FOR PUBLICATION

UNITED STATES COURT OF APPEALS

FOR THE NINTH CIRCUIT

BLADEROOM GROUP LIMITED; No. 19-16583

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

EMERSON ELECTRIC CO.,

Defendant-Appellant,

and

FACEBOOK, INC.; EMERSON

NETWORK POWER SOLUTIONS, INC.;

LIEBERT CORPORATION,

Defendants. BLADEROOM GROUP LIMITED; No. 19-16584

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

EMERSON NETWORK POWER

SOLUTIONS, INC.,

Defendant-Appellant,

and

FACEBOOK, INC.; EMERSON

ELECTRIC CO.; LIEBERT

CORPORATION,

Defendants.

BLADEROOM GROUP LIMITED; No. 19-16585

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

LIEBERT CORPORATION,

Defendant-Appellant,

and

FACEBOOK, INC.; EMERSON

ELECTRIC CO.; EMERSON NETWORK

POWER SOLUTIONS, INC.,

Defendants.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 3

BLADEROOM GROUP LIMITED; No. 19-16730

BRIPCO (UK) LIMITED,

Plaintiffs-Appellants, D.C. No. 5:15-cv-01370- v. EJD

FACEBOOK, INC.,

Defendant,

and

EMERSON ELECTRIC CO.; EMERSON

NETWORK POWER SOLUTIONS, INC.;

LIEBERT CORPORATION,

Defendants-Appellees.

BLADEROOM GROUP LIMITED; No. 20-15758

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

EMERSON ELECTRIC CO.,

Defendant-Appellant,

and

EMERSON NETWORK POWER

SOLUTIONS, INC.; LIEBERT

CORPORATION,

Defendants. BLADEROOM GROUP LIMITED; No. 20-15759

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

EMERSON NETWORK POWER

SOLUTIONS, INC.,

Defendant-Appellant,

and

EMERSON ELECTRIC CO.; LIEBERT

CORPORATION,

Defendants.

BLADEROOM GROUP LIMITED; No. 20-15760

BRIPCO (UK) LIMITED,

Plaintiffs-Appellees, D.C. No. 5:15-cv-01370- v. EJD

LIEBERT CORPORATION,

Defendant-Appellant, ORDER AND

AMENDED

and OPINION

EMERSON ELECTRIC CO.; EMERSON

NETWORK POWER SOLUTIONS, INC.,

Defendants.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 5

Appeals from the United States District Court for the Northern District of California Edward J. Davila, District Judge, Presiding

Argued and Submitted February 4, 2021 San Francisco, California

Filed August 30, 2021 Amended December 21, 2021

Before: Johnnie B. Rawlinson and Patrick J. Bumatay, Circuit Judges, and Stephen J. Murphy, III, * District Judge.

Order; Opinion by Judge Murphy; Concurrence by Judge Rawlinson

SUMMARY **

Non-Disclosure Agreements / Damages and Interest

The panel amended the Opinion and Concurrence filed on August 30, 2021, denied a petition for panel rehearing, and denied on behalf of the court a petition for rehearing en banc. Holding that the district court erroneously interpreted a non-disclosure agreement (“NDA”), the panel (1) reversed the district court’s order granting a motion in limine as it

* The Honorable Stephen J. Murphy, III, United States District Judge for the Eastern District of Michigan, sitting by designation. ** This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. related to the twelfth paragraph of the NDA; (2) vacated the district court’s judgment and post-verdict orders; and (3) vacated the orders awarding attorneys’ and expert witness fees.

BladeRoom Group Limited and Emerson Electric Co. were competitors that began negotiating a sale of BladeRoom to Emerson, and they signed an NDA. The negotiations fell through. Facebook selected Emerson’s proposal for a data center, and BladeRoom sued Facebook and Emerson. Halfway through a jury trial, BladeRoom settled with Facebook, and the case continued against Emerson. Emerson proposed a jury instruction that would have excluded information disclosed or used after August 17, 2013, from its liability for breach of contract, which Emerson argued was the date of the contract’s expiration. The district court denied the instruction but allowed Emerson to make the legal argument to the jury. BladeRoom moved in limine to overturn that ruling. The district court granted the motion and agreed that the NDA’s confidentiality obligations did not expire under paragraph twelve of the NDA. The jury found that Emerson breached the NDA and willfully and maliciously misappropriated BladeRoom’s trade secrets. The jury found that BladeRoom sustained $10 million in lost profits and $20 million in unjust enrichment. The district court later awarded BladeRoom $30 million in punitive damages.

The panel held that the district court erred in interpreting the NDA. The panel applied English law, which interprets contracts to discern the contracting parties’ intent and balances textual and contextual analyses. The panel held that paragraph twelve’s natural meaning unambiguously terminated the NDA and its confidentiality obligations two years after it was signed. The district court therefore erred as

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 7

a matter of law when it granted BladeRoom’s motion in limine.

The panel treated the district court’s error as an error of jury instruction. First, BladeRoom’s motion in limine was a clear attempt to stymie the district court’s jury instruction ruling that allowed Emerson to argue any exclusion from liability it believed should apply.

Second, the district court’s order lacked basic findings necessary for the panel to construe it as a denial of a motion for judgment as a matter of law. Jury instruction error in a civil trial is reviewed for harmless error, and the panel presumed prejudice pursuant to Clem v. Lomeli, 566 F.3d 1177, 1182 (9th Cir. 2009). Thus, the burden shifted to BladeRoom to demonstrate that it was more probable than not that the jury would have reached the same verdict had it been properly instructed. The panel applied Clem’s burden shifting standard because Ninth Circuit precedent so instructed given the abnormal jury instruction error in these appeals. In contrast, for the ordinary civil appeal, the party seeking reversal bears the burden of persuasion, pursuant to Shinseki v. Sanders, 556 U.S. 396, 410 (2009). The panel held that Shinseki’s harmless error review did not apply to these appeals. The panel held further that several Ninth Circuit cases applied Shinseki’s harmless error review instead of Chen’s, but those cases differed greatly from the present appeals.

The panel held that the district court prejudiced Emerson when the jury made its breach of contract, misappropriation, and damage findings. First, the panel vacated the jury’s breach of contract findings. The jury answered only whether a breach of contract occurred, but not when the breach occurred; and, therefore, there was no way to determine whether the jury would have found that Emerson breached the NDA if the jury had known that Emerson’s confidentiality obligations ended after only two years. The holding in Clem required a finding of prejudice in the district court’s error here. Second, the panel vacated the jury’s misappropriation findings because there was no way to know whether the jury found that Emerson misappropriated trade secrets during or after the NDA’s initial two-year period. Without that finding, there was no way to know whether the jury would have found that Emerson acted willfully and maliciously. Finally, the same reasons supported vacating the jury’s damages findings. The jury awarded damages in two lump sums that did not separate damages for breach of contract from misappropriation, and there was no way to know how much damages the jury intended to allocate for each claim. Because it is not known whether the jury would have found Emerson liable under breach of contract or misappropriation, the panel could not tell whether the damages findings would be the same. The panel vacated these jury findings; and remanded for a new trial on those issues.

Given the panel’s order to vacate the judgment and remand for a new trial, the panel vacated the district court’s post-verdict orders on appeal. The panel also vacated the district court’s awards of attorneys’ and expert witness fees.

