I. FACTS AND PROCEDURAL BACKGROUND

A. Factual Background To help explain our conclusions in these appeals, we set out in some detail the factual background of the relationship between HP and Oracle, their products, and the events leading up to the relevant dispute.1

1. HP’s Itanium Servers and Strategic Relationship with Oracle A server is a computer system that performs tasks too big or complex for a personal computer or notebook. HP is a computer technology company that in 2010 manufactured, among other products, computer servers. HP’s high-end “Itanium” servers (hereafter Itanium or Itanium servers) are the technology at the center of these appeals. Oracle is a technology company that develops software for business clients and, in the relevant time period, was a significant supplier of software for the high-end server market, including for HP’s Itanium servers. Itanium servers run on HP’s proprietary Unix (“HP-UX”) operating system and use the Itanium microprocessor, which HP jointly developed with Intel Corporation. The hardware and operating system together form the server “platform.” The platform is fitted with software; together they comprise the “technology stack”—essentially “layers” of hardware and software that work together to deliver an integrated product to HP’s customers. Database software constitutes an essential layer in the technology stack. Oracle is a major provider of database software for the high-end server market. Over 80 percent of 1 These facts are taken from evidence presented in the first and second phases of the trial. Except where noted, the facts presented are not in dispute. 3 HP’s Itanium systems use Oracle’s database software. Oracle also provides “middleware” software, which sits between the database software and the software applications. “Porting” is the process of taking software that has been written on one operating system and processor architecture—like Oracle’s database, middleware, and software applications—and making it available on another system, like HP’s Itanium platform. The porting process is most involved when software must be configured and tested to work with a new server platform, requiring a significant investment of time and resources, especially by the software provider. The process continues even after software is established on a platform, as software vendors constantly prepare new software releases, which need to be tested and tuned to work on the platform. HP started to sell Itanium in 2002. In 2005, the CEOs of HP, Oracle, and Intel jointly launched Itanium on a larger scale. HP and Oracle worked together to port Oracle’s database and middleware products to Itanium and did so without any contract or payment to Oracle. HP and Oracle each bore the costs associated with the work it had done. Similarly, HP and Oracle did not sign a contract or make payments when Oracle ported subsequent releases of database or middleware software to Itanium. HP later contracted to pay Oracle up to $10.3 million to port one of its application software products, the E-Business Suite, to Itanium. HP provided similar funding, under contract, to Oracle to port a few other application products to Itanium. However, most of Oracle’s porting work to Itanium—about 99 percent—happened without any contract or payments between HP and Oracle. Once Oracle ports to a platform, it typically guarantees ongoing support under its lifetime support policy to Oracle’s customers on that platform. A variety of technological, market, and cost factors thus influence the decision to port to a platform. The decision is a discretionary one based on business objectives and made directly by Oracle’s CEO. Only a “small percentage” of Oracle’s porting decisions are governed by 4 contract. Oracle may stop porting to a platform either because the hardware vendor is moving customers to a new platform, or because Oracle has decided not to develop new releases for the platform even though the hardware vendor continues to market and sell it. The only example offered at trial of Oracle choosing to stop developing software for a platform still being sold was for IBM’s Power processor running on Linux. It offered no such examples involving HP. The Itanium product line was highly profitable for HP, generating over $2 billion in annual profits in 2010. According to Ann Livermore,2 who at the time was the senior vice-president of HP’s enterprise business, Itanium was not HP’s biggest business “but in many ways it was [HP’s] most important business from a customer perspective, because . . . customers were running such important applications, and [HP] had a lot of really big customers who used those products.” Oracle’s decision to port its software to Itanium benefitted both Oracle and HP. Oracle was able to deploy its software on HP’s platform, selling database, middleware, and some application products, while HP had “the leading software products” available on its servers and could provide customers with an integrated and functional solution. Marketing material from 2009 designed for use with customers touted the “HP and Oracle Global Alliance” based on “[o]ver 25 years of collaborative partnership,” “[m]ore than 140,000 joint customers,” and “[m]ore than $6 billion” in annual revenue generated by the joint business. It described HP’s and Oracle’s joint support of solutions and summarized each company’s substantial market share in the other’s products: 41 percent of Oracle’s database customers used HP systems, and 84 percent of Itanium servers ran Oracle’s database software.

2 Livermore testified in both trial phases. At the time of her testimony, she served on HP’s board of directors and had worked for the company for 29 years. From 2004 to 2011, during the period covering the dispute with Oracle, Livermore was the senior vice- president of HP’s enterprise business. 5 Oracle ported its software to HP server platforms other than Itanium, including to HP’s platforms that preceded Itanium. For example, when HP stopped selling its PA- RISC servers in December 2008, Oracle continued to port to PA-RISC after that date. These actions enabled Oracle to continue selling software upgrades into the installed base for customers who continued to use the platform. There were a few products that Oracle elected not to port to Itanium unless HP agreed to share costs, including two software applications developed in 2010 (Fusion Applications and Cluster File Software). HP decided that it did not want to pay Oracle for this service, so Oracle never ported those applications.

2. Oracle’s Acquisition of Sun Microsystems and Employment of Mark Hurd Certain events in 2009 and 2010 strained HP’s relationship with Oracle. In April 2009, Oracle announced that it would acquire Sun Microsystems (Sun). Sun was a leading computer hardware company and competitor of HP, including in the server market. The acquisition was completed in January 2010. The acquisition of Sun marked a “potential sea change” in the relationship between HP and Oracle. The central question was whether Oracle would continue to make its software available on HP’s hardware platforms now that it owned one of HP’s direct competitors. At meetings between senior executives for HP and Oracle in the months after the Sun acquisition, Oracle emphasized that it was first and foremost a software company and would continue to offer its software in a platform-neutral way on the HP, IBM, and Sun platforms. At a February 2010 meeting attended by David Donatelli, the executive in charge of HP’s enterprise business, and Thomas Kurian, Oracle’s most senior software executive, the jointly prepared meeting minutes reflect that “HP and Oracle remain strategic partners” and that “Oracle’s intent” was for Itanium to “have release parity” with other major platforms and to “maintain the same functional parity” as other UNIX

6 operating systems.3 HP sought further assurance at an April 2010 meeting attended by senior management, including HP’s senior vice-president Livermore and Oracle’s then co-president Safra Catz. Catz reiterated Oracle’s commitment to the alliance with HP and to “running Oracle Software on HP Infrastructure.” HP understood from these discussions that the companies would compete for hardware business, but Oracle would continue to support its customers’ choice of server platforms, including HP’s platforms. There was no indication that Oracle would refuse to port the next release of its products to Itanium. HP understood that Oracle intended to maintain the same course of action as it had previously, which was to offer its software on Itanium through the life of the platform. This understanding shaped HP’s investment decisions, leading for example to a multi-year commitment to pay Intel for continuing development work on the Itanium chip. In August 2010, HP’s CEO Mark Hurd resigned at the request of HP’s board of directors. Oracle’s CEO Larry Ellison publicly faulted the HP board of directors for asking Hurd to resign, calling it “ ‘the worst personnel decision since the idiots on the Apple board fired Steve Jobs many years ago.’ ” Ellison offered Hurd a position at Oracle as co-president.4 About one month later, Oracle announced on September 6, 2010, that it had hired Hurd as Oracle’s co-president responsible for all marketing and sales activity. This announcement prompted immediate concerns for HP because of Hurd’s separation agreement and his unique knowledge of HP’s confidential financial, technological, and

3 “Release parity” refers to how the software vendor releases its next generation of products. It means that Oracle would release new versions of its software on Itanium in the same time frame as its release on competing server platforms. “Functional parity” refers to HP’s Itanium server maintaining its place among the top tier or “Tier 1” UNIX operating systems. 4 Oracle at the time had two presidents. Safra Catz was the co-president focusing on day-to-day operations. Catz and Hurd later became Oracle’s co-CEOs. Ellison became Oracle’s chief technology officer and chairman of the board of directors. 7 customer information. The following day, HP filed a civil action against Hurd for breach of contract and for threatened misappropriation of trade secrets. HP’s complaint alleged that Hurd would be unable to perform his duties for Oracle in his new position as co- president and member of the board of directors “without necessarily using and disclosing HP’s trade secrets and confidential information.” Oracle responded to HP’s lawsuit against Hurd by issuing a press release. In the press release, Ellison called HP “ ‘an important partner’ ” and said that “ ‘[b]y filing this vindictive lawsuit against Oracle and Mark Hurd, the HP board is acting with utter disregard for that partnership, our joint customers, and their own shareholders and employees. The HP Board is making it virtually impossible for Oracle and HP to continue to cooperate and work together in the IT marketplace.”

3. Settlement Agreement and Reaffirmation of the HP-Oracle Partnership

a. Negotiation of Settlement Agreement The lawsuit against Hurd, and Ellison’s widely reported response, prompted rapid action by HP and Oracle. HP’s senior vice-president Livermore contacted Oracle co- president Catz on September 7, the day HP filed its lawsuit against Hurd.5 That evening, HP’s general counsel, Michael Holston, spoke with Oracle’s general counsel, Dorian Daley. Both lawyers expressed a desire to resolve the lawsuit quickly and, if possible, to preserve the companies’ relationship. Holston and Daley spoke broadly about HP’s desired terms for any settlement, which they contemplated would be confidential but would allow the parties to jointly announce a resolution to the dispute over Hurd. On September 9 and 10, HP and Oracle exchanged term sheets in an attempt to reach an agreement to resolve the Hurd controversy. These first exchanges dealt strictly with terms specific to Hurd and Hurd’s activities at Oracle. The next day, September 11,

5 The events referenced in connection with HP’s lawsuit against Hurd all took place in 2010, unless otherwise specified. 8 Livermore and Catz spoke by phone. Livermore told Catz that HP’s board was concerned about damage to the companies’ relationship—especially from Ellison’s public comments—and wanted any settlement to contractually reaffirm the existing partnership. HP wanted an assurance that the companies would continue to operate as they had before Oracle hired Hurd. Catz did not object. At trial, she characterized reaffirmation of the HP-Oracle partnership as “a very modest ask.” Livermore was not surprised by Catz’s stance, as the HP-Oracle relationship was mutually beneficial and profitable. Over the next two days, Oracle and HP negotiated the language of the proposed agreement. After some back-and-forth, Livermore and Catz concurred that the agreement should not put HP in a better position than it had enjoyed with respect to Oracle prior to Oracle’s decision to hire Hurd. On September 12, Oracle sent a proposed agreement to HP. Oracle had removed language suggesting it would favor HP over other competitors. Oracle’s September 12 version of the reaffirmation clause contained two sentences. It reaffirmed the companies’ “commitment to their longstanding strategic relationship and their mutual desire to continue to support their mutual customers,” and it provided that “Oracle will continue to offer its product suite on HP platforms and HP will continue to support Oracle products (including Oracle Enterprise Linux and Oracle VM) on its hardware in a manner consistent with that partnership.” In her cover e-mail to the September 12 draft, Daley clarified Oracle’s position that the reaffirmation clause was not intended “to put HP in a better position tha[n] it currently enjoys or result in the negotiation of a new contractual commitment.” Daley wrote that the discussions between Catz and Livermore did not address “anything more tha[n] an agreement to continue to work together as the companies have – with Oracle porting products to HP’s platform and HP supporting the ported products and the parties engaging in joint marketing opportunities – for the mutual benefit of customers.” 9 Catz and Livermore spoke again the next morning. Catz was not receptive to the idea of adding more specifics to the reaffirmation clause. Livermore eventually agreed with Catz that “it was simpler and cleaner” to exclude specifics. Livermore testified at trial that she was satisfied the provision addressed her concern about Oracle continuing to offer its products on HP’s Itanium platform. Livermore acknowledged that Catz told her “that she didn’t think that Oracle was taking on any new obligations or work.” Catz told Livermore that Oracle would commit to maintaining the course of action it had done in the past, not any better. Catz e-mailed Oracle’s general counsel Daley after her call with Livermore, saying “she tried to add specifics on the relationship and I wasn’t open.” Later that day, HP’s lawyers sent a revised draft to Daley that added specific terms back into the proposed agreement, including an express reference to porting to Itanium. Oracle forcefully rejected the added language. Daley wrote to Holston that this was “setting us back in a very big way.” Livermore admitted that the draft “got by” her before she had updated HP’s lawyers about her conversation with Catz. Livermore spoke with Catz and told her it was a mistake. They agreed to take out the detail added by HP. Oracle returned a draft the next day that deleted the additions that HP had proposed. It retained the reaffirmation language from the September 12 draft, stating that the companies “reaffirm their commitment” to their partnership and that “Oracle will continue to offer its product suite on HP platforms . . . in a manner consistent with that partnership.” HP made only one more change to the reaffirmation clause, adding that the parties were reaffirming their relationship “as it existed prior to Oracle’s hiring of Mark Hurd.” Daley, Catz, and Holston each confirmed that the added language made it clear that the parties were referring to the partnership and course of dealing before the controversy erupted over Oracle’s hiring of Hurd.

10 b. Text and Announcement of Final Settlement Agreement The parties executed the settlement agreement between HP, Oracle, and Hurd on September 20, 2010 (the agreement). Paragraph 1 of the agreement comprises the reaffirmation clause. It states: “Reaffirmation of the Oracle-HP Partnership. Oracle and HP reaffirm their commitment to their longstanding strategic relationship and their mutual desire to continue to support their mutual customers. Oracle will continue to offer its product suite on HP platforms, and HP will continue to support Oracle products (including Oracle Enterprise Linux and Oracle VM) on its hardware in a manner consistent with that partnership as it existed prior to Oracle’s hiring of Hurd.” Recital B of the agreement, which precedes the provisions forming the parties’ specific commitments, reinforces that “HP, Hurd and Oracle recognize the mutual advantages of the continuation of the HP-Oracle partnership and its benefits to their joint customers and prospects and now desire to further their business relationship and resolve the [lawsuit against Hurd] without the further time and expense of litigation.” The agreement also provides for a joint press release, included as an attachment to the agreement. The press release, issued on September 20, announced that HP and Oracle had resolved the litigation regarding Hurd’s employment at Oracle in a confidential settlement agreement that “also reaffirms HP and Oracle’s commitment to delivering the best products and solutions to our more than 140,000 shared customers.” It quoted Ellison as saying that “ ‘Oracle and HP will continue to build and expand a partnership that has already lasted over 25 years.’ ” Apart from the press release, a confidentiality provision in the agreement prevents further public disclosure about the lawsuit or the terms of the settlement.

4. Course of Dealing Between HP and Oracle In their briefing, the parties present divergent narratives of how each side interpreted and implemented its obligations under the reaffirmation clause. Oracle maintains that HP’s conduct in the months following the agreement was inconsistent with

11 a belief that the reaffirmation clause imposed broad new obligations on either party. HP responds that, just as with the parties’ course of dealing before the agreement, the course of dealing after the agreement focused on continued product development and porting work. The trial court heard extensive evidence in phase 1 of the trial on course of conduct and post-agreement conduct, which we discuss in more detail in the analysis, post (part II.A.2.b.).

a. Prior to the Agreement As stated in the negotiated joint press release announcing the agreement, HP and Oracle had “more than 140,000 shared customers” and had enjoyed a strategic partnership of “ ‘over 25 years.’ ” The core of Oracle’s and HP’s partnership consisted of joint sales, marketing, and mutual support of their products, including the porting of Oracle’s products to HP’s platforms. Although Oracle had ported its software in the past to different HP platforms, as of September 2010, when the agreement was signed, Itanium was the only HP server platform to which Oracle was porting new versions of its software products.6 Nine Oracle software products comprised the “product suite” offered on Itanium at the time of the agreement. Once Oracle completed the initial port of one of these major software products to Itanium, Oracle ported all future releases of the same product to the platform. The vast majority of porting (over 99 percent) occurred without any written contracts between HP and Oracle. The few instances documented at trial in which the companies entered a porting contract involved an initial port to Itanium of a product that

6 Oracle also offered its products on HP’s industry standard servers, which use Intel’s x86 chips and Windows or Linux operating systems (not HP-UX). However, Oracle did not need to port its products to those servers because it developed its software from the outset to work on all industry-standard servers that use Windows or Linux operating systems (of which there are many besides those of HP). 12 had not been previously ported. Oracle made subsequent versions of the product available without a contract. With respect to Itanium’s predecessor platforms, Oracle continued to support the platform by releasing new versions of software on it even after HP stopped sales of that server line.

b. After the Agreement The HP executives responsible for the relationship with Oracle were informed that a settlement had been reached and they should carry on with “business as usual.” They were not told of specific terms of the agreement, beyond the contents of the press release, and understood that HP’s obligation was to behave as it had been doing before the signing of the agreement. HP and Oracle continued to engage in product development and porting work in the post-agreement period. For example, Oracle had begun work in early 2010 to port the next version of its database software (database 12g, later 12c) to Itanium. HP provided hardware, servers, storage, and engineering support as needed to assist with the ports. In September 2010, HP provided additional servers to Oracle to facilitate the ongoing database porting work. HP’s engineers continued to collaborate with Oracle’s engineers on the next release of the database software until as late as June 2011. HP also worked with Oracle to ensure that Oracle’s product suite would continue to run on Itanium, even increasing the level of support to certify the Oracle products on specific HP servers. Livermore and Catz spoke “off and on” in the months after the agreement was signed, including on March 18, 2011, four days before Oracle announced that it would discontinue developing new releases of its products for Itanium. Catz gave no indication to HP that Oracle was going to stop porting to Itanium. Livermore reported in her notes to several HP executives that the phone call had a “[p]ositive tone overall.” Catz testified that she said nothing to Livermore during the call about Oracle’s impending Itanium announcement because Oracle “hadn’t made the final decision yet.”

