. 2012) (“Plaintiffs
12 have been deprived of the use of their funds for nearly ten years, and prejudgment interest
United States District Court
13 is necessary to fully compensate them for their loss.”).
14 Musk argues that there is no basis to infer that the class members would have used
15 their proceeds and received a “predictable rate of return.” MPI Opp’n at 5. But similar
Northern District of California
16 arguments have been rejected by courts awarding prejudgment interest. See, e.g., Hsu v.
17 Puma Biotechnology, Inc., No. SACV 15-00865 AG (SHKX), 2019 WL 4295285, at *2
18 (C.D. Cal. Sept. 9, 2019) (rejecting argument that awarding interest assumes that “class
19 members would have taken $4.50 per share and made a conservative investment”). Musk
20 also contends that interest would serve as a windfall because “class members obtained cash
21 the moment they sold and were able to invest the proceeds as they saw fit.” MPI Opp’n at
22 5. But that is irrelevant because Plaintiffs are seeking interest on the amount of deflation
23 due to Musk’s fraud, not their actual proceeds when selling their stock. See MPI Reply
24 (dkt. 581) at 4.
25 2. Degree of Personal Wrongdoing
26 Musk argues that the verdict did not indicate that he willfully defrauded investors
27 and that he did not personally profit from Plaintiffs selling their shares. MPI Opp’n at 3–4.
28 Musk is correct that the jury could have found that he acted with deliberate recklessness
31
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 32 of 38
1 instead of with knowledge of falsity. See Jury Instructions at 21. But “although scienter is
2 not necessary to award interest, because the focus is on compensating plaintiffs,” Musk’s
3 potentially deliberate recklessness “does not weigh against an award of interest.” Vivendi,
4 284 F.R.D. at 162. Under both standards of intent available to the jury, Musk acted
5 intentionally. And “intentional violations of the federal securities laws support an award
6 of prejudgment interest.” Hsu, 2019 WL 4295285 at *2.
7 The Court does agree, however, that Musk’s lack of personal profit from Plaintiffs’
8 sales does cut against his level of wrongdoing. Nevertheless, on balance, Musk’s
9 fraudulent conduct favors an award. See id. (“[T]here is not a requirement that a defendant
10 be unjustly enriched in order to award prejudgment interest.”); see also Vivendi, 284
11 F.R.D. at 163 (“Vivendi’s argument that it never had use of plaintiffs’ money is irrelevant,
12 because the prejudgment interest inquiry focuses on compensating plaintiffs, not
United States District Court
13 disgorging inequitable gains from defendants.”).
14 3. Fairness
15 Musk also contends that it is unfair to reward investors—particularly those who
Northern District of California
16 “positioned themselves to profit from the alleged fraud”—with interest. MPI Opp’n at 6.
17 But “[s]hares are normally purchased with an eye toward a later sale.” Dura, 544 U.S. at
18 342. It comes as no surprise that investors seek to profit from their investment. In doing
19 so, they buy and sell stock “at the price set by the market” by relying on “the integrity of
20 that price.” Basic, 485 U.S. at 247. And an investor can still be harmed by deflation
21 attributable to fraud even if they also sought to profit from a lower price. Accordingly,
22 “the fair result is [still] to make the victims of [Musk’s] fraud whole.” Hsu, 2019 WL
23 4295285, at *2.
24 4. Appropriate Rate
25 The Court agrees with Musk that the Section 1961 rate is appropriate. The Section
26 6621 rate Plaintiffs propose is “generally applied by the Internal Revenue Service” to
27 “penalize individuals and corporations for the underpayment of taxes.” Hsu, 2019 WL
28 4295285, at *2. In contrast, the Section 1961 rate is in line with the “default rule in the
32
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 33 of 38
1 Ninth Circuit” for securities actions. See id. The interest is based on the “rate equal to the
2 weekly average 1-year constant maturity Treasury yield.” 28 U.S.C. § 1961(a). The Court
3 determines that the Treasury bill rate should be compounded annually pursuant to the
4 statute. See 28 U.S.C.A. § 1961(b).