The panel addressed a number of issues for consideration on the awards of damages and prejudgment interest should they be determined after a new trial.

Under California law, a party cannot collect punitive damages for breach of contract awards. Because the jury awarded a lump sum for breach of contract and misappropriation damages, Emerson correctly argued that the district court could not assume that the whole award went to misappropriation. In awarding punitive damages, the

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 9

district court abused its discretion by changing its reasoning without explaining why, and the inconsistency visibly benefited BladeRoom. On remand, the district court must take several steps to allocate damages: the district court should consider adopting a more-detailed special verdict form; and if the jury cannot allocate damages, then the district court should take another step to explain why it allocated damages.

California Civil Code § 3288 allows the award of prejudgment interest for the breach of an obligation not arising from a contract, whereas California Civil Code § 3287(b) allows an award of prejudgment interest only for breach of contract claims. Awarding prejudgment interest under one section rather than the other can make a sizable difference. The panel held that the district court erred in awarding prejudgment interest on the full compensatory damages award under § 3288, not § 3287(b); and erred in finding that prejudgment interest should run from October 30, 2012. On remand, if the district court awards § 3288 prejudgment interest for BladeRoom’s lost profits, it would have to find when the lost profits began, and the award must run from when the lost profits began. Should the district court be called on to award § 3288 prejudgment interest for unjust enrichment after a new trial, the district court should find whether § 3288 allowed prejudgment interest for unjust enrichment damages; and, if so, the district court should apply the same standard that it applied for lost profits – it should find when the unjust enrichment began.

Judge Rawlinson concurred fully in the majority opinion, and wrote separately to bring two additional matters to the district court’s attention. First, in the event the jury on retrial imposes liability on Emerson for unjust enrichment, it would be BladeRoom’s burden to prove the amount of damages. BladeRoom will not be entitled to future profits absent definitive evidence of a future contract that was lost due to Emerson’s misappropriation of BladeRoom’s trade secret. Second, because BladeRoom conceded that Facebook and Emerson were joint tortfeasors and conspired to misappropriate BladeRoom’s trade secrets, California law required an offset. If the retrial results in the imposition of damages against Emerson, the district court should apply an offset for the amount of the settlement between BladeRoom and Facebook. Correspondingly, Emerson would be entitled to discovery of the settlement terms.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 11

COUNSEL

Carter G. Phillips (argued), Sidley Austin LLP, Washington, D.C.; Christopher M. Egleson, Sidley Austin LLP, Los Angeles, California; Constantine L. Trela Jr., Jillian S. Stonecipher, Taurean K. Brown, and Jason G. Marsico, Sidley Austin LLP, Chicago, Illinois; for Defendant- Appellant/Cross-Appellee Emerson Electric Co.

Rudolph A. Telscher Jr., Steven E. Holtshouser, Kara R. Fussner, and Michael C. Martinich-Sauter, Husch Blackwell LLP, St. Louis, Missouri, for Defendants-Appellants/Cross- Appellees Emerson Network Power Solutions Inc., and Liebert Corporation.

Stephanie P. Skaff (argued), Jeffrey M. Fisher, Carly O. Alameda, Erik C. Olson, Alex Reese, Nadia C. Arid, and Ashley Roybal-Reid, Farella Braun & Martel LLP, San Francisco, California, for Plaintiffs-Appellees/Cross- Appellants BladeRoom Group Limited and Bripco (UK) Limited. The Opinion and Concurrence filed on August 30, 2021 and reported at 11 F.4th 1010 is amended by the Opinion and Concurrence filed concurrently with this Order.

With these amendments, the full court has been advised of the petitions for rehearing en banc, and no judge of the court has requested a vote on the petitions. Fed. R. App. P. 35. The panel unanimously votes to deny the petitions for panel rehearing. Judges Rawlinson and Bumatay vote to deny the petitions for rehearing en banc, and Judge Murphy so recommends. The petitions for panel rehearing and rehearing en banc are DENIED. No further petitions for panel rehearing or rehearing en banc will be accepted.

OPINION

S. MURPHY, III, District Judge:

In these appeals, we consider how English law interprets a non-disclosure agreement (“NDA”). The district court’s interpretation misapplied English law because it conflicts with the NDA’s natural reading. We therefore vacate the district court’s judgment and remand for a new trial.

I.

BladeRoom and Emerson are competitors in modular data center design and building industry. In August 2011, the two began negotiating a sale of BladeRoom to Emerson. To start the process, BladeRoom drafted an NDA and both parties signed it. The parties agreed that English law governed the NDA.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 13

The NDA’s second paragraph included a dozen subparagraphs that detailed Emerson’s confidentiality obligations. One subparagraph, for example, barred Emerson from disclosing information about BladeRoom that it shared with Emerson “at any time.” Another subparagraph stressed that any disclosed information “shall remain the property of” BladeRoom and “shall not confer” “any rights or license whatsoever” to Emerson. The NDA’s third paragraph stated that the confidentiality obligations did not apply to information that was in “or hereafter comes into[] the public domain, otherwise than by reason of breach of” the NDA. With all that in mind, the NDA’s twelfth paragraph is most pertinent to our review.

Paragraph twelve says:

The parties acknowledge and agree that their respective obligations under this agreement shall be continuing and, in particular, they shall survive the termination of any discussions or negotiations between you and [BladeRoom] regarding the Transaction, provided that this agreement shall terminate on the date 2 years from the date hereof.

(emphasis added).

Eventually, the acquisition fell through. And around the same time that the deal fell through, Facebook began plans to build a large data center in Northern Sweden. BladeRoom hoped its technology would catch Facebook’s eye, so BladeRoom pitched a design for the data center in July 2012. Several months later, Emerson also pitched a data center design to Facebook. On October 30, 2012, Facebook verbally approved Emerson’s design. At the time, Emerson’s design was only ten percent done. Despite the approval, Facebook contacted BladeRoom almost a year later to ask about updates to BladeRoom’s proposal.

In the end, only BladeRoom and Emerson competed to design and build the data center. In November 2013, Facebook selected Emerson’s proposal and the two signed a design-build contract in March 2014.

Meanwhile, BladeRoom knew nothing until March 2014 about the data center’s design Emerson pitched. A year later, BladeRoom sued Facebook and alleged that the data center’s design copied BladeRoom technology. BladeRoom ended up filing an amended complaint that asserted claims against Emerson and its subsidiaries. 1 The second amended complaint lodged several claims against Emerson, notably, breach of contract and trade secret misappropriation. In 2018, the parties tried the case before a jury.

Halfway through the trial, BladeRoom settled with Facebook; its case against Emerson went on. Before closing arguments, Emerson proposed a jury instruction that would have excluded any information disclosed or used after August 17, 2013 from its liability for breach of contract— which was, as Emerson argued, the date of the contract’s expiration. 2 The district court denied the instruction but allowed Emerson to make the same legal argument to the jury.

1 We call the defendants “Emerson” collectively. 2 August 17, 2013 was exactly two years after the NDA took effect.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 15

The next day, BladeRoom moved in limine to overturn that ruling. BladeRoom specifically moved to prohibit Emerson from arguing that, as a matter of law, the NDA’s twelfth paragraph allowed Emerson to use BladeRoom’s confidential information two years after signing the NDA.