13 5. Oracle’s March 2011 Announcement According to Oracle, by March 2011 there were clear signals from within the industry that the Itanium platform was losing viability. Ellison testified that technological innovations in other server platforms led to Itanium “falling further and further behind,” and that “Intel had lost interest in Itanium.”7 On March 22, 2011, Oracle issued a press release in which it stated it would stop developing software for Itanium (March 2011 announcement). The March 2011 announcement was titled “Oracle Stops All Software Development For Intel Itanium Microprocessor” and stated that after conversations with Intel about its strategic focus and Itanium’s future, “Oracle has decided to discontinue all software development on the Intel Itanium microprocessor. . . . [¶] . . . [¶] Oracle will continue to provide customers with support for existing versions of Oracle software products that already run on Itanium.” Oracle issued a second announcement the next day reiterating its support of the “current versions” of its software on Itanium and specifying the next versions of Oracle’s software that would not be available on Itanium but would be available on other platforms (i.e., IBM and Oracle/Sun). Oracle issued the March 2011 announcement at 11:00 p.m. (Eastern Standard Time) the night before HP’s annual shareholders’ meeting and caught both HP and Intel unaware. In response, Intel released a press release the next day, March 23, to “directly reiterate” that its work on the Itanium processer was continuing unabated, “ ‘with multiple generations of chips currently in development and on schedule.’ ” Oracle claimed that its decision to discontinue porting future Oracle software versions to Itanium had no effect on most joint customers, since Itanium users made up only “a few thousand” of the 150,000 joint customers, and the remaining joint customers

7 Ellison described in detail the developments he believed signaled Itanium’s demise. He explained by way of one example that “an innovative new product, [Intel’s E7 microprocessor], killed an old, obsolete product, Itanium. That’s the way it works in Silicon Valley. That’s the way it’s always worked.” 14 used x86-based servers, for which Oracle had no plans to stop developing new software. Oracle also emphasized that the decision would have no immediate effect because it would continue to support and develop patches for the products already offered on Itanium in accordance with its standard lifetime support policy. Oracle’s executives nonetheless recognized that the decision “would be big news” to HP. Thomas Kurian, Oracle’s executive vice-president of software development, acknowledged that the March 2011 announcement was unprecedented and was the first time Oracle had decided to stop porting to any server based on a microprocessor architecture that was still being sold and marketed. Customers who were running Oracle’s software on Itanium would have to choose another hardware platform to receive future releases of Oracle’s software. Livermore called Catz immediately after the March 2011 announcement to tell her it contradicted the press release from just six months ago and was a “breach of our contract.” HP initially hoped that pressure from customers would cause Oracle to reverse its decision. However, customer pressure did not change Oracle’s decision, and about two months later HP sent a demand letter to Oracle and subsequently filed this action. B. Procedural Background

1. HP’s Lawsuit HP sued Oracle in June 2011. The complaint alleges that Oracle breached the settlement agreement by refusing to continue to offer its product suite on Itanium just months after it promised to do so and by reneging on its assurances of continued support for customers using Oracle software on the platform. HP asserted claims for declaratory relief, breach of contract, breach of the implied covenant of good faith and fair dealing, promissory estoppel, and various tort and statutory violations.8 HP sought a judicial

8 This appeal involves only HP’s claims for declaratory relief, breach of contract, and breach of the implied covenant of good faith and fair dealing. Although Oracle filed a cross-complaint (and later, an amended cross-complaint) alleging that HP violated federal and state law by misleading Oracle, its customers and investors, and the public,

15 determination of the parties’ rights and obligations under the agreement, which HP asserted “requires Oracle to continue to offer and support” any Oracle product that was offered on the Itanium platform at the time Oracle signed the agreement with HP. HP demanded specific performance and requested direct and consequential damages as a remedy for the alleged breach of contract. The trial court bifurcated trial proceedings into two phases: (1) a phase 1 bench trial to interpret the agreement and decide the issues of declaratory relief and promissory estoppel; and (2) a phase 2 jury trial to decide the breach and damages claims. The trial court agreed to set a short timeline for the phase 1 bench trial, which in HP’s words would allow the court to “resolve the contractual dispute in sufficient time to ensure that, if HP prevails, Oracle can complete necessary development work on its 12g database in a timely fashion so as not to delay the porting of its 12g database to Itanium when it launches 12g on competing server platforms.”

2. Phase 1 Bench Trial The phase 1 bench trial took place over 12 days in June 2012. The trial court heard testimony from 30 witnesses and admitted over 500 exhibits into evidence. The trial court issued its ruling on August 28, 2012. In a detailed, 45-page final statement of decision (statement of decision), the trial court ruled in favor of HP and against Oracle on HP’s claim for declaratory relief. The trial court made factual findings related to the period before the litigation over Oracle’s hiring of Hurd, the negotiation of the settlement agreement, the companies’ historical partnership and course of dealing

about the future prospects of Itanium, and by fraudulently inducing Oracle to enter into the settlement agreement over the Hurd dispute, the jury in the second phase of trial in 2016 rejected Oracle’s claim under the Lanham Act (15 U.S.C. § 1125(a)), and Oracle has not appealed that ruling here. As noted below, after HP succeeded on its breach of contract and implied covenant claims, HP voluntarily dismissed the remaining causes of action. 16 (particularly as to the Oracle product suite on HP’s platforms, porting, and platform support), and the post-agreement period and March 2011 announcement. As to the legal issues, the trial court held that the reaffirmation clause (paragraph 1 of the agreement) was unambiguous as a matter of law and that Oracle had “failed to offer a plausible interpretation” of it. The trial court found the extrinsic evidence, although not admissible to interpret the agreement’s unambiguous meaning, was confirmatory of HP’s interpretation. It reasoned that the contractual language that “ ‘Oracle will continue to offer its product suite on HP platforms’ . . . ‘in a manner consistent with [the HP-Oracle] partnership’ ” was consistent with the parties’ historical relationship and prior course of dealing, and with the language of the joint press release pledging continued support for customers that had for decades relied on the HP-Oracle partnership. The trial court concluded that the agreement “requires Oracle to continue to offer its product suite on HP’s Itanium-based server platforms and does not confer on Oracle the discretion to decide whether to do so or not.” The trial court construed the term “ ‘product suite’ ” to mean “Oracle software products that were offered on HP’s Itanium- based servers at the time Oracle signed the [agreement], including any new releases, versions or updates of those products.” The court also construed Oracle’s obligation under the agreement to apply without charge to HP and “until such time as HP discontinues the sale of its Itanium-based servers.” Over Oracle’s objection, the court declined to construe “a host of collateral issues” that Oracle claimed had to be resolved concurrently, such as the effect of the reaffirmation clause on intellectual property rights in any resulting software and on the parties’ prior porting agreements.

3. Oracle’s August 2012 Press Release and September 2012 Announcement On the day the trial court released its tentative statement of decision for phase I, Oracle issued a press release stating it would appeal (August 2012 press release). The

17 August 2012 press release asserted that “ ‘Oracle did not give up its fundamental right to make platform engineering decisions in the 27 words HP cites from the settlement of an unrelated employment agreement. . . . We plan to appeal the Court’s ruling while fully litigating our cross claims that HP misled both its partners and customers.’ ” HP responded that Oracle “no longer has any basis for refusing to port, and that it should . . . resume porting immediately.” One month later, Oracle publicly announced that it would comply with the trial court’s phase 1 decision and resume porting to Itanium (September 2012 announcement). The September 2012 announcement indicated that “a judge recently ruled that Oracle has a contract to continue porting its software to Itanium computers for as long as HP sells Itanium computers. Therefore, Oracle will continue building the latest versions of its database and other software covered by the judge’s ruling. . . . Oracle software on HP’s Itanium computers will be released on approximately the same schedule as Oracle software on IBM’s Power systems.” Oracle wrote separately in a letter to the trial court that its September 2012 announcement was “without prejudice to [its] rights to appeal” the phase 1 ruling. Oracle told the trial court that it would be able to “meet in a timely manner substantially all of the porting obligations” found by the court to exist under the settlement agreement, since few products had had releases in the interim. Oracle told the trial court that given its decision to resume porting in time to meet its obligations, “HP will need to substantially revise its damages case” which would necessitate a delay of phase 2 of the trial. The trial court allowed the parties to serve supplemental expert reports addressing the impact of Oracle’s September 2012 announcement on HP’s damages claim and granted limited reopening of discovery related to expert reports.9

9 Oracle filed a petition for writ of mandate in this court on January 17, 2013, challenging the trial court’s order allowing HP to submit supplemental expert reports as

18 4. HP’s Expert Testimony on Damages In March 2012, HP’s damages expert, economist Jonathan Orszag, calculated HP’s estimated damages due to Oracle’s breach of contract to be between $3.8 billion and $4 billion. Orszag produced a supplemental written report in December 2012. In his supplemental report, Orszag wrote that the “continued decline in Itanium revenue” reflected in updated projections “shows that any favorable impact from the Phase 1 decision and the Oracle September 2012 announcement has been more than outweighed by the continuing negative impact . . . from the March 2011 Oracle Announcements and the continuing uncertainty created by Oracle’s recent statements regarding its intention to appeal the Phase 1 decision.” At the evidentiary hearing held in March 2013 to determine the admissibility under Sargon Enterprises, Inc. v. University of Southern California (2012) 55 Cal.4th 747 (Sargon) of each party’s designated expert’s testimony, Orszag explained that Itanium’s business “continued to deteriorate” following the September 2012 announcement. Orszag did not assume, in making his damages calculations, that Itanium customers would be unable to get the Oracle products they wanted. He explained that during the period between March 2011 and September 2012, many Itanium customers decided to transition to other server platforms. He noted that Oracle’s statements regarding its intention to appeal the phase 1 decision factored into the “significant uncertainty” about the future availability of Oracle products on Itanium. According to Orszag’s analysis, the “damage had been done.”

well as limitations on the supplemental discovery it allowed Oracle to conduct. This court denied the petition without opinion. (Oracle Corp. v. Superior Court (Mar. 27, 2013, H039210), petn. den.) This court also denied, without opinion, Oracle’s earlier petition for writ of mandate, filed on October 12, 2012, seeking to overturn the trial court’s decision in phase 1. (Oracle Corp. v. Superior Court (Jan. 31, 2013, H038880), petn. den.) 19 The trial court issued a written order after the evidentiary hearing finding that the expert testimony of Orszag, and of Oracle’s proposed expert, Ramsey Shehadeh, met the admissibility standard under Sargon, supra, 55 Cal.4th 747.10

5. Anti-SLAPP Motion and Appeal Shortly before the trial court issued its order finding Orszag’s expert testimony admissible, Oracle filed a motion under the anti-SLAPP statute (Code Civ. Proc., § 425.16) to strike “in whole or in relevant part” HP’s causes of action for breach of contract, breach of the implied covenant of good faith and fair dealing, and promissory estoppel (anti-SLAPP motion). Oracle’s motion asserted that HP had “changed its damages theory from one about harm caused by the unavailability of Oracle’s software to one about harm caused by customer uncertainty . . . allegedly resulting from Oracle’s March 2011 announcement and its subsequent refusal to accept the [statement of decision] as the final, definitive ruling on the meaning of the Hurd Settlement Agreement.” Oracle contended that its announced intention to appeal the trial court’s determination on liability was protected conduct under the anti-SLAPP statute, because it constituted an exercise or attempt to exercise Oracle’s constitutional rights both to freedom of speech and to petition the government for redress of grievances. (See Code Civ. Proc., § 425.16, subd. (b)(1).) A hearing on Oracle’s anti-SLAPP motion was held on April 5, 2013, the last court day before the phase 2 trial was set to begin. That same day, Oracle filed a motion in limine seeking to exclude argument and evidence of lost profits that failed to disaggregate lost profits arising from Oracle’s stated intent to appeal. Oracle claimed that HP was precluded from asserting damages based on Oracle’s constitutional right to petition under the United States and California Constitutions and under California’s

10 We discuss additional facts relevant to the jury’s damages verdict and Oracle’s related contentions on appeal in part II.C., post. 20 litigation privilege. HP opposed the motion in limine, which remained pending while the parties litigated the anti-SLAPP motion. In a written order on April 8, 2013, the trial court denied Oracle’s anti-SLAPP motion as untimely. The court recognized that “as a result of this ruling, Oracle is statutorily entitled to perfect an appeal, and by taking that step, all matters ‘embraced’ or ‘affected’ by the order appealed from are stayed.” Oracle filed a notice of appeal that same day from the order denying its anti-SLAPP motion. The trial court vacated the phase 2 trial date pending the appeal. In a published opinion on August 27, 2015, a panel of this court affirmed the denial of Oracle’s anti-SLAPP motion. (Hewlett-Packard Co. v. Oracle Corp. (2015) 239 Cal.App.4th 1174, 1196 (Hewlett-Packard) (anti-SLAPP appeal).) The opinion deemed Oracle’s appeal to be “utterly without merit.” (Id. at p. 1178.) Addressing only the question of timeliness, this court held that “[t]he motion was late under any reasonable construction of the facts, and it was quite properly denied because it could not possibly achieve the purposes for which the anti–SLAPP statute was enacted.” (Hewlett-Packard supra, 239 Cal.App.4th at p. 1178.) This court rejected Oracle’s proffered excuse for untimeliness (id. at pp. 1193–1194) and raised other concerns with the anti-SLAPP motion, including that it did not target a “cause of action” or “claim” under the statute (id. at pp. 1195–1196) “but part of HP’s intended proof of causation and damages” (id. at p. 1196), rendering it “in effect a motion in limine.” (Ibid.) It observed that the anti-SLAPP motion in fact “raised the same substantive issues” (ibid.) as the above-mentioned motion in limine filed in anticipation of the phase 2 trial, “and had the matter not been derailed, would have produced a ruling that could be reviewed in due course along with any other issues remaining after trial.” (Ibid.) In its affirmance, this court stated that it was declining to assess sanctions against Oracle only to avoid any further delay of the long-deferred phase 2 jury trial. (Id. at p. 1178.)

21 6. Phase 2 Jury Trial Upon remand of the case following the anti-SLAPP appeal, the parties proceeded with the phase 2 trial to decide the breach and damages claims in May and June of 2016.11 Oracle sought, through multiple procedural vehicles, to challenge HP’s theory of breach and to limit HP’s claim to damages. We describe these efforts in our discussion of the jury’s breach of contract verdict, post (part II.B.1.a.). The trial court rejected Oracle’s motions in limine and proposed motion for summary adjudication of these issues. The jury heard testimony over 19 days from 19 witnesses, including from each side’s damages expert. At the close of HP’s case in chief, Oracle moved unsuccessfully for a judgment of nonsuit on HP’s breach of contract claims. Oracle later moved, at the close of all the evidence, for a directed verdict as to HP’s breach of contract claims, contending that the evidence was legally insufficient to permit a jury finding in HP’s favor and HP had not established damages “by any actionable conduct.” The trial court denied these motions and allowed the contested issues to go to the jury. Over Oracle’s objection, the trial court instructed the jury using the interpretation of the agreement found by the court in the phase 1 statement of decision. The introductory breach of contract instruction (CACI No. 300) stated that it was the court’s duty to interpret the meaning of the agreement at issue in the case, and that at the conclusion of the first trial phase, the court “determined that the Hurd Settlement Agreement is a binding contract between HP and Oracle.” The instruction set forth paragraph 1 of the agreement and instructed the jury that it “must accept as true” the court’s findings regarding the meaning of the agreement. These findings in relevant part stated the following: “1. The [agreement] requires Oracle to continue to offer its product suite on HP’s Itanium-based server platforms and does not confer on Oracle the discretion to decide whether to do so or not. [¶] 2. The term ‘product suite’ means

11 The phase 1 and phase 2 trials were conducted before different bench officers. 22 Oracle software products that were offered on HP’s Itanium-based servers at the time Oracle signed the [agreement], including any releases, versions, or updates of those products. [¶] 3. Oracle’s obligation to continue to offer its product suite on HP’s Itanium-based server platforms lasts until such time as HP discontinues the sale of its Itanium-based platforms. [¶] 4. Oracle is required to port its product suite to HP’s Itanium-based servers without charge to HP.” The instruction further stated in part that “HP claims that Oracle breached this contract when it decided and announced in March of 2011 that it would no longer offer new versions of its product suite on Itanium-based servers, thereafter discontinuing software development and porting work for HP’s Itanium-based server platforms and repeatedly telling customers that it would no longer offer its product suite on Itanium- based servers. [¶] . . . . [¶] Oracle denies that it breached the [agreement] and denies that a breach, if any, caused HP any harm.” After deliberations, the jury delivered its verdict finding that Oracle breached the contract and breached the implied covenant of good faith and fair dealing. The jury awarded HP $1.699 billion in damages for “[p]ast lost profits” and $1.315 billion in “[f]uture lost profits” for a total damages award of $3.014 billion. HP moved for an award of prejudgment interest for the period during which the phase 2 trial was delayed due to Oracle’s anti-SLAPP appeal. The trial court denied the motion after a hearing, finding that while the improper delay weighed in favor of an award, other factors related to HP’s “highly contested and uncertain” damages weighed more significantly against it. We address the motion for prejudgment interest in our analysis of HP’s cross-appeal, post (part II.D.). After the parties stipulated to dismiss and voluntarily dismissed the remaining causes of action, HP requested entry of judgment and “elected not to pursue specific performance as a remedy for its breach-of-contract cause of action.” On October 20,

23 2016, the trial court entered judgment in favor of HP and against Oracle in the amount of $3.014 billion and ordered that HP was entitled to recover allowable costs from Oracle. After briefing and argument, the trial court denied a motion for new trial on damages, filed by Oracle on the ground the jury award was excessive and contrary to law. Oracle timely appealed from the final judgment, and HP timely filed its cross-appeal.