5
6 IV. MOTION FOR APPROVAL OF CLASS NOTICE OF VERDICT AND
CLAIMS PROCEDURE
7
Plaintiffs also move for approval of their notice of verdict and claims administration
8
process. MFA. Musk opposes, primarily asserting that he has a right to conduct post-trial
9
discovery to test the individual reliance of class members, with additional jury trials as
10
necessary. MFA Opp’n (dkt. 577). The Court GRANTS the motion. The Court first
11
discusses Musk’s post-trial discovery request and then covers the other components of the
12
United States District Court
claims process.
13
A. Post-Trial Discovery
14
Musk asks the Court to appoint a special master to conduct a second phase of the
15
case that scrutinizes individual issues of reliance. MFA Opp’n at 14. He cites his right to
Northern District of California
16
rebut the fraud-on-the-market presumption and argues that it entitles him to post-trial
17
discovery for rebuttal. MFA Opp’n at 2–6. In Basic, the Supreme Court acknowledged
18
that defendants “may rebut proof of the elements giving rise to the presumption” or show
19
“that an individual plaintiff traded or would have traded despite his knowing the statement
20
was false.” 485 U.S. at 992. The Court certainly agrees that Musk had the right to rebut
21
the presumption in this case. As discussed, he tried and was unsuccessful at trial. But that
22
right does not extend to post-trial discovery.
23
Hsu v. Puma Biotechnology, Inc. is illustrative. After being held liable at trial, the
24
defendants made the same arguments Musk does here. They argued “that they have a right
25
to obtain reasonable discovery to challenge the reliance element of the class action suit,
26
that targeted discovery on individual reliance is necessary to prevent manifest injustice,
27
and that [their] discovery requests are reasonable and proportionate.” Hsu v. Puma
28
33
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 34 of 38
1 Biotechnology, Inc., No. SA CV 15-00865-DOC-SHK, 2021 WL 2644100, at *2 (C.D.
2 Cal. June 11, 2021). The Hsu court disagreed. It held that the defendants did not have “a
3 right to obtain reasonable post-trial discovery to challenge the reliance element of the class
4 action suit.” Id. The court concluded that Supreme Court precedent permitting rebuttal of
5 the fraud-on-the-market presumption on an individual basis did not extend that “right to
6 challenge reliance through reasonable discovery [] post-trial.” Id. The court rejected the
7 post-trial discovery request despite its final pretrial orders “permit[ting] an opportunity for
8 post-trial discovery” because that was “only an opportunity and not a right.” Id.
9 Hsu is on all fours with the instant case. Musk was well-aware of his ability to
10 rebut the presumption before trial—and he actually did so for one plaintiff. The Court
11 agreed that Musk rebutted the presumption when it deemed Steve Garrett “not an
12 appropriate class representative” because he did not rely on Musk’s fraud or the market
United States District Court
13 price when he sold his shares. Order Granting Class Certification (dkt. 106) at 9–10.
14 Musk’s post-trial insistence that he should now get additional discovery and trials runs
15 contrary to the purpose of the class action device. Cf. Smith v. Pac. Pers. Servs., Inc., No.
Northern District of California
16 17-CV-03594-SK, 2018 WL 11437641, at *4 (N.D. Cal. Oct. 11, 2018) (“[R]esolving
17 disputes in a single class action is far more efficient than litigating individual cases.”).
18 Further, Musk’s cited authority in support of his position is inapposite. See MFA
19 Opp’n at 3 (citing Lawrence E. Jaffe Pension Plan v. Household Int’l Inc., 2005 WL
20 3801463 (N.D. Ill. Apr. 18, 2005) and In re Vivendi Universal, S.A. Sec. Litig., 284
21 F.R.D. 144 (S.D.N.Y. 2012)). As the Hsu court noted, “the defendants in those cases
22 sought discovery on individual reliance issues during the pre-trial fact discovery period,
23 the discovery was expressly stayed until after trial, and the final pretrial orders specifically
24 provided for post-trial discovery.”13 2021 WL 2644100, at *2. Musk also cites
25
26 13
At the motion hearing, Musk urged the Court to consider a pretrial dispute where Plaintiffs
27 objected to discovery about unnamed plaintiffs and said it was a post-judgment claims process
issue. See June 18, 2026 Hr’g Tr. at 89:9–91:18 (discussing Joint Discovery Letter Brief (dkt.