The district court granted the motion and agreed that the NDA’s confidentiality obligations did not expire under paragraph twelve. To interpret the NDA, the district court first examined its “purpose and context.” Because “the purpose of the contract [was] to protect information, not provide for its release after 2 years,” the district court found that “a reasonable businessperson in either party’s position would not have contemplated Emerson’s [reading].” And because “Emerson’s [reading] would lead to an absurd result and would create some inconsistency with the rest of the [NDA,]” the district court found that the NDA’s confidentiality obligations survived beyond two years.

After the trial evidence closed, the jury issued a special verdict. The jury found that Emerson breached the NDA and willfully and maliciously misappropriated BladeRoom’s trade secrets. The jury also found that both actions harmed BladeRoom or unjustly enriched Emerson. For damages, the jury found that BladeRoom sustained $10 million in lost profits and $20 million in unjust enrichment. But the jury’s verdict did not separate damages for breach of contract from misappropriation of trade secrets or future lost profits from past lost profits.

The district court later awarded BladeRoom $30 million in punitive damages. The district court chose $30 million because “[t]he trial evidence show[ed] that either [the breach of contract or misappropriation] claim for which the jury found liability could support the [full] amount of compensatory damages.” Yet, in two post-verdict orders, the district court admitted that “there is no way for the parties or the court to know how much was awarded for breach of contract and how much was awarded for misappropriation of trade secrets.”

The district court also awarded BladeRoom prejudgment interest starting from October 30, 2012, the date when the district court found that BladeRoom suffered its “injury . . . [and] its loss.” The district court chose the date because it was then that BladeRoom “was notified” that “it had lost the opportunity to obtain Facebook’s data center contract.” And last, the district court awarded BladeRoom roughly $18 million in attorneys’ and expert witness’ fees. On appeal, Emerson challenged the orders discussed above, along with several other district court orders that do not affect our holding. 3

II.

We review de novo the district court’s order interpreting the NDA. Trs. of S. Cal. IBEW-NECA Pension Tr. Fund v. Flores, 519 F.3d 1045, 1047 (9th Cir. 2008). We first explain why the district court’s interpretation erred. After, we explain why the legal error prejudiced Emerson and why we remand for a new trial.

A.

English courts interpret contracts to discern the contracting parties’ intent. Arnold v. Britton [2015] UKSC 36, 2015 WL 3555408, [15]. English courts determine intent based on “what a reasonable person having all the

3 Emerson appealed the orders related to attorneys’ and expert witness’ fees in 20-15758, 20-15759, and 20-15760.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 17

background knowledge which would have been available to the parties would have understood” the contract’s terms to have meant. Id. (quotation omitted); see also Rainy Sky S.A. v. Kookmin Bank [2011] UKSC 50, 2011 WL 5077782, [14]. English courts therefore “focus[] on the meaning of the relevant [contractual] words . . . in their documentary, factual, and commercial context.” Arnold v. Britton [2015]

UKSC 36, [15].

The focus is a “unitary exercise” that “balance[s]” textual and contextual analyses. Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, 2017 WL 01084489, [12]. To analyze text, English courts read contract terms as a whole, and in their “natural and ordinary meaning.” Arnold v. Britton [2015] UKSC 36, [15]. To analyze context, English courts examine “the overall purpose” of a contract and its clauses. Id. They also look at “the factual background known to the parties at or before the date of the contract.” Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [10]; see also Arnold v. Britton [2015] UKSC 36, [15]. English courts also consider the commercial effects of proposed interpretations using “commercial common sense.” Arnold v. Britton [2015] UKSC 36, [15]. And last, English courts “disregard[] subjective evidence of any party’s intentions.” Id.

English contract interpretation “is not a literalist exercise focused solely on a parsing of the wording of the particular clause.” Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [10]. Rather, some contracts “may be successfully interpreted principally by textual analysis,” but some contracts may require a deeper focus on the facts known to the parties. Id. at [13]. Thus, we interpret the NDA mainly through a textual analysis because when “parties have used unambiguous language, the court must apply it.” Rainy Sky S.A. v. Kookmin Bank [2011] UKSC 50, [23]. The parties’ narrow dispute focuses on the proviso in paragraph twelve. We hold that paragraph twelve’s natural meaning unambiguously terminated the NDA and its confidentiality obligations two years after it was signed. Paragraph 12 says:

The parties acknowledge and agree that their respective obligations under this agreement shall be continuing and, in particular, they shall survive the termination of any discussions or negotiations between you and the Company regarding the Transaction, provided that this agreement shall terminate on the date 2 years from the date hereof.

(emphasis added).

A proviso is “[a] clause in a legal or formal document, making some condition, stipulation, exception, or limitation.” Proviso, Oxford English Dictionary (3d ed. 2007). Both BladeRoom and the district court reasoned that the proviso in paragraph twelve limited only the “discussions or negotiations” clause. Under that reasoning, paragraph twelve therefore deemed “information disclosed [by BladeRoom] during the 2-year lifespan” as “confidential and subject to a continuing obligation against [] disclosure or use, but any information disclosed by [BladeRoom] after 2 years [was] not subject to th[e] restriction.” By contrast, Emerson argued that the proviso limited all of paragraph twelve. Under Emerson’s reading, although the parties’ obligations under the NDA continued through any negotiations, their obligations ended after two years. We find Emerson’s reading best follows the plain text and the whole contract’s natural meaning for several reasons.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 19

First, the term “provided” naturally means “on the condition, supposition, or understanding (that).” Provided, Oxford English Dictionary (3d ed. 2007). Thus, the proviso means: on the condition that this agreement terminates in two years. Preceding the proviso is the statement that the parties’ “obligations under this agreement shall be continuing” and “shall survive the termination of any discussions or negotiations.” Altogether, paragraph twelve plainly mandated that the parties’ obligations created by the agreement were continuing and did not terminate just because negotiations ended—with the condition that the agreement terminated in two years.

What is more, paragraph twelve’s mandate is textually sound. The conjunctive “and” separates the first and second clauses in paragraph twelve. The two clauses are therefore read “side by side.” And, Oxford English Dictionary (3d ed. 2008). In turn, the proviso modifies both clauses. Simply put, paragraph twelve mandated that “this agreement shall terminate”—not that only the “discussions or negotiations” “shall terminate[.]”

The NDA’s other paragraphs also bolster the plain mandate of paragraph twelve. For one, the NDA used the phrase “this agreement” several times and each time the phrase referenced the entire NDA. Thus, the phrase has a fixed natural meaning. It would therefore be unnatural to assume that “this agreement” in paragraph twelve refers to anything but the whole NDA. See Bank of Credit & Comm. Int. S.A. v. Ali [2001] UKHL 8, 2001 WL 171941, [8] (“[T]he [C]ourt reads the terms of the contract as a whole, giving the words used their natural and ordinary meaning in the context of the agreement[.]”).

Likewise, the various confidentiality obligations detailed in subparagraphs of paragraph two do not alter or contradict paragraph twelve’s mandate that “this agreement shall terminate on the date 2 years from the date hereof.” No subparagraphs in paragraph two refer to or limit paragraph twelve. And BladeRoom formatted the NDA so that paragraph twelve is in a different section (and on a different page) than the confidentiality subparagraphs. Thus, a reasonable person reading the NDA would find that the subparagraphs in paragraph two do not affect paragraph twelve’s command, unless it is explicitly stated somewhere else. See Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [12] (“[O]ne [must] read the language in dispute and the relevant parts of the contract that provide its context.”). On the whole, a textual analysis supports Emerson’s position.