II. DISCUSSION

We must first decide whether the trial court erred in interpreting the agreement’s reaffirmation clause to require Oracle to continue to offer its product suite on Itanium until HP discontinues its sale of the platform. We next consider Oracle’s two-pronged contention that (1) the jury verdict finding breach of contract and breach of the implied covenant of good faith and fair dealing must be reversed because the evidence that Oracle resumed porting its software to Itanium precluded liability for breach, and (2) HP waived any alternative claim for anticipatory breach of the contract by seeking and obtaining Oracle’s specific performance of the agreement. We also consider Oracle’s claim that the trial court erroneously allowed HP to introduce evidence in violation of Oracle’s constitutionally protected and privileged statement that it intended to appeal the phase 1 decision, as well as speculative expert testimony in support of HP’s claim for lost profit damages. Lastly, we address HP’s cross-appeal on the issue of prejudgment interest. A. Interpretation of the Reaffirmation Clause Oracle contends that, contrary to the trial court’s interpretation of the settlement agreement, the reaffirmation clause merely restates the historically voluntary, non- contractual relationship between Oracle and HP. It argues that the unambiguous, plain language of the agreement is not reasonably susceptible to the interpretation HP ascribes to it, and furthermore that the undisputed extrinsic evidence confirms Oracle’s own interpretation. HP responds that the reaffirmation clause plainly commits Oracle to continue porting its software to HP’s Itanium server platform, consistent with Oracle’s prior course of conduct as shown by the extrinsic evidence admitted at trial. 24 1. Principles of Contract Interpretation The fundamental goal of contract interpretation is “to give effect to the mutual intention of the parties as it existed at the time of contracting.” (Civ. Code, § 1636.)12 To interpret a contract, we look to its language (§ 1638) and ascertain the intent of the parties, if possible, based solely on the contract’s written provisions (§ 1639). In doing so, we apply the “ ‘clear and explicit’ meaning of these provisions, interpreted in their ‘ordinary and popular sense,’ unless ‘used by the parties in a technical sense or a special meaning is given to them by usage’ [citation] . . . . Thus, if the meaning a layperson would ascribe to contract language is not ambiguous, we apply that meaning.” (AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 822 (AIU Ins.).) At the same time, we “recognize[] the ‘interpretational principle that a contract must be understood with reference to the circumstances under which it was made and the matter to which it relates. (Civ. Code, § 1647.)’ ” (Mountain Air Enterprises, LLC v. Sundowner Towers, LLC (2017) 3 Cal.5th 744, 752.) It is essentially a judicial function to apply the rules of interpretation to a written contract “so that the purposes of the instrument may be given effect.” (Parsons v. Bristol Development Co. (1965) 62 Cal.2d 861, 865, citing §§ 1635-1661 (Parsons).) In so doing, the trial court may “properly admit[] evidence extrinsic to the written instrument to determine the circumstances under which the parties contracted and the purpose of the contract.” (Parsons, at pp. 864–865.) “ ‘Extrinsic evidence is admissible to prove a meaning to which the contract is reasonably susceptible.’ ” (Iqbal v. Ziadeh (2017) 10 Cal.App.5th 1, 8 (Iqbal).) The court, at least initially, considers “all credible evidence offered to prove the intention of the parties. [Citations.] Such evidence includes testimony as to the ‘circumstances surrounding the making of the agreement . . . including the object, nature and subject matter of the writing . . .’ so that the court can ‘place itself in the same situation in which the parties found themselves at the time of

12 Unspecified statutory references are to the Civil Code. 25 contracting.’ ” (Pacific Gas & E. Co. v. G.W. Thomas Drayage etc. Co. (1968) 69 Cal.2d 33, 39–40, fn. omitted (Pacific Gas).) On appeal, we apply a de novo standard of review when construing the contract, including where conflicting inferences may be drawn from undisputed extrinsic evidence, “unless the interpretation turns upon the credibility of extrinsic evidence.” (Parsons, supra, 62 Cal.2d at pp. 865, 866, fn. 2; accord Garcia v. Truck Ins. Exchange (1984) 36 Cal.3d 426, 439.) Put simply, “when the competent extrinsic evidence is not in conflict, the appellate court independently construes the contract.” (Iqbal, supra, 10 Cal.App.5th at p. 8.) To the extent there is conflicting extrinsic evidence requiring credibility determinations by the finder of fact regarding a meaning of which the contract is reasonably susceptible, we will uphold the trial court’s determination if supported by substantial evidence. (Ibid.; see also Tin Tin Corp. v. Pacific Rim Park, LLC (2009) 170 Cal.App.4th 1220, 1225 (Tin Tin).)

2. Analysis Oracle and HP characterize the plain language of the reaffirmation clause as unambiguous, yet ascribe different meanings to it. As we explain, the parties’ disagreement stems not from an ambiguity in the language or from conflicting extrinsic evidence but from distinct views of what defines the HP-Oracle partnership as set forth in the agreement. Under California’s objective theory of contracts, we must determine “ ‘ “what the outward manifestations of consent would lead a reasonable person to believe.” ’ ” (Allen v. Smith (2002) 94 Cal.App.4th 1270, 1277; see also Iqbal, supra, 10 Cal.App.5th at p. 8.) We begin by analyzing the agreed-upon language, focusing on its plain meaning as much as possible and referencing the extrinsic evidence to ascertain whether it supports an interpretation to which the agreement is reasonably susceptible. (Hess v.

26 Ford Motor Co. (2002) 27 Cal.4th 516, 524.) As the key extrinsic evidence here is not in conflict, we engage in independent review. (Iqbal, supra, 10 Cal.App.5th at p. 8.)

a. Plain Language The reaffirmation clause, set forth in the agreement’s first paragraph, consists of only two sentences. First, “Oracle and HP reaffirm their commitment to their longstanding strategic relationship and their mutual desire to continue to support their mutual customers.” Second, “Oracle will continue to offer its product suite on HP platforms and HP will continue to support Oracle products (including Oracle Enterprise Linux and Oracle VM) on its hardware in a manner consistent with that partnership as it existed prior to Oracle’s hiring of Hurd.” The first sentence addresses two points, namely the companies’ “longstanding strategic relationship” and “mutual desire to continue to support their mutual customers.” The only conduct referenced in the first sentence is that HP and Oracle each “reaffirm their commitment.” The plain and unambiguous meaning from a layperson’s perspective (AIU Ins., supra, 51 Cal.3d at p. 822) is that the companies are abiding by the commitment they have historically shown to (1) each other through their strategic relationship and (2) their shared customers. The second sentence elaborates on this commitment by specifying what each side must do to maintain the partnership. Oracle, for its part, “will continue to offer its product suite on HP platforms” while HP “will continue to support Oracle products (including Oracle Enterprise Linux and Oracle VM) on its hardware.” What is more, each agrees to continue performing its part “in a manner consistent with th[e] partnership as it existed” before Oracle hired Hurd. The second sentence transforms the value statement articulated in the first sentence (i.e., we are committed to our partnership and to our mutual customers) into an actionable and enforceable commitment (i.e., Oracle will continue to offer, and HP will continue to support). Oracle’s general counsel Daley, who

27 drafted the main part of the sentence, conceded at trial that use of the term “will” is mandatory in nature. The latter part of the second sentence also provides a temporal point of reference for the action that each side has agreed it “will continue” to perform. The phrase “in a manner consistent with that partnership as it existed prior to Oracle’s hiring of Hurd” eliminates any doubt that the required action—Oracle continuing to offer its product suite on HP servers and HP continuing to support Oracle products on its hardware—must be consistent with the partnership as it existed before tensions flared between the companies in September 2010 over the hiring of Hurd. We conclude that the second sentence, moreover, does more than declare an aspiration or intent to continue working together, as Oracle claims. It commits the parties to continue the actions specified (Oracle offering its product suite and HP supporting the products) as they had done before Oracle hired Hurd. Other provisions of the agreement reinforce the mutual intent expressed in the reaffirmation clause. Recital B states that “HP, Hurd and Oracle recognize the mutual advantages of the continuation of the HP-Oracle partnership and its benefits to their joint customers and prospects and now desire to further their business relationship and resolve the [l]awsuit [against Hurd].” The joint press release, provided for in the agreement, states that the settlement “reaffirms HP and Oracle’s commitment to delivering the best products and solutions to our more than 140,000 shared customers.” Each instance in the agreement that refers to the HP-Oracle partnership (recital B, paragraph 1, and the attached press release) is paired with a reference to supporting the joint customer base. Viewed together, “each clause helping to interpret the other” (§ 1641), the agreement addresses three subjects: purpose, conduct, and public assurance. The recital highlights the purpose for the reaffirmation clause, to continue the partnership and its benefits to the joint customers, the reaffirmation clause articulates the conduct agreed upon, to continue the partnership by offering and supporting Oracle’s 28 products on HP’s hardware as before the Hurd dispute, and the press release provides a public assurance of the joint “commitment to delivering the best products and solutions to our more than 140,000 shared customers.” Oracle’s arguments to the contrary are unconvincing. Oracle maintains that the agreement’s plain language, considering the text of the reaffirmation clause and the agreement as a whole (§ 1641), confirms that HP and Oracle agreed only to continue, not fundamentally alter, their prior, voluntary and mutually beneficial relationship after the Hurd dispute unfolded. Oracle contends that the trial court erred in finding the language to be consistent with “a continued porting obligation” because no “obligation” ever existed outside of a few contracts applicable to specific products, let alone for an indefinite period of time and without payment. Oracle submits it is undisputed that the parties’ relationship before the hiring of Hurd was based on “porting at will” without any payment or condition when mutually beneficial. It points to the negotiated porting agreements, executed a few times for specific software, as the only instances in which Oracle obligated itself to port a product. Thus, Oracle asserts that the relationship the parties agreed to continue was one in which each company had discretion to engage, and “[i]f their views did not align, Oracle would simply decline to port.” Oracle contends that reaffirming the arrangement as it already existed does not create new commitments or enlarge their scope. We disagree. The plain language of the agreement not only ties the notions of partnership and support for joint customers together but also describes, in the second sentence of the reaffirmation clause, the behavior to which each side commits for that purpose. Oracle relies on Barham v. Barham (1949) 33 Cal.2d 416 (Barham), for the proposition that a contract “[r]eaffirm[ing]” prior agreements “will preserve their legal effect” (id. at p. 425) but “will not operate to enlarge their scope” (id. at p. 426). But Barham involved the interpretation of a divorce settlement provision that “expressly

29 ratified and reaffirmed” several preceding agreements concerning property distribution between the estranged spouses. (Id. at p. 425.) The contractual provision here bears no resemblance to the one examined in Barham because the agreement here does not reaffirm a legal agreement with delineated duties, but instead references a noncontractual relationship defined by the parties’ past course of dealing. The language providing that Oracle “will continue to offer its product suite on HP platforms . . . in a manner consistent with that partnership” maintains—not enlarges—the parties’ scope of activities as conducted before the Hurd dispute. But unlike the express ratification of a prior contractual agreement that serves only to preserve the prior agreement’s legal effect (cf. Barham, supra, 33 Cal.2d at p. 425), this mandatory language, which appears in a contractual setting where previously no contract had existed, necessarily creates new legal duties. The parties in the agreement, therefore, have bound themselves by contract to “continue” certain conduct, the scope of which is defined by past, voluntary practices. Oracle next points to the agreement’s overarching purpose—resolution of Hurd’s employment dispute—and its employment of an integration clause as support for its proposed interpretation. Oracle questions how a 17-page agreement that is “overwhelmingly dedicated” to resolving the controversy over the terms of Hurd’s employment can be read to surrender “in just two sentences” each company’s control over its development obligations. Oracle argues that to read a “sweeping porting obligation” into the agreement is inconsistent with the integration clause stating that the agreement “constitutes the entire agreement among the Parties regarding the resolution and settlement of the [l]awsuit” (italics added) over Hurd, especially because specific

30 porting agreements were still operative between the parties and contained detailed provisions regarding duration, remedies, intellectual property rights, and payment.13 We perceive no contradiction between the commitments made in the affirmation clause and the settlement of the Hurd dispute. Oracle’s hiring of Hurd precipitated a tumultuous period in which the parties’ partnership—and particularly Oracle’s continued commitment to offering its product suite on Itanium—was in question. As Oracle states in its opening brief, quoting the trial court’s findings, “the reaffirmation clause was a reaction to Oracle’s public threat to end all business collaboration just days before, and arose from HP’s concern that ‘the litigation might permanently damage its relationship with Oracle.’ ” Recital B of the agreement, in which the companies expressly recognized “the mutual advantages of the continuation of the HP-Oracle partnership and its benefits to their joint customers” suggests that reaffirming the partnership was central to settling the dispute over Hurd’s hiring.14 By assuring Oracle’s continued offering of its product suite on HP’s platforms, the reaffirmation clause furthers the overarching goal of the entire agreement to continue the companies’ partnership for the benefit of their joint customers, particularly as expressed in the joint press release provision and statement. Finally, Oracle challenges the trial court’s determination that to construe paragraph 1 as having reserved for the companies the “absolute discretion not to work together” would be “essentially illusory.” Oracle claims it was error to characterize its interpretation as illusory when it both constrained Oracle from altogether ending the