28 175)). But as Musk conceded, that dispute “wasn’t about reliance.” Id. at 89:15. And Musk
never sought review after Magistrate Judge Ryu ruled on the issue. Id. at 94:7–10;
34
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 35 of 38
1 Jaroslawicz v. Engelhard Corp. MFA Opp’n at 1. Although that court ordered post-trial
2 rebuttal proceedings, it decided that issue before trial alongside a motion to dismiss.
3 Jaroslawicz v. Engelhard Corp., 724 F. Supp. 294, 303 (D.N.J. 1989).
4 Accordingly, the Court denies Musk’s request.14
5 B. Schedule
6 The Court adopts Plaintiffs’ proposed notice and claims schedule. See MFA at 1–2.
7 Musk only objects to the extent it does not permit him sufficient time (30 days) to evaluate
8 claims to test individual reliance or damages and asks for 120 days to review claims. MFA
9 Opp’n at 14. But as Plaintiffs note, Musk will actually have more than 30 days to review
10 claims. MFA Reply (dkt. 582) at 9. Starting 60 days after publication of the notice, the
11 Claims Administrator will start providing the parties with a summary update on the
12 received claims. See MFA at 1. The Administrator will then send updates to the
United States District Court
13 summaries every 30 days. Id. This builds in enough time for advanced preparation.
14 The approved schedule is reproduced below:
15
Northern District of California
16
17
18
19
20
21
22
23
24
25
26
27
14
28 To the extent Musk makes his post-trial discovery argument to challenge other parts of the
claims process, the Court rejects it.
35
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 36 of 38
1 C. Claims Administrator
2 As both parties agree, the Court appoints Epiq Class Action and Claims Solutions as
3 the Claims Administrator. See MFA at 2; MFA Opp’n at 10.
4 D. Notice of Verdict and Proof of Claim Form
5 The Court also adopts Plaintiffs’ proposed notice of verdict and proof of claim
6 form. The proposed notice concisely covers the necessary information regarding the case,
7 judgment, and class member rights. Musk’s request for additional information would only
8 create confusion and potential intimidation. Aside from his rejected right to post-trial
9 discovery, Musk asks the Court to include language that he will have access to the
10 submitted information and has the right to challenge submissions. MFA Opp’n at 13. The
11 Court agrees with Plaintiffs that the proposed claim form adequately explains that claims
12 and documentation will be reviewed without raising the specter that Musk—a significant
United States District Court
13 public figure—will be waiting in the wings for a potential challenge. See MFA Reply at 8.
14 Plaintiffs must, however, fix the mistaken inclusion of “Settlement” on page three of the
15 notice, as Musk identified. See MFA Opp’n at 10.
Northern District of California
16 In addition to his post-trial discovery argument, Musk makes a host of challenges to
17 the proof of claim and its underlying process.
18 First, he argues that the Court should make clear to Epiq that its “validation” of
19 claims is not adjudication. MFA Opp’n at 11. That is easy enough. Plaintiffs point out
20 that the “validation” Musk is referring to is simply verification of information for the
21 Court’s eventual approval. MFA Reply at 6.
22 Second, Musk asserts that the proposed process does not actually require claimants
23 to submit documentation in support of their claim. MFA Opp’n at 11. That is wrong.
24 Epiq requires documentation and flags ineligibility when it is missing. Azari Decl. (dkt.
25 567-3) ¶¶ 51–52. Even claim forms submitted electronically are flagged for deficiencies
26 and follow up. Id. ¶577.
27 Third, Musk asks the Court to require claimants to submit government-issued
28 identification to verify identity. MFA Opp’n at 12. That is unnecessary; the
36
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 37 of 38
1 documentation requested in the claim form is sufficient. Claimants must already submit
2 their names, contact information, addresses, Social Security Numbers or Taxpayer
3 Identification Numbers, and trading account numbers. See Claim Form (dkt. 567-5).
4 Musk provides no reason why additional verification would be required.
5 Fourth, Musk requests that Epiq provide the parties with “read-only” access to the
6 claims database so he can prepare in advance for any challenges. MFA Opp’n at 12. As
7 Plaintiffs note, the process already contemplates real-time access and monthly reports.
8 MFA Reply at 7; Azari Decl. ¶ 68 (party access to file transfer protocol site for document
9 storage).