In contrast, BladeRoom’s textual analysis not only twisted the ordinary meaning of words, but it also spawned absurdity. BladeRoom proposed that paragraph twelve mandated only the actual “discussions or negotiations” between it and Emerson terminated after two years—not Emerson’s confidentiality obligations. Yet the proposal passes over the clear phrase that “this agreement shall terminate”—not that “discussions or negotiations” “shall terminate.” And ignoring that phrase creates bizarre outcomes. If, for example, either party stopped negotiating within two years, then the party would have breached the NDA. Similarly, if the parties kept negotiating after two years, then both parties would have breached the NDA. Either way, BladeRoom’s textual analysis is inapposite.

Beyond the textual flaws, BladeRoom’s reliance on two English cases that interpreted different NDA provisions also lacked merit. One case interpreted a provision that plainly stated, “[t]he terms and conditions contained in this Agreement shall continue to apply whether or not the Parties conclude an agreement for joint participation in the

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 21

Business.” Pers. Hygiene Servs. Ltd & Ors v. Rentokil Initial UK Ltd. [2014] EWCA Civ 29, 2014 WL 287664, [9] (emphasis added). And the natural reading of that specific provision is a far cry from the clause in BladeRoom’s NDA. In the second case, the English court relied on equitable principles to create a confidentiality duty. BBC v. HarperCollins Publishers Ltd. [2010] EWHC 2424 (Ch), [2011] E.M.L.R. 6, [47–50]. Yet BladeRoom never asserted an equitable argument below. In each case, BladeRoom failed to persuasively connect its textual analysis with English case law.

The district court’s reasoning also failed to save BladeRoom’s analysis. First, the district court never analyzed paragraph twelve’s natural and ordinary meaning. Instead, the district court relied on three contextual factors: “the purpose and context of the [NDA,]” “evidence in the trial record[,]” and whether Emerson’s construction would “create some inconsistency with the rest of the [NDA.]” 4 Although “[t]extualism and contextualism are not conflicting paradigms” under English contract law, the district court must “balance[] the indications given by” closely examining the relevant text and the factual background. Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [12–13] (emphasis added). Thus, the district court’s sole reliance on contextual reasoning was insufficient.

4 Deciding whether a proposed reading aligns with the rest of the NDA is a common textual tool used to interpret contracts. See Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [12] (explaining that “read[ing] the language in dispute and the relevant parts of the contract” is separate from a contextual analysis focusing on the factual background). But the district court never relied on the NDA’s words or structure to conclude whether Emerson’s reading aligned with the rest of the NDA. The district court’s reasoning therefore relied only on context. Admittedly, English courts may place “greater emphasis” on contextual interpretations for more informal contracts. Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [13]. Those informal contracts often lack “skilled professional assistance” or completeness. Id. But the NDA is far from informal. Indeed, mature companies negotiating a large international deal and prefacing their negotiations with a detailed NDA exemplifies formality. The NDA should therefore be “interpreted principally by textual analysis, [] because of [its] sophistication and complexity and because [it has] been negotiated and prepared with the assistance of skilled professionals.” Id.

After properly conducting both textual and contextual analyses, BladeRoom’s proposed reading remains unreasonable. As the district court found, the NDA’s purpose was to “allow the exchange of confidential information in connection with a possible acquisition.” And the confidentiality obligations in paragraph two highlight that purpose. But the NDA plainly limits its overall purpose to only two years. If the purpose were to last beyond two years, then the parties would have never agreed to the straightforward proviso in paragraph twelve. Rather, the parties would have rewritten it to reflect a different timespan. In fact, the parties did just that in another paragraph that barred Emerson from soliciting BladeRoom’s employees or clients for “a period of 18 months.” (emphasis added). In sum, although the NDA purported to protect BladeRoom’s technology, paragraph twelve limited its overall purpose to two years. See Arnold v. Britton [2015] UKSC 36, [15] (assessing a contract’s meaning “in light of . . . the overall purpose of the clause and the [contract]”) (emphasis added).

The surrounding facts known to the parties also failed to support BladeRoom’s reading of the NDA. Although few

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 23

facts exist in the record from which we can draw inferences, the district court noted that one party intended the NDA’s confidentiality obligations to last beyond two years. That evidence is irrelevant because English contract law disregards a party’s subjective intent. Arnold v. Britton [2015] UKSC 36, [15]. Instead, the relevant facts known to the parties show that BladeRoom drafted the NDA, their agent signed it, sent it to Emerson, whose agent then signed it. English courts presume ambiguity in a commercial contract against the drafting party. Persimmon Homes Ltd. v. Ove Arup & Partners Ltd. [2017] EWCA Civ 373, 2017 WL 02212888, [52]. So, if paragraph twelve were ambiguous, then the factual context would create a presumption against BladeRoom. 5 But paragraph twelve’s command is straightforward and thus the inferences or presumptions that would arise from the factual context are unhelpful.

The last relevant presumption in English contract interpretation is “commercial common sense.” Arnold v. Britton [2015] UKSC 36, [15]. When faced with “two possible constructions, the court is entitled to prefer the construction which is consistent with business common sense . . . .” Rainy Sky S.A. v. Kookmin Bank [2011] UKSC 50, [21]. But the presumption is also unhelpful here because Emerson’s reading is the only sensible reading grounded in textual and contextual analyses.

Although the district court reasoned that Emerson’s reading would create “an uncontemplated windfall[,]” our

5 At the same time, the presumption “has a very limited role” in commercial contracts between parties with equal bargaining power. Persimmon Homes Ltd. v. Ove Arup & Partners Ltd. [2017] EWCA Civ 373, [52]. analysis remains sound. In fact, English courts are “very slow to reject the natural meaning of a provision as correct simply because it appears to be a very imprudent term for one of the parties to have agreed.” Arnold v. Britton [2015] UKSC 36, [20]. It could very well be the case that BladeRoom “may have agreed to something which with hindsight did not serve [its] interest.” Wood v. Capita Ins. Servs. Ltd. [2017] UKSC 24, [11] (citing Arnold v. Britton [2015] UKSC 36, [20, 77]). Accordingly, we will not read into whether the parties’ interpretations make commercial common sense.

In all, a reasonable person in the parties’ situation would have read the NDA and understood that the confidentiality obligations terminated after two years. The district court therefore erred as a matter of law when it granted BladeRoom’s motion in limine. We reverse the district court’s order.

B.

We also vacate the district court’s judgment and remand for a new trial.

At its core, the district court’s error prevented the jury from hearing Emerson’s chief defense. Given the legal error, the panel must “give judgment after an examination of the record without regard to errors or defects which do not affect the substantial rights of the parties.” 28 U.S.C. § 2111. Put differently, if the error is harmless, then we must affirm the judgment.

The parties disagree about how to label the district court’s error and they disagree on our standard for reviewing the error. Emerson compared the error to a jury instruction error. But BladeRoom seemed to suggest that the error was

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 25

like an incorrect denial of a motion for judgment as a matter of law (“JMOL”). We treat the district court’s action as an error of jury instruction for two reasons.