13 For example, in 2006 HP contracted to pay Oracle up to $10.3 million to port an application software product, the E-Business Suite, to Itanium (EBS agreement). The EBS agreement is a 14-page commercial contract with terms defining the products covered by the agreement, Oracle’s porting and maintenance responsibilities, contract duration (set to expire December 31, 2013), the amount HP would pay, and limitation of liability provisions, among other standard commercial contract terms. 14 The other two recitals, Recital A and Recital C, address the procedural history of HP’s lawsuit against Hurd and specify the effect of the agreement on Hurd’s obligations under his prior separation agreement with HP. 31 relationship with HP and abandoning Itanium (which it submits it was free to do in September 2010) and prevented it from allowing the Hurd dispute to interfere with its relationship. Oracle relies on Third Story Music, Inc. v. Waits (1995) 41 Cal.App.4th 798 (Third Story) and other authorities to argue that illusoriness is determined from the entire contract, not a single provision, and that even entirely discretionary provisions in contracts are frequently upheld. Oracle further disputes that its interpretation renders the affirmation clause illusory, since it conferred on HP the “temporary benefit” of preventing Oracle from terminating the relationship. Oracle cites Asmus v. Pacific Bell (2000) 23 Cal.4th 1, 16 (Asmus), in support of the proposition that a short-term benefit is not illusory where the promisee obtains the benefit of the promise while it remains in force. Asmus concerned an employer’s implied power to terminate a unilateral contract in the employment context and does not help Oracle here, both because the contract in question does not confer the implied-in-fact unilateral power discussed in Asmus and because Oracle’s exercise of its purported discretion under the agreement lacked any of the procedural protections identified in that case. (See Asmus, supra, 23 Cal.4th at p. 18 [concluding, based on the application of contract principles, that “an employer may terminate a unilateral contract of indefinite duration, as long as its action occurs after a reasonable time, and is subject to prescribed or implied limitations, including reasonable notice and preservation of vested benefits”].) More to the point, in our view the issue is not whether the agreement was illusory, or even contained an illusory promise, but whether the reaffirmation clause may be construed to mean that Oracle reserved absolute discretion to decide whether to offer its next generation products on HP’s platform. In stark contrast with Third Story and other cases cited by Oracle, the contractual language does not expressly grant a discretionary power. (Cf. Third Story, supra, 41 Cal.App.4th at pp. 801–802 [marketing contract promised to market music, or to refrain from doing so, at the election of the promisor]; Carma Developers (Cal.), Inc. v. 32 Marathon Development California, Inc. (1992) 2 Cal.4th 342, 376 (Carma) [lease provision granting the lessor the right to terminate and recapture tenant’s lease “to claim for itself appreciated rental value of the premises was expressly permitted by the lease and was clearly within the parties’ reasonable expectations”].) The trial court in this case used the phrase “essentially illusory” to mean that interpreting the reaffirmation clause as “ ‘an agreement to continue to work together . . . that reserves the absolute discretion not to work together . . . gives [it] no real meaning.’ ” Applying de novo review to the agreement’s terms, we agree with the trial court’s interpretation. Oracle’s proposed construction of the reaffirmation clause as an obligation-free reaffirmation of an entirely voluntary partnership is contrary to the plain language of the second sentence and renders the provision superfluous, since the affirmatory public restatement of the partnership was addressed by paragraph 15, which required issuance of a joint press release. “An interpretation which gives effect is preferred to one which makes void. (§ 3541; see also Rest.2d Contracts, § 203, subd. (a) (1981) [interpretation which gives a reasonable meaning is preferred to one which renders a part of no effect].) Oracle’s interpretation runs contrary to the statutory preference for a construction of the contract which gives meaning to all the terms. We conclude that the phrase “ ‘in a manner consistent with that partnership as it existed’ ” is not reasonably susceptible to an interpretation that would transform the “ ‘will continue’ ” language in the first part of the sentence into language meaning “ ‘may continue’ ” or “ ‘has the discretion not to continue.’ ” Simply put, the language of the reaffirmation clause and the agreement as a whole do not support Oracle’s claim to absolute discretion over whether to continue offering its products on HP’s platform. To the contrary, as noted ante, in agreeing that it “will continue to offer its product suite on HP platforms . . . in a manner consistent with th[e] partnership as it existed prior to Oracle’s hiring of Hurd” (italics added), Oracle ceded its discretion over whether to offer those products by taking on a new legal obligation to do so. 33 Our analysis thus far has focused on ascertaining the parties’ mutual intent based solely on the agreement’s plain language. Though the words of the affirmation clause are clear and explicit and may be understood in an ordinary, non-technical way (§§ 1638, 1644), words alone “do not have absolute and constant referents.” (Pacific Gas, supra, 69 Cal.2d at p. 38.) In our view, more information about the “partnership as it existed prior to Oracle’s hiring of Hurd” is needed to understand what that arrangement entailed and the resulting expectations or limitations it imposed on each company. Paragraph 1 of the agreement thus exemplifies the need to appraise “ ‘circumstances surrounding the making of the agreement . . . including the object, nature and subject matter of the writing . . .’ so that the court can ‘place itself in the same situation in which the parties found themselves at the time of contracting.’ ” (Id. at p. 40.) To accomplish this task, we turn to the extrinsic evidence offered during the phase 1 trial.

b. Extrinsic Evidence Oracle maintains that the uncontroverted course-of-dealing evidence, set forth in the trial court’s findings of fact, confirms the voluntary nature of the partnership arrangement, whereby Oracle conducted “over 99% of all porting to HP-UX/Itanium” voluntarily and without contractual obligation. It asserts that even HP’s witnesses testified that each company had “ ‘discretion’ ” under the arrangement not to offer products, in Oracle’s case, and not to support products, in HP’s case. Oracle further contends that the parties’ negotiations over the agreement terms and HP’s conduct after executing the agreement confirm that the parties did not intend for the reaffirmation clause to impose any new obligations. It is well settled that courts may consider extrinsic evidence insofar as it sheds light on a meaning to which the contract is reasonably susceptible. (Iqbal, supra, 10 Cal.App.5th at p. 8.) “Extrinsic evidence is ‘admissible to interpret the instrument, but

34 not to give it a meaning to which it is not reasonably susceptible.’ ” (Parsons, supra, 62 Cal.2d at p. 865.) Accordingly, we review the extrinsic evidence to understand the objective intent behind the agreed-upon point of reference for the reaffirmation clause, namely the “partnership as it existed prior to Oracle’s hiring of Hurd.” Because the extrinsic evidence relevant to our analysis is largely uncontradicted, we apply independent review, even when conflicting inferences may be drawn from the evidence. (Wolf v. Walt Disney Pictures & Television (2008) 162 Cal.App.4th 1107, 1126 (Wolf).) To the limited extent that we refer to disputed evidence, we defer to the trial court’s determination of witness credibility if supported by substantial evidence in the record. (See Tin Tin, supra, 170 Cal.App.4th at p. 1225.) The record is replete with descriptive evidence of how the companies viewed and portrayed their strategic partnership. For example, marketing material from October 2009 touted the companies’ more than “25 years of collaborative partnership” and “140,000 joint customers” and highlighted their joint support of solutions and market share in each other’s products. Even after the acquisition of Sun, Oracle’s top software executives told their HP counterparts that the company was committed to releasing new versions of software on Itanium in parity with their release on competitor platforms. Joint meeting minutes from the February 2010 meeting (after the Sun acquisition) attended by executives on both sides listed “[c]ommitment to the partnership” as a “key decision[] reached” (capitalization omitted) and covered topics like release parity of Oracle’s products on Itanium. And meeting minutes from the “joint alliance executive review” in April 2010 reflect Oracle co-president Catz’s opening statement about the companies’ “[s]trong relationship from many years of working together” and Oracle’s“ commit[ment] to [the] HP relationship and running Oracle Software on HP Infrastructure.” As these examples illustrate, the parties’ own conception of their partnership in the timeframe prior to Oracle’s hiring of Hurd centered on serving their

35 joint customers by continuing to offer Oracle’s software on HP’s hardware in a way that was competitive with other platforms. Evidence specific to the parties’ course of dealing in the years and months before Hurd’s hiring further demonstrates that the porting of Oracle’s products to HP’s platforms, the companies’ joint sales and marketing of those products, and coordinated joint support constituted the defining features of the partnership. Most notably, the partnership history between HP and Oracle contained no instance of Oracle offering a product on an HP platform, carrying out the initial port, and then electing not to continue to offer the product by porting subsequent versions or releases to the platform. Oracle’s executive vice-president of software development confirmed that “once we made the ports available[,] subsequent versions were available on Itanium.” When customers expressed concern about a rumored change in Oracle’s commitment to Itanium after it acquired Sun, Oracle’s “updated drawer statement” for HP to share with prospective customers highlighted Oracle’s products that were available on HP’s Itanium servers and reiterated Oracle’s “target[]” to ship upcoming Oracle releases to Itanium in parity with “the other strategic UNIXes.” We conclude this evidence of Oracle’s unbroken practice of porting the newest versions and releases of its product suite to existing HP platforms supports the only plausible reading of the agreement’s commitment to furthering the companies’ business relationship (recital B) and “delivering the best products and solutions to [their] more than 140,000 shared customers” (appended press release). It would be objectively unreasonable for the parties to agree to further their business relationship and to publicly announce their commitment to delivering the best products to their customers while giving Oracle unbounded discretion to refuse to port to their shared customers the latest versions of those products, which as the trial court correctly noted “are perforce its ‘best products,’ ”—an authority that Oracle had not previously exercised with respect to HP’s server platforms. 36 The evidence surrounding formation of the agreement also does not support Oracle’s construction of the reaffirmation clause as having preserved its discretion to decline to port products that were already being offered on the platform. From the outset, both sides expressed a desire to preserve the strategic relationship. That relationship, as just discussed, was premised on a course of dealing in which Oracle made available, typically without contract or fee, and through mutual cooperation and joint efforts, the latest versions of its product suite on HP’s server platforms. When Livermore and Catz spoke on September 11, 2010, after HP initiated the Hurd lawsuit, Livermore conveyed HP’s desire that the settlement include “reaffirmation of the partnership” “as part of the contract.” Catz testified that she viewed reaffirming the partnership as “a very modest ask.” Second, the parties quickly agreed that the strategic relationship was not bounded by their preexisting written contracts. HP deleted Oracle’s initial, proposed partnership language that limited the reaffirmation commitment to the “longstanding strategic partnership as established under” the companies’ “existing contractual commitments and their mutual desire to continue to support their joint customers.” Oracle accepted the deletion of the existing contracts limitation and adopted HP’s modified version of what became the first sentence of paragraph 1. Third, the parties agreed to omit language that was preferential to HP or would have put HP in a better position than it had been prior to Oracle’s decision to hire Hurd. Oracle firmly rejected HP’s proposal that would have favored HP over other competitors, with Catz and Livermore agreeing that the purpose of reaffirmation was not to put HP in a better position than it currently enjoyed. In a draft proposal on September 12, 2010, Oracle retained HP’s proposed language that “Oracle will continue to offer its product suite on HP platforms and HP will continue to support Oracle products . . . on its hardware” but replaced the term that would have advantaged HP over other competitors with the phrase “in a manner consistent with that partnership.” Daley highlighted 37 Oracle’s intent on this point in her cover e-mail to HP, stating that the reaffirmation clause “was intended to reaffirm and continue the existing relationship and not to put HP in a better position tha[n] it currently enjoys or result in the negotiation of a new contractual commitment.” Fourth, the parties ultimately agreed to exclude specific commitments and instead used their “partnership as it existed prior to Oracle’s hiring of Hurd” as the point of reference for the strategic partnership. This point was closely tied to HP’s early efforts to include preferential language. Daley’s September 12 e-mail rejected this approach, stating that Catz and Livermore “did not discuss anything more tha[n] an agreement to continue to work together as the companies have – with Oracle porting products to HP’s platform and HP supporting the ported products and the parties engaging in joint marketing opportunities – for the mutual benefit of customers.” Livermore agreed the next day not to add more specific language into the agreement. Livermore testified that she was satisfied at that point that the draft provision addressed her concerns about Oracle continuing to offer its product suite on HP’s server platform. Livermore also understood from Catz that Oracle was committing to continue its past course of action and was not taking on new obligations. Consistent with this understanding, Livermore quickly stepped in when HP’s lawyers tried to add specific terms for Oracle to “continue to support all ongoing versions of HP-UX with Oracle’s relevant database, middleware and application products with the availability, marketing and pricing in competitive terms that Oracle has provided HP for the past five years.” Livermore told Catz that the lawyers’ language got past her and was a mistake. In the final exchanges surrounding the negotiation of the agreement, Oracle deleted language about supporting “all ongoing versions of HP-UX . . . in competitive terms that Oracle has provided HP for the past five years,” and HP modified the partnership reference to specify that the parties were reaffirming their relationship “as it existed prior to Oracle’s hiring of Mark Hurd.” Daley, Catz, and HP’s general counsel 38 Holston each testified that the parties intended to refer to the partnership and course of dealing before the controversy erupted over Oracle’s hiring of Hurd. Our observations here reinforce the plain meaning of the reaffirmation clause and the extrinsic evidence pertaining to the parties’ partnership. Oracle repeatedly refused to commit to granting preferential terms to HP in relation to the marketing or pricing of its products, but it did not protest the inclusion of an explicit commitment to continue offering its product suite on HP’s platform in the same manner it had immediately prior to the Hurd dispute. Oracle adopted the proposed language and framed it in terms of the preexisting partnership, which was undeniably premised—both in external messaging and internal planning—on the consistent and prospective availability of the latest versions of Oracle’s software on HP’s platforms. Oracle contends that this interpretation is inconsistent with its rejection, during negotiations, of language proposed by HP “that would have included a new duty to port,” as well as with the parties’ mutual understanding that Oracle “was not ‘taking on any new obligations or work.’ ” Oracle cites its rejection on September 12, 2010, of language that would have committed it “to continue to offer its product suite on HP Platforms on terms that are as good or better than any other platform,” and on September 13 of language that would have committed it to “continue to support all ongoing versions of HP-UX with Oracle’s relevant database, middleware and application products with the availability, marketing and pricing in competitive terms that Oracle has provided HP for the past five years.” Oracle suggests the common element of these spurned clauses was the porting obligation and claims it was implausible for the trial court to instead infer that the rejection pertained to the preferential terms. But these arguments ignore what the record plainly shows. Most significantly, Oracle rejected the proposed language on September 12 that related to “terms that are as good or better than any other platform” but accepted the obligation to “continue to offer its product suite on HP Platforms . . . .” Oracle in fact adopted that language in its 39 responsive proposal and added to it, inserting HP’s commitment to “continue to support Oracle products (including Oracle Enterprise Linux and Oracle VM) on its hardware in a manner consistent with that partnership.” Oracle’s further rejection of the language proposed by HP’s lawyers on September 13, which again contained preferential language regarding pricing and marketing, does not change its express acceptance and adoption of the “will continue to offer its product suite” language. In accepting that language, Oracle manifested an objective intent to “continue [] offer[ing]” (i.e., porting) its product suite to existing HP server platforms. The record does not support Oracle’s attempt to draw a contrary inference, particularly given Daley’s contemporaneous e-mail in which she disclaimed Catz and Livermore discussing “anything more tha[n] an agreement to continue to work together as the companies have – with Oracle porting products to HP’s platform and HP supporting the ported products and the parties engaging in joint marketing opportunities – for the mutual benefit of customers.” Daley’s testimony at trial that she “made a reference to porting and joint marketing as examples of the kinds of things that would not become obligatory as a result of that affirmation provision” cannot be reconciled with the text of the e-mail. We defer to the trial court’s credibility finding on this point, as articulated in the statement of decision, that “Daley’s testimony is the precise opposite of what [her] email actually stated and is entitled to no weight.” Nor does the evidence of Oracle’s assertion that it was taking on no new obligations help its case. There is no question that Catz emphasized, and Livermore understood, that Oracle would not commit to “any new obligations or work.” Oracle contends that this manifestation of intent confirms the parties did not agree to a new porting obligation. We agree there were no “new” obligations in the reaffirmation clause because the agreement was for each party to continue to do what it had done in the past. In Oracle’s case, it agreed that it “will continue to offer its product suite on HP

40 platforms . . . in a manner consistent with that partnership as it existed prior to Oracle’s hiring of Hurd.” We do not construe the invocation of no “new obligations or work” as tantamount to disavowing any legal commitments altogether. Taken to its logical conclusion, Oracle’s insistence on no new obligations would have precluded it from entering into a contract, which by definition entails a “new” legal obligation. By agreeing to continue to offer its product suite on HP’s platforms in a manner consistent with the partnership as it existed prior to Oracle’s hiring of Hurd, Oracle did exactly what the parties at the outset had agreed upon, which was to “reaffirm their commitment to their longstanding strategic relationship and their mutual desire to continue to support their mutual customers.” In Oracle’s case, the uncontested evidence is that its past practice was to continue porting new versions of its product suite to the HP servers that HP continued to sell to its customers. Oracle complains that in construing the agreement to require Oracle to “port all its software, indefinitely and without charge,” the trial court contravened the parties’ mutual intent not to put HP “in a better position” than it had enjoyed prior to the agreement. Yet the trial court did not construe the agreement to require Oracle to port all its software but to continue to offer the software, including new versions, releases, or updates, that made up the product suite that was already offered for Itanium at the time the parties signed the agreement. The trial court did not construe Oracle’s obligation to apply without charge indefinitely but “until such time as HP discontinues the sale of its Itanium-based servers.” Each of these obligations reflected the continuation of past practices, consistent with the objective intent of the agreement. Oracle contends that HP’s conduct in the months after signing the agreement, before the present controversy arose, demonstrates the parties’ understanding that the reaffirmation clause did not impose upon Oracle a duty to port. Oracle cites, for example, evidence that HP never told its executives responsible for the Oracle-HP 41 relationship that Oracle was required to port. Oracle also suggests that HP did not act in conformity with the understanding of the agreement it professes, because in December 2010 it offered to support Oracle Enterprise Linux and Oracle VM products—both which were already expressly listed in paragraph 1 of the agreement—if Oracle agreed to pay $5 million and guarantee continued porting to Itanium. Oracle advances a form of “practical construction” placed upon a contract by the parties before a controversy has arisen as to its meaning, which under appropriate circumstances may be entitled to great weight by the court. (Crestview Cemetery Ass’n v. Dieden (1960) 54 Cal.2d 744, 753–754; see Universal Sales Corp. v. Cal. etc. Mfg. Co. (1942) 20 Cal.2d 751, 761–762.) HP responds that this approach “ ‘may be considered only when the acts of the parties were positive and deliberate and done in attempted compliance with the terms of the agreement.’ ” (United States Liab. Ins. Co. v. Haidinger-Hayes, Inc. (1968) 263 Cal.App.2d 531, 538 (Haidinger-Hayes).) The Restatement, which “ ‘California usually follows’ ” (Airs Aromatics, LLC v. CBL Data Recovery Technologies Inc. (2020) 50 Cal.App.5th 1009, 1014), indeed suggests that such evidence of acceptance or acquiescence in a course of performance requires “repeated occasions for performance by either party with knowledge of the nature of the performance and opportunity for objection to it by the other.” (Rest. 2d, Contracts § 202(4).) An exemplary application of the doctrine may be found in Oceanside 84, Ltd. v. Fidelity Federal Bank (1997) 56 Cal.App.4th 1441, 1450–1451, where the court determined that a borrower’s payments on a loan for five years, without objection, in the face of periodic rate adjustments may be considered acquiescence in the lender’s interpretation of the method for calculating the interest rate adjustments. Here, however, none of the conduct cited by Oracle meets the specifications of the interpretational doctrine of practical construction because, unlike in Oceanside 84, there was no meaningful and sustained opportunity prior to the onset of this litigation for the parties to engage in or test their performance under the agreement. For example, while it 42 is true that the key individuals at HP responsible for the Oracle relationship were informed only that a settlement had been reached and they should carry on with “business as usual,” that messaging was wholly consistent with the agreement’s terms (including its confidentiality provision), because the HP employees understood their obligation was to keep doing what they had been doing before the agreement. Regarding HP’s offer to support Oracle’s Enterprise Linux and VM products, Oracle omits to mention that the evidence adduced at trial on this subject, which included testimony from several fact witnesses, showed that HP’s offer was in reference to providing “enhanced OEL/OVM support and marketing” (italics added) after actions taken by Oracle on pricing appeared to disadvantage HP’s competitiveness. Oracle also appears to misconstrue HP’s communications as an attempt to “negotiate a porting agreement for Itanium” (which would be unnecessary had it believed it already had one). But the evidence Oracle relies on shows only a discussion about Oracle’s commitment to “continued parity for HP-UX” in timing its software releases and does not seek a commitment to porting generally. HP’s conduct in this context may not be fairly construed as evidence of a deliberate effort to comply with the terms of the agreement. (Cf. Haidinger-Hayes, supra, 263 Cal.App.2d at p. 538.)

c. Definite and Enforceable Terms In addition to these arguments grounded in extrinsic evidence, Oracle also challenges the agreement as lacking sufficiently definite and enforceable terms. It contends that the reaffirmation clause omits material terms including the scope of the duty involved, performance limits, duration, and compensation, as well as other terms necessarily included in prior porting agreements between the parties. HP responds that the parties used their course of dealing to supply the material terms for the agreed-upon porting and product support obligations. We agree that the parties’ past practices afford an adequate basis to ascertain and enforce the obligations under the agreement.