10 And fifth, Musk objects that the Court should not permit late claims. MFA Opp’n
11 at 13. Plaintiffs propose that Epiq should be able to accept all late, but valid, claims that
12 are filed before it submits its final claims report with the Court. MFA at 6. Plaintiffs
United States District Court
13 assert that this process would sweep in otherwise valid claims from claimants who may be
14 abroad or difficult to reach. MFA Reply at 8. As there is still a hard deadline for such late
15 claims, the Court agrees with Plaintiffs that the process is appropriate. Cf. Lemus v. H &
Northern District of California
16 R Block Enters., LLC, No. C 09-03179 SI, 2013 WL 3831866, at *2 (N.D. Cal. July 23,
17 2013) (“A district court has discretion to allow late claims to a settlement fund.”).
18 E. Calculating Damages
19 The parties have two disputes as to damages calculation. They disagree on which
20 accounting method to use and whether class members who did not purchase stock during
21 the Class Period need documentation for when they bought it.
22 As to the accounting method, courts apply either the first-in-first-out method
23 (“FIFO”) or the last-in-first-out method (“LIFO”). Both methods aim to evaluate the order
24 in which securities sold during the Class Period will be matched against securities held or
25 bought in the period. Under FIFO, sales of stock during the Class Period are matched first
26 against holdings at the start of the Class Period and then purchases made during the Period
27 in chronological order. LIFO operates in reverse, matching sales against the most recent
28 shares purchased. Plaintiffs urge the Court to adopt FIFO. MFA at 9. Musk asks for
37
Case 3:22-cv-05937-CRB Document 590 Filed 07/06/26 Page 38 of 38
1 LIFO. MFA Opp’n at 8.
2 The Court adopts FIFO. It is an accepted method of calculating damages. See, e.g.,
3 Plumbers & Pipefitters Loc. 572 Pension Fund v. Cisco Sys., Inc., No. C 01-20418 JW,
4 2004 WL 5326262, at *3 (N.D. Cal. May 27, 2004) (FIFO is “a legitimate method for
5 computing losses or gains from stock purchases or sales.”); Gruber v. Gilbertson, 647 F.
6 Supp. 3d 100, 120 (S.D.N.Y. 2022) (approving FIFO). Musk rightly points out that many
7 courts prefer LIFO. See MFA Opp’n at 9. But the concerns some courts have about using
8 FIFO are not applicable here. Courts prefer LIFO because it accounts for “gains that might
9 have accrued to plaintiffs during the class period due to the inflation of the stock price.”
10 Cf. Gruber, 647 F. Supp. 3d at 119–20 (internal quotation omitted). That is not an issue in
11 a seller case with deflation. And Plaintiffs have also proposed adjustments to offset
12 windfalls. See MFA Reply at 5.
United States District Court
13 Musk also urges the Court to require class members who only purchased before the
14 Class Period to provide records of when they bought their shares as it would be relevant to
15 damages calculation. MFA Opp’n at 10. The Court disagrees. As Plaintiffs point out, for
Northern District of California
16 shares purchased before the Class Period, the date of purchase does not matter as there was
17 no fraud-related deflation at the time—a deflated price upon sale is the damage. See MFA
18 Reply at 6.
19 V. CONCLUSION
20 For the foregoing reasons, the Court GRANTS Musk’s motion for judgment as a
21 matter of law for Musk’s May 17 tweet and DENIES the remainder of the motion. The
22 Court also DENIES Musk’s motion to decertify the class. The Court GRANTS both
23 Plaintiffs’ motion for prejudgment interest and motion for approval of class notice of
24 verdict and claims administration procedure.
25 IT IS SO ORDERED.
26 Dated: July 6, 2026
CHARLES R. BREYER
27 United States District Judge
28
38