First, parties are “entitled to an instruction about [their] theory of the case if it is supported by law and has foundation in the evidence.” Clem v. Lomeli, 566 F.3d 1177, 1181 (9th Cir. 2009) (quotation omitted). Emerson therefore proposed a jury instruction that would have excluded any information disclosed or used after August 17, 2013 from its liability for breach of contract. But the district court denied the proposal. It reasoned that “the [NDA] is in evidence, [so] Emerson can argue any exclusion from liability it believes should apply.” BladeRoom then moved in limine the very next day to preclude Emerson from arguing that it could use confidential information after August 17, 2013. Thus, BladeRoom’s motion in limine was a clear attempt to stymie the district court’s jury instruction ruling that allowed Emerson to “argue any exclusion from liability it believes should apply.” And in turn the district court reversed its jury instruction order as to Emerson’s ability to advocate by prohibiting Emerson from arguing that the NDA’s twelfth paragraph excluded it from breach of contract liability.

Second, the district court’s order lacked basic findings necessary for us to construe it as a denial of a motion for JMOL. An order denying JMOL would have found that a “reasonable jury would not have a legally sufficient evidentiary basis to find” that Emerson did not breach the NDA under either party’s proposed reading. Fed. R. Civ. P. 50(a)(1). Yet the district court never made that finding. For those two reasons, we treat the district court’s error as an error of jury instruction.

And we review a jury instruction error in a civil trial for harmless error. Clem, 566 F.3d at 1182. We first presume prejudice. Id. Thus, “the burden shifts to [BladeRoom] to demonstrate that it is more probable than not that the jury would have reached the same verdict had it been properly instructed.” Id. (internal quotations marks and citation omitted).

We apply Clem’s burden shifting standard because Ninth Circuit precedent instructs us to do so given the abnormal jury instruction error in these appeals. In contrast, for the ordinary civil appeal, “the party seeking reversal” bears the burden of persuasion. Shinseki v. Sanders, 556 U.S. 396, 410 (2009). But Shinseki’s harmless error review does not apply to these appeals for several reasons.

First, the Supreme Court said nothing in Shinseki about the standard the panel must apply to determine whether there was prejudice for a civil jury verdict. Shinseki involved appellate review of administrative proceedings. Id. at 406. Harmless error review in Shinseki therefore occurred under 38 U.S.C. § 7261(b)(2). Id. As the Court explained, the § 7261(b)(2) standard is merely “the same kind of ‘harmless- error’ rule that courts ordinarily apply in civil cases.” Id.

Second, Shinseki explained that what ordinarily happens in civil appeals is “the party seeking reversal normally must explain why the erroneous ruling caused harm.” Id. at 410. Indeed, “it normally makes sense to ask the party seeking reversal to provide an explanation, say, by marshaling the facts and evidence showing the contrary.” Id. (emphasis added). Shinseki explained that appellate courts should avoid “mandatory presumptions and rigid rules” and favor a “case-specific application of judgment,” id. at 407, and courts can employ “generalizations about what kinds of errors are likely” based on experience, id. at 411 (citing Kotteakos v. United States, 328 U.S. 750, 760–61 (1946)). So long as the generalizations “influence, though not control,

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 27

future determinations,” Shinseki allows courts to employ different harmless error review frameworks. See id. In short, Shinseki did not create a bright-line rule that precludes us from applying Clem’s burden shifting provision.

Admittedly, the Sixth Circuit recently held:

In Shinseki, the Supreme Court expressly rejected the idea of placing the burden of proving harmlessness on the appellee in a civil case, announcing that “we have placed such a burden on the appellee only when the matter underlying review was criminal.” 556 U.S. at 410–11. In other words, Shinseki holds that in a civil case, the tie goes to the verdict.

Kocher v. Durham Sch. Servs., L.P., 969 F.3d 625, 630 (6th Cir. 2020) (citation truncated).

But the Sixth Circuit’s reading instilled words into Shinseki that the majority declined to use. See id. at 630 n.4 (detailing the so-called “tie-to-the-verdict” rule in civil cases). For one, Shinseki never used the term “tie-to-the- verdict.” For another, Shinseki never endorsed a rigid rule in civil cases like the “tie-to-the-verdict” rule—it endorsed the opposite. See 556 U.S. at 407 (“We have previously warned against courts[] determining whether an error is harmless through the use of mandatory presumptions.”).

Even if Shinseki were to create the rule, the rule would apply only in “the ordinary civil case.” Id. at 411 (emphasis added). Because the Shinseki majority described only harmless error review in a “normal” or “ordinary” civil appeal, Shinseki did not create the bright-line “tie-to-the- verdict” rule for all civil cases. See generally id. at 410–11. At its core, Ninth Circuit precedent under Clem addresses the abnormal civil cases when there is an errant jury instruction. 566 F.3d at 1182. The jury instruction error here, like in Clem, was abnormal because it affected Emerson’s substantive rights. 566 F.3d at 1181 (“Each party is . . . ‘entitled to an instruction about [their] theory of the case if it is supported by law and has foundation in the evidence.’”). Likewise, Clem recognizes the unique circumstances when “[p]rejudice is also generally more likely than not if nothing about the jury’s verdict indicates that the result would have been the same without the error.” 566 F.3d at 1183 (cleaned up). It follows that Clem’s harmless error review adheres to Shinseki because it is a generalized framework that does not “control[] future determinations.” 556 U.S. at 411. And other Ninth Circuit cases have employed Clem’s harmless error standard when there is a substantive jury-instruction error without controversy. See, e.g., Tekoh v. Cnty. of Los Angeles, 985 F.3d 713, 725 (9th Cir. 2021); Gantt v. City of Los Angeles, 717 F.3d 702, 707 (9th Cir. 2013).

Although several Ninth Circuit cases have applied Shinseki’s harmless error review instead of Clem’s, those cases differ greatly from the present appeals. We have, for example, applied Shinseki in cases that reviewed agency adjudications—not civil jury trials like the present appeal. 6 Still, Shinseki’s analysis explained there is no “relevant distinction between the manner in which reviewing courts

6 Al Haramain Islamic Found., Inc. v. U.S. Dep’t of Treasury, 686 F.3d 965, 989 (9th Cir. 2012) (reviewing the Office of Foreign Assets Control designating an Islamic organization a ‘specially designated global terrorist’); Ludwig v. Astrue, 681 F.3d 1047, 1054 (9th Cir. 2012) (reviewing social security benefits denial); Molina v. Astrue, 674 F.3d 1104, 1118–21 (9th Cir. 2012) (same).

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 29

treat civil and administrative cases.” 556 U.S. at 407. But our cases applying Shinseki did not involve the abnormal situation of jury instruction errors like the errors in Clem or the present appeals. 7 And the Sixth Circuit’s Kocher decision involved evidentiary errors. 969 F.3d at 628–29. At bottom, Clem recognized the abnormality that arises from jury instruction errors because the errors prevent the jury from “determin[ing] the issues presented intelligently.” 566 F.3d at 1182 (quoting Fikes v. Cleghorn, 47 F.3d 1011, 1013 (9th Cir. 1995)). Our other case law, therefore, does not conflict with Clem.

In all, Shinseki cautions appellate courts to avoid “mandatory presumptions and rigid rules.” 556 U.S. at 407. Ninth Circuit case law has created a framework that properly influences—not controls—how panels review abnormal jury instruction errors. See, e.g., Tekoh, 985 F.3d at 725; Gantt, 717 F.3d at 707; Clem, 566 F.3d at 1182–83. With all that in mind, BladeRoom failed to meet the burden outlined in Clem.