43 The object of the agreement, defined as “the thing which it is agreed, on the part of the party receiving the consideration, to do or not to do” (§ 1595) is, in this case, entirely ascertainable. Oracle agreed to continue to offer and update its product suite—those software products already available on Itanium when the agreement was signed—on HP’s existing platform. Even so, Oracle complains that the agreement fails to identify products or platforms, does not discuss porting fees or address project management, maintenance, or support, and includes no warranties, disclaimers, damages limitations, or provisions on intellectual property ownership. Oracle contrasts the sparseness of the reaffirmation clause against more detailed provisions of the agreement, like the multi- paragraph “standstill agreement” pertaining to possible takeover activity, or the duration- specific provisions limiting Hurd’s activities at Oracle. Oracle maintains, citing Bustamante v. Intuit, Inc. (2006) 141 Cal.App.4th 199, 215 and Vons Companies, Inc. v. United States Fire Ins. Co. (2000) 78 Cal.App.4th 52, 58 (Vons), that the trial court improperly supplied terms the contract lacked and did so in a manner inconsistent with the express terms of the parties’ written porting agreements and with the terms that would have been specified had two, sophisticated parties sought to negotiate an “unconditional, comprehensive porting agreement.” Oracle’s argument rests on a faulty premise. The reaffirmation clause is not an unconditional, comprehensive porting agreement; nor do we believe the trial court’s construction of the agreement made it one. Simply put, the reaffirmation clause creates an obligation to continue with an expressly identified course of dealing (offering and supporting Oracle’s products on HP’s existing platforms as long as they are sold by HP) no different from the course of dealing that had defined their strategic partnership for years prior to Oracle’s hiring of Hurd. Our independent review of the plain language and of the relevant extrinsic evidence, including the past course of dealing and circumstances surrounding formation of the agreement, confirms the trial court’s interpretation that

44 paragraph 1 of the agreement requires Oracle to continue to offer its product suite on the HP Itanium platform, which HP continued to sell. To be sure, the trial court articulated terms in its order granting HP declaratory relief. But it did so by inferring from the parties’ past practices and longstanding course of dealing on each of those topics. This was entirely acceptable. “Unexpressed provisions of a contract may be inferred from the writing or external facts.” (Cal. Lettuce Growers v. Union Sugar Co. (1955) 45 Cal.2d 474, 482 (Cal. Lettuce).) While courts may not “create for the parties a contract that they did not make” (Vons, supra, 78 Cal.App.4th at p. 59), courts may look to the nature and circumstances of the contract to effectuate the intent of the parties where it can be reasonably ascertained. “ ‘The law does not favor, but leans against the destruction of contracts because of uncertainty; and it will, if feasible, so construe agreements as to carry into effect the reasonable intentions of the parties if that can be ascertained.’ ” (Cal. Lettuce, at p. 481.) California case law provides numerous examples of the use of extrinsic evidence to infer or clarify a contractual term. (See, e.g., Cal. Lettuce, supra, 45 Cal.2d at pp. 484–485 [inferring price setting features and obligation to purchase from prior dealings of the parties]; Consolidated Theatres, Inc. v. Theatrical Stage Employees Union (1968) 69 Cal.2d 713, 729 [inferring term of duration from “the nature of the contract and the circumstances surrounding it”]; Okun v. Morton (1988) 203 Cal.App.3d 805, 818 [rejecting uncertainty of business ventures contract where extrinsic evidence introduced at trial was “sufficient to establish . . . the ways in which future ventures were to be financed, owned, and operated by the parties”].) In this case, extrinsic evidence of the strategic partnership provided detailed information about the parties’ course of dealing as it related to Oracle’s porting practices to HP’s existing server platforms, summarized ante (part II.A.2.b.). Oracle does not question that the products it offered for Itanium when the agreement was signed properly defined the product suite. It does not dispute that over 99 percent of porting at the time 45 had occurred or was occurring without a written contract or exchange of payment, and that once it made a product available on Itanium by porting, it continued to release subsequent versions of the software on the platform as long as HP offered the platform. These practices comprised the course of dealing for many years and formed the basis for the strategic partnership. Because the reaffirmation clause identified the “partnership as it existed prior to Oracle’s hiring of Hurd” as the reference point for Oracle’s commitment to continue offering its product suite on HP platforms, the trial court did not err in construing the agreement to require the contractual partnership to continue—and end—on those same terms.

d. Summary For these reasons, we conclude that the agreement is not reasonably susceptible to Oracle’s proposed interpretation. To construe the reaffirmation clause as affirming only the voluntary or discretionary basis of the HP-Oracle partnership is inconsistent with both the plain language of the agreement and with the record of the parties’ past course of dealing, which we view as the defining feature of their strategic partnership. While there may not have been a contractual porting obligation prior to the signing of the agreement, there was an established and uninterrupted porting practice in which Oracle partnered with HP to make each subsequent release of software available on Itanium, without payment or limitation. Oracle’s stated desire to enter into a contract without taking on any new obligations does not supersede the mutually expressed intent of the parties, as evidenced by the words of the contract and confirmed by robust extrinsic evidence, to continue their practice of offering and supporting Oracle’s product suite on HP’s existing platform as long as HP continued to sell the platform, as that practice existed before the Hurd dispute arose. Accordingly, the trial court did not err in its construction of paragraph 1 of the agreement as set forth in its statement of decision following the phase 1 trial.

46 B. Breach of Contract and Breach of Implied Covenant Claims Oracle attacks the jury verdict for breach of contract and breach of the implied covenant of good faith and fair dealing. With respect to breach of contract, Oracle contends that the undisputed evidence at the phase 2 trial established that there was never a breach of the agreement because Oracle resumed porting to Itanium in time to complete the next significant port of a product version around the same time frame it made that version available on competitor platforms. Oracle maintains that HP’s actual claim, based on the March 2011 announcement to end porting to Itanium, was for anticipatory repudiation. Further, Oracle asserts that HP waived any right to damages for anticipatory breach by seeking, obtaining, and accepting performance. As for the verdict on breach of the implied covenant, Oracle contends that because the reaffirmation clause had no express porting provision and was not a “porting agreement,” the implied covenant could not impose a substantive duty to port and could not be breached by Oracle’s (later retracted) decision to cease porting. HP responds that Oracle failed to assert the anticipatory repudiation argument in the trial court and has forfeited its claim that HP waived the right to contract damages. HP otherwise contends that substantial evidence supports the jury’s conclusion that Oracle’s March 2011 announcement and the ensuing conduct breached the express terms of the agreement and the implied covenant of good faith and fair dealing.

1. Breach of Contract We begin with Oracle’s challenges to the jury verdict for breach of contract. We address and reject HP’s contention that Oracle has forfeited this argument on appeal. We next apply the law governing contract repudiation and breach to uncontroverted facts in the record to decide as a matter of law whether HP waived any claim to damages based on Oracle’s March 2011 announcement.

47 a. Forfeiture HP frames Oracle’s anticipatory repudiation claim as an issue newly asserted on appeal. HP contends that nowhere in motion practice, in limine motions, or in any other filings in the trial court, including in its proposed jury instructions, did Oracle argue that HP could not pursue contract damages because the purported basis for Oracle’s liability was anticipatory repudiation of the agreement. Oracle responds that HP’s forfeiture argument is “perplexing” given that Oracle attempted at numerous points in the phase 2 trial to limit HP’s contract claim and available damages based on the fact that HP received and accepted performance under the contract. “ ‘As a general rule, theories not raised in the trial court cannot be asserted for the first time on appeal; appealing parties must adhere to the theory (or theories) on which their cases were tried. This rule is based on fairness—it would be unfair, both to the trial court and the opposing litigants, to permit a change of theory on appeal.’ (Eisenberg et al., Cal. Practice Guide: Civil Appeals and Writs (The Rutter Group 2015) ¶ 8:229; p. 8– 167.) ‘New theories of defense, just like new theories of liability, may not be asserted for the first time on appeal.’ ” (Nellie Gail Ranch Owners Assn. v. McMullin (2016) 4 Cal.App.5th 982, 997.) Having reviewed the record, we agree with Oracle that it raised the relevant issues of breach of contract and anticipatory repudiation numerous times before and during the phase 2 trial. For example, Oracle argued in briefing filed before the phase 2 trial that the trial court should adopt Oracle’s proposed special instructions on damages, in part because HP’s failure to plead “any theory of breach by anticipatory repudiation . . . is fatal to any effort by HP to claim lost profits damages based on the statement contained in Oracle’s March 2011 announcement.” Oracle asserted that if HP were permitted to pursue damages based on the theory that the March 2011 announcement constituted an anticipatory repudiation, it could not claim lost profits from a period before performance was due, because when the repudiating party retracts its repudiation before performance

48 is due, “ ‘the repudiation is nullified and the injured party is left with his remedies, if any, invocable at the time of performance.’ ” (Citations omitted.) Oracle argued that since both parties acknowledge the agreement is still in force, any belated attempt by HP to claim damages based on an anticipatory repudiation theory “cannot succeed.” Oracle also sought leave to move for summary adjudication of HP’s cause of action for breach of contract before the phase 2 trial. Oracle argued that because it fulfilled its stated intent after the phase 1 trial to deliver its software for Itanium on the same schedule as other platforms, the court should determine prior to trial whether the “real-world contractual performance of the [agreement], as interpreted by this Court, voids HP’s breach of contract and promissory estoppel claims.” Oracle asserted in the proposed summary adjudication motion that HP’s only possible theory of breach was based on Oracle’s March 2011 announcement that it would not discharge its future obligation (as defined by the trial court in phase 1); however, that theory of anticipatory breach was incompatible with HP’s subsequent demand for and acceptance of Oracle’s performance under the contract. The trial court denied Oracle leave to file the proposed motion for summary adjudication. These instances demonstrate Oracle’s attempts before and during the phase 2 trial to challenge HP’s contract claim on the grounds that (1) Oracle fully performed under the contract, precluding liability for breach of contract, and (2) HP had waived any claim based on anticipatory repudiation by demanding and accepting performance. Oracle’s showing more than satisfies the minimum standard for preservation of claims on appeal, which “ ‘is that the asserted error must have been brought to the attention of the trial court.’ ” (DiPirro v. Bondo Corp. (2007) 153 Cal.App.4th 150, 178.) Both the trial court and HP were aware of Oracle’s grounds for challenging HP’s breach of contract claim by the phase 2 trial. We decide that Oracle has not forfeited its right to assert on appeal its anticipatory repudiation argument; therefore we address its merits.

49 b. Breach of Contract and Anticipatory Repudiation Oracle’s contention that HP’s breach of contract claim is actually a claim for anticipatory repudiation, for which HP has waived its right to damages, rests on two premises—first, Oracle’s interpretation of the parties’ obligations under the agreement; second, application of the laws governing breach of contract and anticipatory repudiation to certain undisputed statements made by Oracle. We begin by summarizing the governing rules and principles. As the facts relevant to ascertaining the nature of HP’s breach claim are not in dispute, we generally exercise independent review. (Ghirardo v. Antonioli (1994) 8 Cal.4th 791, 799.) To the extent there are factual disputes concerning any of the relevant historical facts, we defer to the decision of the trial court where supported by substantial evidence. (Id. at p. 800.) Oracle relies on principles of contracts, anticipatory repudiation, and double recovery. “California law recognizes that a contract may be breached by nonperformance, by repudiation, or a combination of the two.” (Central Valley General Hospital v. Smith (2008) 162 Cal.App.4th 501, 514, fn. omitted; see Rest.2d Contracts, § 236 (1981).) “Any breach, total or partial, that causes a measurable injury, gives the injured party a right to damages as compensation.” (Witkin, Summary of Cal. Law (11th ed. 2020) Contracts, § 877.) Nonperformance typically refers to an unjustified or unexcused failure to perform a material contractual obligation when performance is due. (Central Valley, at p. 514, fn. 3.) But “[t]here can be no actual breach of a contract until the time specified therein for performance has arrived.” (Taylor v. Johnston (1975) 15 Cal.3d 130, 137 (Taylor).) By contrast, “an anticipatory breach of contract occurs when the contract is repudiated by the promisor before the promisor’s performance under the contract is due.” (Central Valley, at p. 514, citing Taylor, at p. 137.) In other words, “if a party to a contract expressly or by implication repudiates the contract before the time for his or her performance has arrived, an anticipatory breach is said to have occurred.” (Romano v. Rockwell Internat., Inc. (1996) 14 Cal.4th 479, 489 (Romano).)

50 Oracle maintains that HP’s breach of contract claim is, in fact, a claim for anticipatory breach involving an express repudiation. An express repudiation “is a clear, positive, unequivocal refusal to perform.” (Taylor, supra, 15 Cal.3d at p. 137.) The California Supreme Court’s 1975 opinion in Taylor illustrates the principles of repudiation and anticipatory breach. The plaintiff in Taylor contracted to breed his two thoroughbred mares to the defendants’ stallion. (Taylor, supra, 15 Cal.3d at p. 133.) The defendants sold the stallion and shipped it to buyers in Kentucky before the time for performance under the breeding contracts was due. The plaintiff refused to accept the defendants’ written “ ‘release []’ ” from his “ ‘reservations’ ” for the stallion and insisted on performance. (Ibid.) The plaintiff shipped the mares to Kentucky where the defendants arranged for the breeding to go forward. (Id. at pp. 133–134.) But after numerous failed attempts to secure a date for the breeding, the plaintiff abandoned the effort and bred his mares to a different stallion for a substantially higher stud fee. (Id. at pp. 134–135.) In the breach of contract action that followed, the trial court awarded the plaintiff damages after finding the defendants liable for breach. (Taylor, supra, 15 Cal.3d at p. 135.) The California Supreme Court in Taylor reversed. The court noted that while the trial court found the defendants’ continuous course of conduct from the sale of the stallion through the plaintiffs’ last attempted effort to schedule a breeding “amounted to a repudiation which [the] plaintiff was justified in treating as an anticipatory breach,” the conduct “cannot be treated as an undifferentiated continuum amounting to a single repudiation but must be divided into two separate repudiations.” (Id. at p. 138.) The court explained that where the plaintiff did not treat the defendants’ repudiation (selling the stud to buyers in Kentucky) as an anticipatory breach and instead acted on the defendants’ retraction of the repudiation (by arranging for breeding in Kentucky), the “retraction nullified the repudiation.” (Ibid.)