The district court’s error prejudiced Emerson when the jury made its breach of contract, misappropriation, and damages findings. Based on the special verdict, there is no way to know whether the jury would return the same answers if the district court had allowed Emerson to present its chief defense. See id. at 1183 (“Prejudice is also generally more likely than not if nothing about the jury’s verdict indicates that the result would have been the same without the error.”) (cleaned up). We therefore vacate the jury’s findings relating to breach of contract, misappropriation, and

7 Al Haramain Islamic Found., Inc., 686 F.3d at 988 (procedural due process issues); Ludwig, 681 F.3d at 1055 (ex parte communications); Molina, 674 F.3d at 1122 (evidentiary-related issues). damages and remand for a new trial on those issues. See People v. Cepeda, 851 F.2d 1564, 1568 (9th Cir. 1988) (“[An] appellate court has broad discretion to grant a new trial on all or only some issues.”) (citations omitted). We now address in turn why we must vacate each of those jury findings.

First, and most important, we vacate the jury’s breach of contract findings. The jury answered only whether a breach of contract occurred—not when the breach occurred. Without the crucial finding of when the breach occurred, there is no way to determine whether the jury would have found that Emerson breached the NDA if the jury had known that Emerson’s confidentiality obligations ended after only two years. See Clem, 566 F.3d at 1182. In short, the jury could have found breach under two theories—either during the open-ended period after Emerson’s confidentiality obligations end, or during the initial two-year period—but we do not know which one the jury actually found. See Tex. Advanced Optoelectronic Sols, Inc. v. Renesas Elecs. Am., Inc., 895 F.3d 1304, 1316 (Fed. Cir. 2018) (“The general rule is that ‘if a jury could find liability according to multiple theories, and one of them is [legally] erroneous, we reverse unless we can tell that the jury came to its decision using only correct legal theories.’”) (alterations in original) (quoting Rodriguez v. Riddell Sports, Inc., 242 F.3d 567, 577 (5th Cir. 2001)).

Our holding in Clem requires us to find prejudice in the district court’s error here. In Clem, the district court misinformed the jury that it could only find liability under one theory, when, in reality, the jury could have found liability under a second theory. See 566 F.3d at 1180–81 (explaining that the deliberate indifference instruction misstated that the defendant must “act” even though “failure

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 31

to act” could also establish liability). Because the jury’s verdict failed to establish liability under the instructed theory and because the jury may have found liability under the uninstructed theory, we reversed and remanded. Id. at 1183.

Admittedly, the errant instruction in Clem essentially added an extra element to the plaintiff’s burden of proof, and that did not happen here. Id. But the distinction does not sway our overall analysis. The reversal in Clem was grounded in two reasons, each of which was sufficient to find prejudice. We first pointed out “that when the trial court erroneously adds an extra element to the plaintiff’s burden of proof, it is unlikely that the error would be harmless.” Id. at 1182 (cleaned up). And second, we explained that “prejudice is also generally more likely than not if nothing about the jury’s verdict indicates that the result would have been the same without the error.” Id. at 1183 (cleaned up). Based on the second reason, it is clear that we must find prejudice here.

To be sure, the evidence on the timing of when breach occurred is murky. Some evidence shows that breach could have occurred during the NDA’s initial two-year span. Other evidence shows that breach could have occurred after the two years passed. Thus, no party can demonstrate what the jury found as to when breach occurred, and BladeRoom cannot refute the presumption of prejudice. See Gambini v. Total Rental Care, Inc., 486 F.3d 1087, 1093 (9th Cir. 2007) (“[T]he prevailing party is not entitled to have disputed factual questions resolved in [its] favor . . . .”) (quoting Swinton v. Potomac Corp., 270 F.3d 794, 805–06 (9th Cir. 2001)). As a result, “not only is it impossible to determine that the jury would have reached the same result, there are signs that the jury might very well have returned a different verdict.” Hoard v. Hartman, 904 F.3d 780, 791 (9th Cir. 2018). Accordingly, we vacate the jury’s breach of contract findings.

Likewise, we vacate the jury’s misappropriation findings. Under California law, misappropriation cannot occur if someone “discloses his trade secret to others who are under no obligation to protect the confidentiality of the information.” Altavion, Inc. v. Konica Minolta Sys. Lab., Inc., 226 Cal. App. 4th 26, 57 (2014) (quotations and citations omitted). Because the misappropriation claim hinged on Emerson’s written—not implied—obligations in the NDA, the same reasoning that lead us to vacate the breach of contract verdict applies. There is simply no way to know whether the jury found that Emerson misappropriated trade secrets during or after the NDA’s initial two-year period. Without that finding, we also do not know whether the jury would have found that Emerson acted willfully and maliciously. After all, the jury never heard Emerson’s chief defense. For those reasons, “nothing about th[e jury’s] verdict” suggests that the jury would still find willful and malicious misappropriation without the error excluding the NDA evidence and so we vacate the jury’s misappropriation findings. Clem, 566 F.3d at 1183 (alterations in original).

Finally, the same reasons support vacating the jury’s damages findings. The jury awarded damages in two lump sums that did not separate damages for breach of contract from misappropriation. As a result, we have no way to know how much damages the jury intended to allocate for each claim. And because we do not know whether the jury would have found Emerson liable under breach of contract or misappropriation, we cannot tell whether the damages findings would be the same. Id. at 1183. We therefore vacate the jury’s damages findings.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 33

“[W]e do not take lightly the decision to reverse a jury verdict.” VHT, Inc. v. Zillow Grp., Inc., 918 F.3d 723, 748 (9th Cir. 2019). But the district court’s error “went to the heart of” Emerson’s defense and improperly precluded it from the jury’s consideration. Hoard, 904 F.3d at 791. “[W]e are convinced that this verdict for [BladeRoom], if allowed to stand, would be a legally unjustified windfall to [BladeRoom] and a miscarriage of justice.” Autohaus Brugger, Inc. v. Saab Motors, Inc., 567 F.2d 901, 915 (9th Cir. 1978); see also Obrey v. Johnson, 400 F.3d 691, 701 (9th Cir. 2005) (recognizing that we find harmful error when “the error itself had substantial influence . . . or if one is left in grave doubt”) (alterations in original) (quoting Kotteakos v. United States, 328 U.S. 750, 765 (1946)). “Because this error likely prejudiced the outcome of the case and—left uncorrected—would contribute to a miscarriage of justice, we vacate the district court’s judgment and remand for a new trial.” Hoard, 904 F.3d at 787.

Given our order to vacate the judgment and remand for a new trial, we vacate the district court’s post-verdict orders on appeal. 28 U.S.C. § 2106 (An appellate court “may . . . vacate, set aside or reverse any judgment, decree, or order of a court lawfully brought before it for review, and may remand the cause and . . . require such further proceedings to be had as may be just under the circumstances.”). We also vacate the district court’s attorneys’ and expert witness fees orders in the 20-15758, 20-15759, and 20-15760 appeals. Id.

III.

Given the complex issues on appeal, we also address a couple of issues for consideration on the awards of damages and prejudgment interest should they be determined after a new trial. California Civil Code § 3426.3(a) allows an aggrieved party to recover actual damages and any unjust enrichment on a trade secret misappropriation claim. Because the jury here found that Emerson willfully and maliciously misappropriated BladeRoom’s trade secrets, the district court awarded punitive damages equal to the $30 million in compensatory damages.

Under California law, however, a party cannot collect punitive damages for breach of contract awards. Applied Equip. Corp. v. Litton Saudi Arabia Ltd., 7 Cal. 4th 503, 516 (1994). Because the jury awarded a lump sum for breach of contract and misappropriation damages, Emerson correctly argued that the district court could not assume that the whole award went to misappropriation.

Courts that apply California law use three factors to award punitive damages. Neal v. Farmers Ins. Exch., 21 Cal. 3d 910, 928 (1978). The second factor, “the amount of compensatory damages awarded,” id., helps balance a punitive damages award with actual harm that is caused. See id. (“[I]n general, even an act of considerable reprehensibility will not be seen to justify a proportionally high amount of punitive damages if the actual harm suffered [] is small.”). When weighing the second factor, the district court here merely noted that under Neal, 21 Cal. 3d at 928, “compensatory damages are a ‘relevant yardstick’ for [punitive] damages.” And, in a footnote, the district court explained that “[t]he trial evidence shows that either [the breach of contract or misappropriation] claim for which the jury found liability could support the amount of compensatory damages it awarded.”

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 35

The district court misapplied the second factor for two reasons. One, the district court abused its discretion based on an earlier order denying Emerson’s motion to compel. And second, the district court never explained why it allocated compensatory damages the way it did.

First, in the earlier order, the district court found that “there is no way for the parties or the court to know how much was awarded for breach of contract and how much was awarded for misappropriation of trade secrets.” To be sure, the district court confirmed that “only the jury knows the number.”

A district court abuses its discretion when it “reaches a result that is illogical, implausible, or without support in inferences that may be drawn from the facts in the record.” United States v. Hinkson, 585 F.3d 1247, 1262 n. 21 (9th Cir. 2009) (collecting cases from other circuits). It was illogical for the district court to find how much the jury awarded for each claim when it already found that it had “no way” to do just that. See id. The district court should have explained why its reasoning changed. Instead, after awarding punitive damages, the district court issued another order that adopted the same reasoning in the order that denied Emerson’s motion to compel. In brief, the district court flipped its reasoning twice without explaining why. And the inconsistency visibly benefited BladeRoom because it cost Emerson millions in punitive damages that the district court would not have awarded otherwise. In sum, the inconsistency was “beyond the pale of reasonable justification” and an abuse of discretion. Harman v. Apfel, 211 F.3d 1172, 1175 (9th Cir. 2000).

Second, apart from the logical inconsistency, the district court never explained why it allocated compensatory damages under the second punitive damages factor. We review damages allocations for abuse of discretion. Hemmings v. Tidyman’s Inc., 285 F.3d 1174, 1195 (9th Cir. 2002). 8 Even though the evidence “show[ed] that either claim for which the jury found liability could support the amount of compensatory damages it awarded[,]” the district court allocated the full $30 million in compensatory damages to the misappropriation claim. But the allocation was inadequate because although it “may well represent a permissible allocation of the damages between the two theories, . . . the [district] court gave no explanation for its decision, leaving us guessing whether this exercise of discretion is a permissible one.” Gibson v. Moskowitz, 523 F.3d 657, 667 (6th Cir. 2008).

On remand, the district court must take several steps to allocate damages. First, the district court should consider adopting a more-detailed special verdict form. The verdict form may, for example, ask the jury to allocate damages between breach of contract and misappropriation. If the jury can allocate damages accordingly, then the district court would avoid trouble on remand. Yet, if the jury cannot allocate damages, then the district court should take another step to explain why it allocated damages. That way, the parties are not left guessing at the allocated amounts.

8 The week before oral argument, BladeRoom filed a letter of supplemental authority under Circuit Rule 28(j). In the letter, BladeRoom cited—for the first time—Passantino v. Johnson & Johnson Cons. Prods., Inc., 212 F.3d 493, 509 (9th Cir. 2000), to suggest that the district court did not abuse its discretion to award punitive damages. Rule 28(j) prohibits a party from raising new issues not raised in the briefs. United States v. LaPierre, 998 F.2d 1460, 1466 n. 5 (9th Cir. 1993). Thus, BladeRoom should have raised the argument in its response brief, or, at the very least, it should have raised the argument earlier than a week before oral argument. Id.

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 37

B.

California Civil Code § 3288 allows the award of prejudgment interest “[i]n an action for the breach of an obligation not arising from contract, and in every case of oppression, fraud, or malice.” 9 California Civil Code § 3287(b) allows a court to award prejudgment interest but only for breach of contract claims.

Awarding prejudgment interest under one section rather than the other can make a sizeable difference. For instance, § 3288 prejudgment interest runs from the date that property was lost. Greater Westchester Homeowners Assn. v. City of Los Angeles, 26 Cal. 3d 86, 103 (1979). By contrast, § 3287(b) prejudgment interest runs from “no [] earlier than the date the action was filed.”

The district court awarded prejudgment interest on the full compensatory damages award under § 3288, not § 3287(b). It also found that prejudgment interest should run from October 30, 2012. Those decisions were erroneous.

First, § 3288’s plain text bars prejudgment interest awards for breach of contract claims. See Greater Westchester Homeowners Assn., 26 Cal. 3d at 102 (“[U]nlike [§ 3287], which relates to liquidated and contractual claims, [§ 3288] permits discretionary prejudgment interest for unliquidated tort claims.”). Yet the district court flouted § 3288’s plain bar when it awarded prejudgment interest for a lump sum involving a breach of contract claim. If the district court grants prejudgment interest on remand, then it must award prejudgment interest

9 The parties stipulated for the court to award prejudgment interest under § 3288. for any breach of contract damages under § 3287(b), not § 3288.

Moreover, to award § 3288 prejudgment interest for BladeRoom’s lost profits, the district court relied on a factual error. Section 3288 prejudgment interest runs from the date that property was lost. Greater Westchester Homeowners Assn., 26 Cal. 3d at 103. The district court found that BladeRoom’s injury and loss occurred on October 30, 2012, because “BladeRoom was notified it had lost the opportunity to obtain [the Facebook contract].” But the evidence refutes that finding. The evidence showed that “Facebook gave a verbal approval . . . for [Emerson’s] design and concept.” on October 30, 2012. At the time, the proposal was only “10 percent done.” Almost a year later, Facebook asked BladeRoom for an update on its proposal. And, in March 2014, Facebook awarded the contract to Emerson. In all, no evidence suggests that BladeRoom became aware that it had lost the contract on October 30, 2012. Instead, the evidence may suggest that date was when Emerson knew that it could win the contract. The district court’s October 30, 2012 finding therefore lacked a factual basis and a link to BladeRoom’s lost profits.