51 The high court disagreed with the trial court’s conclusion that the defendants’ repeated, subsequent cancellations of the plaintiffs’ attempts to reserve the stallion constituted a second repudiation. (Taylor, supra, 15 Cal.3d at pp. 138–139.) The court observed that there was no evidence the defendants had expressly refused to perform, nor did their conduct amount to an unequivocal refusal to perform, since at the time the plaintiffs abandoned their efforts there still remained time in the breeding season to attempt performance. (Id. at p. 139.) The court reasoned that while the defendants’ conduct may have “cast doubt upon the eventual accomplishment of performance[,] it did not render performance impossible.” (Ibid.) The court concluded that “as a matter of law this conduct did not amount to an unequivocal refusal to perform and therefore did not constitute an anticipatory breach of the contract.” (Id. at p. 141.) In reaching its conclusion that there had been no anticipatory breach of the contract, the California Supreme Court in Taylor explained the legal consequences of an anticipatory repudiation. “When a promisor repudiates a contract, the injured party faces an election of remedies: he can treat the repudiation as an anticipatory breach and immediately seek damages for breach of contract, thereby terminating the contractual relation between the parties, or he can treat the repudiation as an empty threat, wait until the time for performance arrives and exercise his remedies for actual breach if a breach does in fact occur at such time. [Citation.] However, if the injured party disregards the repudiation and treats the contract as still in force, and the repudiation is retracted prior to the time of performance, then the repudiation is nullified and the injured party is left with his remedies, if any, invocable at the time of performance.” (Id. at pp. 137–138.) The court, in a later case, described the principle of anticipatory breach as a recognition that “the promisor has engaged not only to perform under the contract, but also not to repudiate his or her promise.” (Romano, supra, 14 Cal.4th at p. 489.) Oracle contends that its March 2011 announcement that it was discontinuing software development for Itanium was not a breach of the agreement because the time for 52 performance had not yet arrived. Oracle maintains that the March 2011 announcement equally cannot serve as the basis of a claim for anticipatory breach for which HP can recover damages because HP did not consider the announcement as a repudiation but instead treated the contract as in force and accepted Oracle’s eventual performance. Oracle argues that for HP to treat the repudiation as a breach and seek damages while also demanding continuing performance conflicts with the established rule whereby the injured party, upon repudiation, “faces an election of remedies.” (Taylor, supra, 15 Cal.3d at p. 137.) Oracle compares HP’s posture to that of the plaintiff in Taylor after the breeding arrangements in Kentucky effectively nullified the retraction. Oracle claims that because HP pressed for performance of the agreement and accepted and relied upon Oracle’s September 2012 announcement retracting any repudiation, it was limited to its “remedies that might arise at the time of performance” (Taylor, supra, 15 Cal.3d at p. 138) if Oracle failed to perform. Oracle further contends that as the non-repudiating party suing for breach of contract damages, HP had to reject the breaching party’s tender of performance or else waive its right to pursue damages for anticipatory breach. Oracle’s attempt to characterize HP’s claim as a waived claim for anticipatory repudiation rests on two assumptions about the agreement. First, Oracle assumes that the reaffirmation clause, interpreted by the trial court to require “Oracle to continue to offer its product suite on HP’s Itanium-based server platforms” sets the time of performance as Oracle’s release of the next versions of software in its product suite available on Itanium. In Oracle’s words, “[t]he time for performance had not yet arrived, and would not arrive, until Oracle actually released the software versions that it purportedly had agreed to port to Itanium.” Second, Oracle assumes that the agreement defines the mandatory conduct strictly in terms of the obligation to port, ignoring the broader context and purpose to publicly reaffirm the continuation of the strategic partnership in service of the joint customers. 53 These assumptions, if correct, would support the application of an anticipatory repudiation framework to Oracle’s March 2011 announcement, as anticipatory repudiation exists to address a repudiation before the promisor’s performance under the contract is due. (Romano, supra, 14 Cal.4th at p. 489.) However, the agreement’s text, as we have construed it above, does not support these assumptions. The agreement obligates Oracle to “continue to offer its product suite on HP platforms, . . . in a manner consistent with th[e] [companies’] partnership” before the Hurd dispute arose. (Italics added.) As we have explained above, the timeframe for performance was immediate because the agreement provided for the continuation of conduct that was ongoing at the time the parties signed the agreement. The evidence at trial, which described the nature and function of the partnership before the Hurd dispute, was for all intents and purposes uncontroverted. This evidence established that the process for Oracle to “offer its product suite” on the Itanium platform required ongoing coordination and collaboration around porting that did not have a definitive endpoint since software requires new releases, updates, and constant tuning to run on the platform. Performance here is defined as continuing conduct that was already established between the parties and was in full swing when Oracle made its March 2011 announcement. It is difficult to imagine circumstances in which a company’s public announcement that it was discontinuing all software development for a specified platform would not constitute actual breach of an agreement in which the company had agreed to “continue to offer” its software products on that platform consistent with the above- detailed prior practice. Furthermore, Oracle’s claim that the March 2011 announcement did not constitute a breach of contract ignores a fundamental feature of the agreement, which was its public reaffirmation of the HP-Oracle partnership. Here, as HP points out, the announcement by its very nature undermined customer confidence in the HP-Oracle relationship. Immediately after the March 2011 announcement, Oracle posted on its customer website 54 a list of the “next software release[s]” that would not be available for Itanium. In the months that followed, Oracle ceased all ongoing work to port new versions of its software to the Itanium server platforms. HP introduced evidence that Oracle capitalized directly on the opportunity to further weaken HP’s position by mobilizing its sales force to try to convert Itanium customers who depended on Oracle’s software to Oracle’s Sun platforms. HP also introduced evidence that customer uncertainty about the future availability of Oracle’s database software (in particular) on Itanium provoked many customers to switch their business to other platforms, and that by the time Oracle released its September 2012 statement recommitting to Itanium, many customers “had made alternate plans.” HP asserts that its Itanium business suffered immediate and irreversible harm as a result of the March 2011 announcement and subsequent actions taken by Oracle. Under these circumstances, Oracle’s March 2011 announcement did more than assert “a clear, positive, unequivocal refusal to perform” (Taylor, supra, 15 Cal.3d at p. 137) as in an express repudiation. The announcement precipitated nearly 18 months of conduct (until Oracle’s subsequent September 2012 statement in which it reversed course) during which Oracle arguably did not perform what under the terms of the agreement was a continuing contractual obligation to offer its software on the Itanium platform, and during which time customers were revaluating their positions and choices regarding next-generation database and other software applications. Although Oracle’s course reversal in September 2012, after the phase 1 ruling, allowed it to complete the port of Oracle’s database application and other products, substantial evidence in the record supports HP’s position that the damage from Oracle’s initial decision to discontinue porting and the conduct that attended that decision had an “immediate and devastating” impact. This evidence provided a sufficient basis for the trial court to allow the breach of contract claim to go to the jury.

55 We conclude that because the time for performance began when the parties signed the agreement, and the nature of the performance was both continuous and public, the March 2011 announcement did not constitute an anticipatory repudiation but was substantial evidence from which a reasonable jury could find Oracle had committed an actual breach of the agreement.

2. Breach of Implied Covenant of Good Faith and Fair Dealing Oracle similarly challenges the jury verdict on breach of the implied covenant of good faith and fair dealing. Oracle does not appear to challenge the sufficiency of the evidence; rather, it attacks the legal basis for the judgment in HP’s favor, claiming that there was no porting contract from which a covenant of good faith and fair dealing could be implied and that any purported breach arose only from Oracle’s anticipatory repudiation. Oracle also contends that the trial court’s erroneous interpretation of the reaffirmation clause “necessarily infected” the jury’s consideration of whether Oracle acted in good faith and consistently with the purposes of the agreement. The California Supreme Court has articulated the relevant framework. “The covenant of good faith and fair dealing, implied by law in every contract, exists merely to prevent one contracting party from unfairly frustrating the other party’s right to receive the benefits of the agreement actually made.” (Guz v. Bechtel Nat. Inc. (2000) 24 Cal.4th 317, 349, italics omitted.) The implied covenant “finds particular application in situations where one party is invested with a discretionary power affecting the rights of another. Such power must be exercised in good faith.” (Carma, supra, 2 Cal.4th at pp. 371–372.) The implied covenant cannot, however, “impose substantive duties or limits on the contracting parties beyond those incorporated in the specific terms of their agreement.” (Guz, at pp. 349–350.) In other words, “the scope of conduct prohibited by the covenant of good faith is circumscribed by the purposes and express terms of the contract.” (Carma, at p. 373.) It “will only be recognized to further the contract’s

56 purpose; it will not be read into a contract to prohibit a party from doing that which is expressly permitted by the agreement itself.” (Wolf, supra, 162 Cal.App.4th at p. 1120.) We perceive no error in the trial court’s application of the implied covenant of good faith and fair dealing. We have addressed and rejected Oracle’s contention that the agreement imposes no duty to port. The reaffirmation clause requires Oracle to “continue to offer its product suite on HP platforms . . . in a manner consistent with th[e] partnership as it existed prior to Oracle’s hiring of Hurd.” As discussed in detail above, the extrinsic evidence overwhelmingly confirms that to “offer its product suite” is synonymous with porting and impossible without porting. Oracle is therefore incorrect to assert that there was no porting obligation from which a covenant of good faith and fair dealing could be implied. This is not a case in which a court has read a contract’s implied terms to vary or impermissibly expand upon the express terms (see Carma, supra, 2 Cal.4th at p. 374), because the express terms affirmatively identify the conduct required by the contract. Oracle’s decision to cease porting its products to Itanium directly contradicted the contractual term that it would continue to offer those products on HP’s platform in a manner consistent with the parties’ partnership before the Hurd dispute. This arrangement stands in contrast with cases like Wolf, where the contract expressly granted “unfettered discretion” to a party (Wolf, supra, 162 Cal.App.4th at p. 1121), rendering any attempt to limit that discretion by use of an implied covenant improper as a matter of law (id. at pp. 1120–1121). Here, we have little difficulty concluding that Oracle’s decision to stop porting activities, and its subsequent conduct, could properly serve as the basis for the jury to consider breach of the implied covenant as “contrary to the contract’s purposes and the parties’ legitimate expectations.” (Carma, at p. 373.) Oracle also contends that the trial court’s “erroneous interpretation of the reaffirmation clause” that Oracle was required to port and had no discretion to do otherwise “necessarily infected the jury’s consideration of whether Oracle’s conduct was 57 in good faith and consistent with the purposes of the agreement.” Oracle submits that the verdict on the breach of the implied covenant of good faith and fair dealing claim must be reversed because HP’s theory that Oracle “lacked subjective good faith in the validity of its act” (Wolf, supra, 162 Cal.App.4th at p. 1123) was, in effect, predetermined by the trial court’s erroneous evidentiary rulings and jury instructions. Oracle focuses on the jury instruction that conveyed the trial court’s findings regarding the meaning of the agreement and on evidentiary rulings during Catz’s phase 2 trial testimony which prevented her from explaining her reasons for rejecting HP’s proposed terms on porting. It is true that over Oracle’s objection, the jury instruction for HP’s breach of contract cause of action repeated the trial court’s findings in the phase 1 statement of decision. As noted ante, the jury was instructed that it was the court’s duty to interpret the meaning of the agreement at issue in the case, and that at the conclusion of phase 1, the court determined the agreement “is a binding contract between HP and Oracle.” The jury was instructed that it “must accept as true” the court’s findings regarding the meaning of the agreement, including that the agreement “requires Oracle to continue to offer its product suite on HP’s Itanium-based server platforms and does not confer on Oracle the discretion to decide whether to do so or not.” Oracle contends that this instruction prevented the jury from even considering Oracle’s subjective good faith, especially because Oracle says the trial court’s evidentiary rulings during the phase 2 trial excluded evidence that showed why, even if mistaken, Oracle believed it was not required to continue porting. Oracle points out that the trial court sustained certain objections and did not allow Oracle to elicit testimony from Catz regarding her decision to reject the porting term that HP had proposed during the agreement negotiations, nor did it allow the jury to view the redlined draft agreement striking out HP’s proposed terms even though that exhibit had been admitted in phase 1. Oracle contends that having instructed the jury that Oracle was required to port its software to Itanium and having prevented the jury from hearing the already admitted 58 evidence that supported Oracle’s contrary view of its obligations, the jury’s consideration of the implied covenant claim was “indelibly tainted.” Oracle has not on appeal challenged the trial court’s evidentiary rulings or jury instructions; therefore, we do not examine those decisions for error. We note in relation to Catz’s testimony that although the trial court sustained an objection to a question asking why she struck language from a draft proposal during negotiations and did not allow the redlined draft to be published to the jury, it otherwise allowed Catz to testify in detail about her decision to reject HP’s proposed language regarding Itanium and development commitments. What is more, Oracle calls our attention in a perfunctory manner, in what amounts to a few paragraphs of summary argument, to issues that were fully litigated before and during the phase 2 trial. For example, the extent of the binding effect of the phase 1 findings at the phase 2 trial was the subject of extensive briefing and argument to the trial court. We recognize that in challenging the implied covenant verdict, Oracle seeks to highlight more broadly what it sees as the cumulative effect of the erroneous phase 1 rulings and interpretation of the agreement imposing an affirmative porting obligation on Oracle. Yet having determined that the trial court’s interpretation of Oracle’s obligations under the contract was not erroneous, and that the agreement did not authorize Oracle to discontinue porting to Itanium but expressly required that it continue offering its product suite on HP’s Itanium platform in a manner consistent with the partnership before Oracle’s hiring of Hurd, we see no basis, as a matter of law, for reversal of the verdict on this ground. C. Damages Oracle requests that we reverse the jury’s damages award of $3.014 billion for HP’s lost profits. Oracle’s two claims center on the testimony of HP’s damages expert, economist Jonathan Orszag. First, Oracle claims that Orszag’s calculation of damages was predicated in part on Oracle’s announced intent to appeal the trial court’s phase 1 59 ruling, effectively penalizing Oracle for exercising its right of appeal under the Petition Clause of the First Amendment of the United States Constitution, and infringing on Oracle’s litigation privilege (§ 47, subd. (b)) and the fair-and-truthful reporting privilege (§ 47, subd. (d)(1)). Second, relying primarily on Sargon, supra, 55 Cal.4th 747, Oracle claims that the trial court should have excluded Orszag from testifying because his testimony about HP’s lost profit damages was impermissibly speculative.

1. Standard of Review We apply de novo review to Oracle’s claim that the jury’s damage award erroneously included lost profits based on Oracle’s constitutionally protected and statutorily privileged statement that it would appeal. (See In re Taylor (2015) 60 Cal.4th 1019, 1035.) With respect to Oracle’s challenge to the trial court’s admission of Orszag’s expert testimony, “[e]xcept to the extent the trial court bases its ruling on a conclusion of law (which we review de novo), we review its ruling excluding or admitting expert testimony for abuse of discretion.” (Sargon, supra, 55 Cal.4th at p. 773.) “A ruling that constitutes an abuse of discretion has been described as one that is ‘so irrational or arbitrary that no reasonable person could agree with it.’ . . . [¶]. . . ‘The scope of discretion always resides in the particular law being applied, i.e., in the “legal principles governing the subject of [the] action . . . .” Action that transgresses the confines of the applicable principles of law is outside the scope of discretion and we call such action an “abuse” of discretion. . . . [¶] The legal principles that govern the subject of discretionary action vary greatly with context. . . . They are derived from the common law or statutes under which discretion is conferred.’ ” (Ibid., citations omitted.)