If the district court awards § 3288 prejudgment interest on remand for BladeRoom’s lost profits, then it would have to find when the lost profits began. See Greater Westchester Homeowners Assn., 26 Cal. 3d at 102 (“[Section 3288] prejudgment interest [is] ‘awarded to compensate a party for the loss of his or her [p]roperty.’”) (quoting Bullis v. Sec. Pac. Nat. Bank, 21 Cal. 3d 801, 815 (1978)). Put simply, § 3288 prejudgment interest awards “represent[] the accretion of wealth which money or particular property could have produced during a period of loss.” Id. at 102–03. Although the district court relied on Lakin v. Watkins

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 39

Associated Indus., 6 Cal. 4th 644, 663 (1993), to award § 3288 prejudgment interest from “the date of the injury[,]” the case law rejects that approach. First, the California Supreme Court has explained that § 3288 prejudgment interest runs from “the date of loss of the property.” Greater Westchester Homeowners Assn., 26 Cal. 3d at 103 (quoting Bullis, 21 Cal. 3d at 815). And second, the California Supreme Court’s holding in Lakin did not specifically apply § 3288. Rather, Lakin opined about the “general” purpose of prejudgment interest. 6 Cal. 4th at 663. And the distinction makes sense because California’s Civil Code has three prejudgment interest statutes that each deal with different legal claims. Cal. Civil Code §§ 3287, 3288, 3291. Thus, any § 3288 prejudgment interest award for BladeRoom’s lost profits must run from when the lost profits began.

Should the district court be called on to award § 3288 prejudgment interest for unjust enrichment after a new trial, the district court should make two findings. First, it should find whether § 3288 allows prejudgment interest for unjust enrichment damages. And second, if so, then the district court should apply the same standard that it applied for lost profits—it should find when the unjust enrichment began. See Greater Westchester Homeowners Assn., 26 Cal. 3d at 103.

IV.

We REVERSE the district court’s order granting the motion in limine as it relates to the NDA’s twelfth paragraph. We VACATE the district court’s judgment and post-verdict orders on appeal and REMAND for a new trial consistent with this opinion. Given the new trial, the orders related to attorneys’ and expert witness’ fees in the 20- 15758, 20-15759, 20-15760 appeals are also VACATED. RAWLINSON, Circuit Judge, concurring:

I concur fully in the majority opinion. I write separately to bring two additional matters to the attention of the district court.

1. In the event that the jury on retrial again imposes liability on Emerson Electric for unjust enrichment, it would be BladeRoom’s burden to prove the amount of damages. See Ajaxo, Inc. v. E*Trade Fin. Corp., 48 Cal. App. 5th 129, 185 (2020) (indicating that plaintiff has the “burden of proving damages [in a trade secret case] by showing the misappropriation, the subsequent commercial use, and . . . evidence by which the jury can value the rights the defendant has obtained”). It is not sufficient for the damages expert to attribute the entire value of the sale of Emerson Electric’s Hyperscale division to the misappropriation of BladeRoom’s trade secret despite admitting that the Hyperscale division had business and value that did not involve use of the misappropriated trade secret. See 02 Micro Int’l Ltd. V. Monolithic Pwr. Sys., 399 F. Supp. 2d 1064, 1076 (N.D. Cal. 2005) (rejecting expert damages testimony that calculated damages “based on an assumption that all of the trade secrets were misappropriated” when the jury found that only some of the trade secrets were misappropriated). In 02, the district court determined that a damage calculation that was not consistent with the evidence “was useless to the jury” because any jury award would be “based on speculation and guesswork, not on evidence.” Id. at 1077.

Similarly, in Hilderman v. Enea Teksci, Inc., No. 05cv1049BTM (AJB), 2010 WL 546140 at *2 (S.D. Cal. Feb. 10, 2010), the district court found the expert’s damages opinion “unreliable because it rest[ed] on [an] unfounded assumption[]” that the “entire Goodwill value” was

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 41

attributable to the allegedly misappropriated trade secrets (internal quotation marks omitted). See also Wyatt Tech. Corp. v. Malvern Instr. Inc., No. CV 07-8298 (ABC)(MANx), 2010 WL 11505684 at *3 (C.D. Cal. Jan 25, 2010) (“In calculating lost profits, a plaintiff must prove in a reasonable manner the link between the injury suffered and the illegal practices of the defendant.”) (quoting City of Vernon v. S. Cal. Edison Co., 955 F.2d 1361, 1371 (9th Cir. 1992) (emphasis in the original). The court determined that the damages calculations were “flawed and inadmissible at trial because they were not limited to [Defendant’s] alleged wrongdoing.” Id. at *4.

In addition, “lost anticipated profits cannot be recovered if it is uncertain whether any profit would have been derived at all from the proposed undertaking.” Food Safety Net Svcs. v. Eco Safe Sys. USA, Inc., 209 Cal. App. 4th 1118, 1132 (2012) (citation and alteration omitted). More specifically, “lost profits based on a future contract cannot be recovered when the contract is uncertain or speculative.” Id. (citation omitted). Consequently, BladeRoom will not be entitled to future profits absent definitive evidence of a future contract that was lost due to Emerson Electric’s misappropriation of Bladeroom’s trade secret. See Sargon Enters., Inc. v. Univ. of S. Calif., 55 Cal. 4th 747, 775 (2012) (explaining that there is no recovery for “claimed lost profits” that are “uncertain, hypothetical and entirely speculative”) (citation and internal quotation marks omitted). If the asserted loss profit testimony is not based on “objective evidence of past volume of business or any other provable data,” the testimony should be excluded. Id. at 778 (citation and internal quotation marks omitted). At bottom, there must be “a substantial similarity between the facts forming the basis of the project projections and the business opportunity that was destroyed.” Id. (citation and internal quotation marks omitted).

2. BladeRoom concedes that Facebook and Emerson were joint tortfeasors and that they “conspired” to misappropriate BladeRoom’s trade secrets. With that concession, California law required an offset. See Calif. Civ. Proc. Code § 877(a); see also Dell’Oca Bank of NY Trust Co., N.A., 159 Cal. App. 4th 531, 561 (2008). In the event a retrial results in the imposition of damages against Emerson Electric, the court should apply an offset for the amount of the settlement between BladeRoom and Facebook. See Calif. Civil Proc. Code § 877(a) (providing that when one tortfeasor settles a case, that settlement “shall reduce the claims against the other[] [tortfeasor] in the amount [of the settlement]”); see also Dell’Oca, 159 Cal. App. 4th at 561 (construing § 877 broadly to allow “an offset for sums paid to settle plaintiffs’ claims against the other defendants”).

Correspondingly, Emerson Electric would be entitled to discovery of the settlement terms. See Fed. R. Civ. P. 26(b)(1) (authorizing discovery of “any nonprivileged matter that is relevant to any party’s claim or defense and proportional to the needs of the case”); see also Phillips ex rel. Ests. of Byrd v. Gen. Motors Corp., 307 F.3d 12506, 1212 (9th Cir. 2002) (indicating that “confidential settlement information” may be produced under appropriate circumstances); Burke v. Regalado, 935 F.3d 960, 1048 (10th Cir. 2019) (holding that the district court “erred when it declined to order disclosure of the settlement agreement” when the settlement agreement was “relevant” and “necessary” to resolving the case). Any concerns regarding unauthorized disclosure of the settlement terms may be addressed by a protective order fashioned by the district court. See Phillips, 307 F.3d at 1211 (explaining that courts

BLADEROOM GROUP LTD. V. EMERSON ELECTRIC 43

have “broad discretion . . . to decide when a protective order is appropriate and what degree of protection is required:); see also St. Bernard Par. v. Lafarge N. Am., Inc., 914 F.3d 969, 975 (5th Cir. 2019) (explaining that “discovery of confidential settlement agreements is generally available under an appropriate protective order”).