2. Additional Background

a. HP Expert Orszag’s Testimony Prior to the phase 2 jury trial in 2016, HP’s damages expert Orszag produced three written reports, two in 2012 and one in 2015. His original report was finalized in March

60 2012 and calculated HP’s estimated damages from Oracle’s breach of contract to be between $3.8 billion and $4 billion. As described above, in its September 2012 statement Oracle announced a reversal of its previous decision and stated it would resume porting to Itanium. Oracle wrote in a letter to the trial court that its September 2012 statement was “without prejudice to [its] rights to appeal” the rulings relevant to phase 1. In an earlier press release from August of that year, Oracle had also referenced its right to appeal when it criticized the trial court’s “preliminary opinion” and stated that “[w]e plan to appeal the Court’s ruling.” In order to address the impact of Oracle’s decision to resume porting to HP products, the trial court allowed the parties to serve supplemental expert reports and engage in additional discovery. In December 2012, Orszag produced a supplemental written report on damages. Shortly thereafter, the trial court conducted an evidentiary hearing in March 2013 on the admissibility of expert testimony, including that of Orszag, under the standards set out in Sargon, supra, 55 Cal.4th 747. Orszag testified at the March 2013 evidentiary hearing. Orszag stated that, in formulating his opinions, he relied on HP documents and projections, industry data, Oracle documents, industry analyst reports, and press releases or announcements, among other materials. In his original March 2012 report, Orszag divided his analysis into two time periods: (1) a “pretrial” period from March 2011 until “effectively today” and (2) from “today through 2020” and analyzed what HP’s revenues for the Itanium business would have been but for Oracle’s breach of contract. For the pretrial period, Orszag stated that he used three alternative approaches to estimate damages. The first approach was the “Itanium constant market share” approach, in which he assumed that Itanium’s market share in 2010 remained constant through 2020. His second approach used multiple versions of a regression (which he defined as a “statistical or econometric analysis of the relationship between variables”) to project 61 HP’s damages. His third approach relied on an internal HP forecast of the Itanium business, referred to as the “kinetic plan,” which HP had completed shortly before the Oracle announcement in March 2011. For the posttrial period, of the three methodologies he had used for the pretrial period, Orszag employed only those based on constant market share and kinetic plan. Orszag gave several reasons for his decision to extend future damages to 2020, including his use of data from a “roadmap” in place between HP and Intel (the supplier of the Itanium microprocessor) that went through 2020. In projecting future damages, he testified that he took into account factors other than Oracle’s March 2011 announcement, including market and sales trends. Orszag noted that the high-end server market in which Itanium competed against IBM’s and Sun’s proprietary servers (the RISC/EPIC market) had performed worse than projected in March 2011, which caused him to adjust down his projection of HP’s damages. He also considered in his calculations that some of the sales HP had lost were recaptured by other parts of HP’s business, including HP’s x86 server. Regarding the monetary damages that he calculated using the three methodologies, Orszag stated that in his original March 2012 report he had calculated damages to be between $3.8 billion and $4 billion depending on the methodology. Based on his experience as an economist, the numbers resulting from the different approaches were “robust” and showed a “pretty tight range for an estimate of damages.” In his December 2012 supplemental report, Orszag updated his prior opinion from March 2012 to account for new data from industry analysts and Oracle’s announcement that it “would reverse their previous decision and now port their Database software to the Itanium product.” Orszag did not assume that HP’s business would have grown exponentially but for Oracle’s conduct; rather he assumed that HP’s Itanium revenue “would have shrunk.” Orszag incorporated more recent industry data that revealed a “slightly more pessimistic view than [analysts] previously had” and which caused him to reduce in the December report his calculation of the amount of estimated damages. 62 Orszag discussed the impact of Oracle’s September 2012 announcement that it would resume porting to Itanium. Based on his review of the data, HP’s Itanium business continued to deteriorate following Oracle’s September 2012 announcement. Orszag did not assume that any Itanium customer would be unable to get the Oracle product it wanted. However, he concluded that HP had still suffered damages because of the gap between March 2011 and the September 2012 announcement. He relied in part on industry analysts’ observations that “[t]he damage had been done.” Orszag noted there had been a drop in sales because of the change in the mix of information to customers who “want to have reliability and assurance that the products they need for years are going to be there, and it wasn’t there.” On cross-examination, he stated, “I am not assuming there are customers who lost software today” but rather “[t]here are customers who believed that they were going to lose software tomorrow, that affected their decisions today because they are buying servers for a multi-year use period.” On cross-examination, Orszag acknowledged that one of the factors he took into account in his December 2012 supplemental report was Oracle’s August 2012 press release stating Oracle’s intent to appeal and its filing of a petition for writ of mandate. Orszag’s supplemental report highlighted the fact that “Oracle will appeal and there is no guarantee of the outcome.” Following Orszag’s testimony at the pretrial hearing on the admissibility of expert testimony, Oracle’s damages expert Ramsey Shehadeh also testified. He criticized Orszag’s reasoning as “circular.” For example, addressing Orszag’s constant market share methodology, Shehadeh stated that Orszag “assumes that HP would maintain a constant market share of what he describes as the RISC EPIC market” and that “his regression analysis evaluates the statistical correlation relationship between HP sales and that same RISC EPIC market.” According to Shehadeh, Orszag’s regression analysis “is constructed to yield the exact same assumption about the performance of the HP business that his constant market share approach does.” However, Shehadeh testified that in his 63 own analysis he, like Orszag, had relied on a constant market approach and industry analyst data to make certain of his calculations. Shortly after the evidentiary hearing, on March 20, 2013, the trial court issued a written order finding that Orszag’s expert testimony was admissible and rejecting Oracle’s argument that it failed to meet the Sargon standard. The trial court found that “[u]nlike the expert’s ‘employer’ in Sargon, HP is the veritable definition of an ‘established’ business, and Orszag was entitled to look at company records and statistics” and that, furthermore, “the analysis he undertook appears to be based on that data.” The trial court found that “Orszag’s testimony provides a logical basis for his conclusions, meets the standards proscribed in Sargon and the [Code of Civil Procedure], and will be admitted into evidence.” In its analysis of Orszag’s testimony, the trial court did not reference Orszag’s consideration of Oracle’s statement that it intended to appeal in calculating HP’s damages. However, the trial court did address Oracle’s intent to appeal in the context of its discussion of Oracle’s expert Shehadeh. The trial court stated, “whether the ‘we intend to appeal’ statement has a probative effect or not is for the jury to decide. It is but another prong in HP’s damages argument which may, or may not be persuasive at trial.”

b. Oracle’s Motion in Limine Regarding its Constitutional Right to Petition On April 5, 2013, Oracle filed a motion in limine seeking to exclude argument and evidence of lost profits that failed to disaggregate lost profits arising from Oracle’s plan to appeal (Oracle’s motion in limine). Oracle argued that HP was precluded from asserting such damages based on Oracle’s constitutional right to petition under the United States and California Constitutions, California’s litigation privilege, and California’s substantive law governing breach of contract and promissory estoppel. HP opposed Oracle’s motion in limine, arguing Orszag’s testimony should be presented to the jury. Oracle’s motion in limine remained pending while the parties litigated Oracle’s anti-

64 SLAPP motion. The trial court dismissed Oracle’s anti-SLAPP motion as untimely, and this court affirmed that order in 2015. (Hewlett-Packard, supra, 239 Cal.App.4th at p. 1196.) (See ante, part I.B.5.) In November 2015, following the delay in the litigation due to Oracle’s anti- SLAPP appeal, Orszag prepared his third and final report. The purpose of his final report was to update his December 2012 supplemental report with “actual data of what had happened over a three-year period.” He calculated a new damages estimate of $3.014 billion. As he testified at trial, Orszag reduced his calculation of damages from his previous estimate of approximately $4 billion because the RISC/EPIC market had shrunk much faster than analysts had forecast. On May 23, 2016, prior to the phase 2 trial, the trial court addressed the parties’ pending motions in limine, including Oracle’s motion to exclude arguments and evidence of lost profits that failed to disaggregate potential damages arising from Oracle’s exercise of its appellate rights. At the hearing, relying on its constitutional rights to freedom of speech and petition and the litigation privilege, Oracle argued that the jury could not base any damages on Oracle’s announcement that it would appeal the trial court’s phase 1 ruling. After hearing argument from counsel, the trial court denied Oracle’s pretrial motion.

c. Expert Testimony on Damages and Jury Award Orszag testified at the phase 2 jury trial on June 20, 2016 and June 21, 2016. On direct examination, he stated that he had based his damages opinion on Oracle’s decision in March 2011 to discontinue porting, its public announcement of that decision, and “the reaffirmation of that announcement over the next 17 and half months or so.” Orszag testified that under the constant market approach—his “preferred approach for damages”—he calculated that HP’s damages for the period of March 22, 2011, until

65 October 31, 2015, were $1.699 billion.15 Regarding future damages from November 1, 2015 to October 31, 2020, Orszag estimated the damages at roughly $1.3 billion. On cross-examination, Orszag confirmed that he was “not measuring damages based on a failure to ultimately deliver software.” Rather, in Orszag’s view, HP had suffered damages because Oracle had created uncertainty in the marketplace, and the risk that Oracle would win on appeal was part of the conduct that created the marketplace uncertainty. According to Orszag, “There was uncertainty in the marketplace about the fact that they were porting their software under protest, and that’s reflected in the real world decisions of businesses buying mission-critical hardware products.” When asked if, assuming Oracle had a legal right to appeal, Orszag’s current damage estimates could tell how much damages were caused by the uncertainty of the appeal, Orszag responded that that was “not an analysis that [he had] undertaken.” Orszag stated that he had not examined the damages caused by Oracle separate from its right to appeal and could not “give you an answer one way or the other.” The jury in the phase 2 trial ultimately awarded HP $1.699 billion in damages for “[p]ast lost profits” and $1.315 billion in “[f]uture lost profits” for a total damages award of $3.014 billion. In November 2016, Oracle filed a motion for new trial requesting a new trial on damages, or, in the alternative, a reduction of the damages to the maximum amount Oracle’s expert (Shehadeh) had testified was supported by the evidence ($559 million). Oracle asserted that the jury award was both excessive and contrary to law. It argued that Orszag’s statistical models were flawed in that they did not include “even one other causative factor that would have played a significant role in the declining market share” for Itanium but rather concluded that Oracle’s “announcement caused 100% of this decline in market share.” Oracle further argued that Orszag “did not attempt to

15 Orszag segregated past from future damages using October 31, 2015, as the final date for which he was able to include past data when preparing his third and final report in November 2015. 66 separate the impact to Itanium of other factors that by law cannot form the basis of any damages award, such as Oracle’s statements and actions that it would appeal the Phase 1 decision” and that Orszag “admitted that, if Oracle’s assertion of its right to appeal was legally proper, then he had no opinion on the amount of damages in this case.” The trial court denied Oracle’s motion for new trial. Among its other findings, the trial court addressed Oracle’s argument “that the litigation privileges immunize[] it from any liability arising from its statement in the September 2012 announcement that it would appeal the Phase [1] decision in this action.” The trial court found that “HP did not claim that this statement caused its damages, merely that this and other circumstances created uncertainty surrounding Oracle’s commitment to Itanium, explaining why the September 2012 announcement did not cause HP’s market share to recover.”

3. Constitutional Claim We begin with Oracle’s claim that the trial court erred as a matter of law in allowing Orszag to base his calculation of damages in part on Oracle’s constitutionally protected and privileged statement that it would appeal the trial court’s phase 1 decision. Oracle broadly asserts that its statement of intent to appeal was protected by the First Amendment right to petition the courts and the California litigation and fair-and-truthful reporting privileges. Oracle submits that Orszag both attributed damages to the intent to appeal and conceded he could not disaggregate those damages from damages caused by other factors. “[I]t is a fundamental principle of appellate procedure that a trial court judgment is ordinarily presumed to be correct and the burden is on an appellant to demonstrate, on the basis of the record presented to the appellate court, that the trial court committed an error that justifies reversal of the judgment.” (Jameson v. Desta (2018) 5 Cal.5th 594, 608– 609.) This principle of appellate practice is founded in the constitutional doctrine of reversible error. (Ibid.; see Cal. Const., art. VI, § 13.) It precludes reversal of the

67 judgment on the ground of the improper admission of evidence unless, after examining the entirety of the matter before us, we conclude that “the error complained of has resulted in a miscarriage of justice” (Cal. Const., art. VI, § 13) which, in this case, may be conceived as damages that penalize Oracle for exercising a constitutional right. We recite these basic principles in observance of several, unaddressed gaps in Oracle’s arguments for reversal of the jury’s damages verdict on constitutional and privilege grounds. As explained further below, we decide that Oracle has failed to establish both that there was legal error in admission of this testimony and that the jury’s damages award was actually based on an impermissible consideration of Oracle’s protected or privileged conduct. As a threshold matter, to demonstrate that the jury impermissibly assigned damages based on constitutionally protected conduct, Oracle must first establish that its conduct was protected. Oracle relies on the Petition Clause in the First Amendment to the United States Constitution and cites a single line from a single case, Borough of Duryea v. Guarnieri (2011) 564 U.S. 379 (Guarnieri), to support the premise that its September 2012 announcement was an exercise of its constitutionally protected right to petition. Guarnieri, however, bears little relationship to the dispute at issue here. Guarnieri addressed “the extent of the protection, if any, that the Petition Clause grants public employees in routine disputes with government employers.” (Guarnieri, supra, 564 U.S. at p. 382.) In Guarnieri, the United States Supreme Court described the right of petition as “the right of individuals to appeal to courts and other forums established by the government for resolution of legal disputes.” (Id. at p. 387.) This general statement does not establish whether invoking the intent to appeal in a press release in a private, contractual, commercial dispute implicates the Petition Clause. (See Guarnieri, at pp. 388–389 [“A petition conveys the special concerns of its author to the government and, in its usual form, requests action by the government to address those concerns.”].) 68 Other leading cases from the United States Supreme Court on the Petition Clause are similarly far afield. (See e.g., Bill Johnson’s Restaurants, Inc. v. N.L.R.B. (1983) 461 U.S. 731, 733 [considering whether the National Labor Relations Board may issue a cease-and-desist order to halt the prosecution of a state court civil suit brought by an employer to retaliate against employees for exercising federally protected labor rights]; California Motor Transport Co. v. Trucking Unlimited (1972) 404 U.S. 508, 511 [concluding that the right of access to the courts is an aspect of the First Amendment right to petition the Government for redress of grievances and construing the antitrust laws as not prohibiting the filing of a lawsuit, regardless of the plaintiff’s anticompetitive intent or purpose in doing so, unless the suit was a “ ‘mere sham’ ” filed for harassment purposes].) In short, Oracle’s conclusory reference to the right of petition, accompanied by neither argument nor application to the facts presented, is insufficient to establish that its press release raised a constitutionally protected right. But even if we assume that Oracle’s statement of intent to appeal comes within the ambit of the Petition Clause of the First Amendment, Oracle offers no authority connecting the exercise of the right of petition in the context of a breach of contract claim with what it presents as “black letter law” that the First Amendment prohibits courts from awarding damages that would penalize protected speech. The cases upon which Oracle relies bear no resemblance to the circumstances here. For example, Oracle cites Snyder v. Phelps (2011) 562 U.S. 443, in which the United States Supreme Court considered whether the First Amendment shields church members from tort liability for picketing near a soldier’s funeral service. The court examined the nature of the speech, whether of public or private concern, as determined by the circumstances of the case (id. at pp. 450–451), and concluded that the First Amendment barred the plaintiff’s recovery on tort claims like intentional infliction of emotional distress (id. at p. 459) and intrusion upon seclusion (id. at p. 460). The Court 69 reasoned that the church members’ speech was “at a public place on a matter of public concern” and therefore “entitled to ‘special protection’ under the First Amendment.” (Id. at p. 458.) Oracle does not demonstrate how its press release falls within this rubric. Oracle cites another case along this vein of protected speech, Freitag v. Ayers (9th Cir. 2006) 468 F.3d 528, 532, in which a jury found three prison administrators liable for retaliating against a former correctional officer for engaging in constitutionally protected speech. After determining that a jury instruction erroneously listed examples of unprotected speech as well as protected speech, the Ninth Circuit remanded to the district court to decide whether the error was harmless (i.e. whether the jury verdict finding retaliation was affected by the erroneous inclusion of the two or three examples of unprotected speech) (id. at p. 546) and whether the compensatory damages award remained valid (id. at p. 547). In its decision, the Ninth Circuit relied on a then-recent decision from the United States Supreme Court, which held that the First Amendment protects speech by public employees only when “the speech in question addresses a matter of public concern” and is not made pursuant to their official duties. (Id. at pp. 543–544 [analyzing Garcetti v. Ceballos (2006) 547 U.S. 410].) Oracle does not attempt to explain the relevance of these cases to the issues presented here, beyond the unremarkable generalization that the First Amendment prohibits penalizing protected speech. The reliance on this overarching principle does not establish constitutional error in the damages award—particularly because Snyder and Freitag involve the protection of speech, whereas here Oracle claims a violation of its right to petition. Further, there was no dispute in Snyder and Freitag that the speech had directly resulted in significant penalties (in the case of Snyder a substantial tort judgment and in the case of Freitag the termination of employment), whereas here the link between the asserted right to petition and the money judgment for breach of contract and the implied covenant of good faith and fair dealing is—at best—substantially more attenuated. 70 The only other case in support of its constitutional claim that Oracle refers to is an unreported opinion from the Central District of California. That case examines the contours of “[s]ubstantial truth” as a defense to defamation and trade libel/commercial disparagement claims under California and Illinois law. (See Aurora World, Inc. v. Ty Inc. (C.D. Cal., Aug. 24, 2010) 2010 WL 11506546, at *13.) We fail to see its relevance to this appeal. In sum, in failing to develop a reasoned argument supported by legal authority for its First Amendment claim, Oracle improperly leaves this court to decode what amounts to little more than “ ‘a bare assertion that the judgment, or part of it, is erroneous’ ” . . . .’ (Eisenberg et al., Cal. Practice Guide: Civil Appeals and Writs (The Rutter Group 2018) ¶ 8:17.1, p. 8-6).” (Lee v. Kim (2019) 41 Cal.App.5th 705, 721.) For this reason alone, Oracle’s First Amendment argument cannot prevail. Nonetheless, recognizing the possibility that Oracle’s stated intent to appeal may fall within its First Amendment right of petition (Guarnieri, supra, 564 U.S. at p. 387) and the likelihood that damages arising from the protected conduct are prohibited (see, e.g., NAACP v. Claiborne Hardware Co. (1982) 458 U.S. 886, 926–927, 933 [reversing judgment where state court imposed liability on organizers of a boycott for business losses resulting in part from nonviolent, protected speech and assembly]), we examine Oracle’s contention that the jury was permitted to award damages based on Oracle’s exercise of its constitutional right to petition the courts. Oracle claims that Orszag’s approximately $3 billion damages figure was based on “two Oracle statements”: the March 2011 announcement that it would discontinue future product development on Itanium and the September 2012 statement that it would comply with the trial court’s order, though it was appealing the court’s decision. Oracle further states that Orszag attributed all of HP’s damages to these announcements and admitted that he could not allocate damages between them. To assess the accuracy of Oracle’s claim, we consider Orszag’s testimony. 71 There is no dispute that Orszag’s testimony about how he calculated HP’s damages included consideration of Oracle’s stated intention to appeal the phase 1 ruling. Orszag told the jury that HP had suffered damages because Oracle had created “uncertainty” in the marketplace, and the risk that Oracle would win an appeal was part of the conduct that created the marketplace uncertainty despite Oracle’s announcement that it would resume porting to HP’s products. As summarized by Orszag to the jury on direct examination, “When the decision [after the phase 1 trial] had been reached they said that they were disappointed in the decision and that they intended to appeal that decision. Then some time after that, within a few weeks, . . . they announced that they would resume porting. So, they would resume the future development of—future software development of their Oracle Database for Itanium. And that they would do so, effectively under protest, because they were still appealing the decision.” Orszag confirmed that he took those announcements into consideration in calculating damages, explaining that in the “actual world” of data he reviewed after the September 2012 announcement, customers were “making buying decisions with the knowledge that Oracle is currently porting its database software, A. And B, that they are appealing the decision. So, I’m taking that into account. And those buying decisions are the buying decisions that we observed in HP’s data.” Oracle argues, based on this testimony and similar statements, that the claimed uncertainty was the sole basis for HP’s claim that Itanium would continue to lose market share even after Oracle resumed porting “and is the entire basis for Orszag’s damages from that point in time forward.” But this assertion substantially oversimplifies Orszag’s testimony. For example, Oracle’s counsel pressed Orszag in cross-examination to identify whether the damages he calculated after Oracle’s September 2012 announcement “are 100 percent based upon the effect of the announcement and zero percent based upon the actual availability or nonavailability of the software? [¶] That was your position; 72 correct?” Orszag responded, “And it’s the same position I have had prior to that date and the same position I have today.” In other words, Orszag testified that the September 2012 announcement that Oracle would resume porting its software to Itanium did not alter Orszag’s opinion regarding the market effect of the original, March 2011 announcement. Orszag clarified his statement by explaining that “there was an 18-month period when they stopped software development, and then they restarted software development, and then they delivered software to customers. But during that 18-month period, they had been telling the market that they had discontinued software development, and that had a significant adverse effect on the marketplace, as I measure in my damage calculation.” We observe that Orszag’s testimony on this point was consistent throughout his direct and cross-examination. Orszag identified the same sources of damages at the outset of the jury trial when he explained that he had based his damages opinion on “the decision not—to discontinue porting, the March 2011 announcement of that decision, and the reaffirmation of that announcement over the next 17 and a half months or so.” Orszag testified that based on the market data he had collected, the September 2012 announcement that Oracle would resume porting did not cause HP’s market to “bounce back” because, for the preceding 18 months, the market had operated with the knowledge that Oracle’s next generation of database and other applications would not be available on Itanium. Orszag opined that “there was uncertainty in the marketplace about Oracle’s commitment to the future development of its database for the Itanium products.” Orszag’s testimony thus falls far short of attributing damages to the stated intent to appeal, rather than to market uncertainty about the availability of Oracle’s software on HP’s Itanium platform after Oracle’s March 2011 announcement. The following colloquy illustrates Orszag’s response at trial to a question clarifying the relationship between Oracle’s stated intent to appeal and the market’s uncertainty leading to damages: “[Counsel for Oracle]: You’re saying that’s the way the world was. The way the world 73 was, was Oracle had the right to appeal, and that was part of the uncertainty; correct? [¶] [Orszag]: I think that’s a fair proposition. There was uncertainty in the marketplace about the fact that they were porting their software under protest, and that’s reflected in the real world decisions of businesses buying mission-critical hardware products. [¶] [Counsel for Oracle]: And that uncertainty that you’ve described in your report is one part of the cause of damages in this case; correct? [¶] [Orszag]: Again, it’s — I'm looking at what actually happens. Part of the uncertainty would tend to reduce sales. The fact that they decided to port their software would tend to increase sales, and I looked at the world as it actually was from 2012 to 2015.” We conclude from the testimony at trial that Orszag explained HP’s damages in terms of the real-world data that showed what consumers were buying, or not buying, starting in March 2011, through late 2012 following Oracle’s September 2012 announcement, and through the time of trial in 2016. Orszag acknowledged that Oracle’s vow to appeal failed to reverse the market’s uncertainty about the future of Oracle’s product suite on Itanium, even after the September 2012 announcement that Oracle would resume porting. At no point, however, did Orszag ascribe damages to Oracle’s statement of its intent to appeal standing alone or suggest the uncertainty reflected in the market data was solely, or even predominantly, the result of Oracle’s announcement about its appeal. More accurately, Oracle’s appeal was a factor in the calculation of damages only insofar as it reduced any mitigation from the resumption of porting in September 2012, not because the exercise of the right to appeal was a source of harm in and of itself. Orszag’s testimony cannot reasonably be interpreted as having based his calculation of HP’s damages on Oracle’s statement of intent to appeal. Our observations are reinforced by the trial court’s ruling on Oracle’s motion for new trial. In addressing whether Oracle was immune from liability arising from its 2012 statement regarding its plan to appeal the phase 1 decision, the trial court found that “HP did not claim that this statement caused its damages, merely that this and other 74 circumstances created uncertainty surrounding Oracle’s commitment to Itanium, explaining why the September 2012 announcement did not cause HP’s market share to recover.” Though the trial court framed this aspect of Oracle’s new trial motion as largely a legal issue, it rendered its finding fully informed by its evaluation of the evidence at trial. To the extent that the basis for HP’s damages claim represents a factual issue about the nature of the testimony heard by the jury, we defer to the trial court’s finding. (Ghirardo, supra, 8 Cal.4th at p. 800.) We also note that although Oracle proposed special jury instructions to address the “constitutional prohibition of damages based on” Oracle’s petitioning activity, Oracle has not claimed instructional error in the instructions given, none of which directed the jury to consider Oracle’s September 2012 announcement in ascertaining damages. The jury was instructed to decide reasonable compensation “for the harm caused by the breach” (CACI No. 350) and defined breach, as claimed by HP, only in terms of Oracle (1) making its March 2011 decision and announcement, and (2) repeatedly telling customers that it would no longer offer its product suite on Itanium (CACI No. 300). This differentiates the circumstances here from a case like Freitag, supra, 468 F.3d at p. 532, where the jury instruction on retaliation expressly identified instances of unprotected speech as well as protected speech. We recognize the circumstances here present an apparently novel issue, but Oracle’s arguments lack both legal and factual support. Oracle has cited no case authority that addresses the implications of a party’s exercise under the Petition Clause of the First Amendment, where that constitutional exercise affects the market behavior underlying an expert’s calculation of damages. Nor have we identified a case directly on point—or even one that is analogous. As a general proposition, Oracle may be correct when it argues that “[n]o one could seriously contend that a plaintiff could recover damages because the defendant noticed an appeal.” (Cf. Bordenkircher v. Hayes (1978) 434 U.S. 357, 363 [“To punish a person because he has done what the law plainly allows 75 him to do is a due process violation of the most basic sort.”].) But Oracle has not persuaded us that issuance of a press release announcing an intent to appeal an interim legal finding on a contractual claim in a business dispute implicates the Petition Clause. In addition, based on our review of the record, Oracle has not shown that HP recovered damages based on Oracle’s stated intention to appeal the phase 1 ruling. That the jury in the phase 2 trial awarded HP the same amount of damages set forth in Orszag’s preferred damages model does not demonstrate that Oracle was penalized for exercising its constitutional right of petition, because Orszag did not testify that he considered Oracle’s intent to appeal as a “cause” of damages but as a factor in the affected market’s failure to “bounce back” after the resumption of porting. For these reasons, we conclude that Oracle has not sustained its burden to demonstrate error requiring reversal of the jury’s damages verdict on First Amendment grounds. Lastly, although Oracle references the California litigation and fair-and-truthful reporting privileges (§ 47, subds. (b), (d)(1)) as an additional source of immunity from liability for its exercise of its right of petition, it fails to support its claim with adequate argument or authority. Oracle relies on a single line from the California Supreme Court’s decision in Silberg v. Anderson (1990) 50 Cal.3d 205 (Silberg), which explained the broad application of the litigation privilege. The quoted sentence states that the privilege applies to any publication, or communication, “required or permitted by law in the course of a judicial proceeding to achieve the objects of the litigation, even though the publication is made outside the courtroom and no function of the court or its officers is involved.” (Id. at p. 212.) Oracle’s conclusory assertion that the litigation privilege applies here is insufficient. Oracle does not attempt to apply the formulation typically used to determine whether the privilege applies under the circumstances in a given case. (See Silberg, supra, 50 Cal.3d at p. 212 [articulating the “usual” four-part formulation]; accord Rusheen v. Cohen (2006) 37 Cal.4th 1048, 1057.) The sheer breadth of the privilege is 76 not enough to establish its application to Oracle’s statement of intent to appeal. We observe that Oracle wholly fails to address that it remains unsettled under California law whether or to what extent the privilege precludes liability for contract claims.16 The broad formulation of the privilege set forth in Silberg, and in more recent cases like Rusheen, discuss the privilege only in terms of immunity to tort claims. (Silberg, at p. 212 [stating “the privilege is now held applicable to . . . all torts except malicious prosecution”]; Rusheen, at p. 1057 [“ ‘[C]ommunications with “some relation” to judicial proceedings’ are ‘absolutely immune from tort liability’ by the litigation privilege”].) Although Oracle points to the outcome in Action Apartment Assn., Inc. v. City of Santa Monica (2007) 41 Cal.4th 1232 (Action Apartment) as additional authority for reversing, as a matter of law, a damages award purportedly based on protected litigation conduct, that case does not address whether the litigation privilege serves as a bar to contract damages.17 Meanwhile, Oracle cites no authority and provides no legal analysis to support its argument that the fair-and-truthful reporting privilege applies to its September 2012 announcement. Because Oracle has failed to provide reasoned argument or legal analysis in support of its assertion that Oracle’s stated intent to appeal was absolutely privileged under the California litigation and fair-and-truthful reporting privileges, we are unable to ascertain any reversible error as to this issue.

16 Indeed, the California Supreme Court is currently considering whether the litigation privilege of section 47, subdivision (b), applies to contract claims, and if so, under what circumstances. (See Doe v. Olson (Aug. 30, 2019, B286105) [nonpub. opn.], review granted Nov. 20, 2019, S258498 [2019 WL 4127263].) 17 In Action Apartment, the California Supreme Court held that the litigation privilege partially preempted a city ordinance authorizing civil and criminal penalties against a landlord for maliciously serving a notice of eviction, and fully preempted a provision of the ordinance authorizing penalties against a landlord for bringing an action to recover possession of a rental unit without a reasonable factual or legal basis. (Action Apartment, supra, 41 Cal.4th at p. 1237.) 77 4. Sargon Claim Oracle asserts, relying on Sargon, supra, 55 Cal.4th 747, that Orszag’s testimony was impermissibly speculative and should have been excluded from the jury’s consideration in the phase 2 trial. Oracle argues primarily that Orszag’s approach to damages was fundamentally flawed because he attributed a single causal factor— Oracle’s conduct—to HP’s actual and projected decline in its share of the relevant market. Oracle argues that Orszag failed to consider other key events in the complex market for servers, including Intel’s announcement of a new microprocessor (i.e., the Xeon E7 microprocessor for x86 servers), and other events related to HP’s overall business. Courts have “a substantial ‘gatekeeping’ responsibility” to exclude unreliable expert testimony. (Sargon, supra, 55 Cal.4th at p. 769.) The California Supreme Court explained in Sargon that the trial court’s gatekeeping responsibility is required by Evidence Code sections 801, subdivision (b), and 802, and that the trial court must exclude expert opinion testimony “that is (1) based on matter of a type on which an expert may not reasonably rely, (2) based on reasons unsupported by the material on which the expert relies, or (3) speculative.” (Sargon, at pp. 771–772.) The trial court’s gatekeeping role as to expert testimony, including as to lost profits, is to determine “whether the expert opinion is founded on sound logic” rather than to assess its “persuasiveness.” (Id. at p. 772.) In Sargon, the California Supreme Court explained that “[l]ost profits need not be proven with mathematical precision, but they must also not be unduly speculative” and concluded that the trial court had acted within its discretion to exclude “opinion testimony that [Sargon] would have become extraordinarily successful had the university completed the clinical testing.” (Sargon, supra, 55 Cal.4th at p. 753.) Although it affirmed the trial court’s exclusion of the expert in that case, the California Supreme Court in Sargon also provided the following caution: “The lost profit inquiry is always

78 speculative to some degree. Inevitably, there will always be an element of uncertainty. Courts must not be too quick to exclude expert evidence as speculative merely because the expert cannot say with absolute certainty what the profits would have been. Courts must not eviscerate the possibility of recovering lost profits by too broadly defining what is too speculative. A reasonable certainty only is required, not absolute certainty.” (Id. at p. 775.) A “trial court’s gatekeeping role does not involve choosing between competing expert opinions.” (Id. at p. 772.) Rather, a trial court “must simply determine whether the matter relied on can provide a reasonable basis for the opinion or whether that opinion is based on a leap of logic or conjecture.” (Ibid.) We have carefully considered Oracle’s arguments with respect to Orszag’s methodology and assumptions and decide that Oracle has not shown error in the trial court’s admission of Orszag’s testimony. Oracle’s arguments criticizing Orszag’s methodology are factors the jury could properly consider, but they do not mandate exclusion of the evidence altogether. For example, Oracle argues—as it did in the trial court— that Orszag failed to isolate the impact of Intel’s announcement about the Xeon E7 chip, which occurred approximately three weeks after the Oracle announcement in March 2011. Oracle portrays Intel’s announcement as the “death knell” of the Itanium platform. Oracle contends that Orszag’s analysis was “fatally speculative” because he did not measure the effect of the introduction of the Xeon E7 on Itanium’s market share. But at trial, Orszag testified that he had considered the impact of similar product announcements on Itanium in past years and noted that “when we look at competing products, every time they came out Itanium did fine.” In his view, when a competitor came out with a faster chip, there was not a significant impact because “[i]n this market there’s a lot of . . . installed base.” Regarding the Xeon E7 chip in particular, Orszag characterized it as only “an incremental improvement over the previous year’s chip” and opined that its introduction would not affect HP’s market share because “there’s no

79 evidence that this product would [a]ffect one company in the RISC/EPIC marketplace differently than the others.” The record thus does not support Oracle’s assertion that Orszag’s approach, like the expert testimony deemed inadmissible in Sargon, was impermissibly speculative because it failed to consider the relevant variables. The issue in Sargon was that the expert’s “attempt to predict the future was in no way grounded in the past.” (Sargon, supra, 55 Cal.4th at p. 780.) Here, Orszag relied on past data to explain his conclusions regarding Itanium’s predicted market share and provided a “logical basis to infer” that his conclusions were supported. (Cf. Sargon, at p. 781.) Orszag’s testimony explained his consideration of other dynamic factors that contributed to the drop in Itanium revenue— independent of Oracle’s conduct—including not only the impacts of competing and faster technology like the Xeon E7 chip but also the “shrinking” market and market-wide movement to x86 platforms. This evidence also supports a conclusion that Orszag considered multiple variables and did not attribute lost profits exclusively to a single factor. (Cf. Camper v. McDermott (1968) 266 Cal.App.2d 41, 46 [reversing damages award that relied exclusively on one factor and failed to consider “many” other relevant factors].) Therefore, we conclude that Orszag’s testimony provided a “reasonable basis for [his] opinion” (Sargon, supra, 55 Cal.4th at p. 772), and the trial court did not abuse its discretion in admitting the testimony. Oracle contends that Orszag’s calculation of projected future damages from 2015 to 2020 “is even more unsound.” Orszag testified that he used the most recent forecasts from HP’s quarterly financial data that HP had as of November 2015. He stated that he had projected his calculations to October 31, 2020, for “a number of reasons” including because the Itanium collaboration agreement between Intel and HP that was “signed in October of 2010 had a roadmap out past 2020” and there would be “chips available to be sold until about 2022.” In addition, he relied on statements that “they’re going to support 80 the product through 2025” and noted that for the RISC/EPIC market, “[o]ne of the witnesses for Oracle said that they had a roadmap through 2020 as well.” Oracle does not argue that Orszag lacked evidentiary support for these considerations. Instead, Oracle contends that it was speculative to award five years of projected future damages based on a constant market share given the “likely introduction of new products and technology and appearance of new competitors.” However, we disagree with Oracle’s characterization of Orszag’s projections into 2020 as “sheer fantasy.” Based on the market information that Orszag considered, and given the technology roadmaps upon which he relied, we cannot say his projections were illogical or lacked a reasonable basis. (See Sargon, supra, 55 Cal.4th at p. 772.) It was for the jury to consider the probative value of the evidence and the persuasiveness of Orszag’s projections, and we may not second-guess its conclusions on that score. We also do not agree that the evidence here mirrored that found inadmissible in Sargon. As the California Supreme Court stated, “[a]n expert might be able to make reasonably certain lost profit estimates based on a company’s share of the overall market.” (Sargon, supra, 55 Cal.4th at p. 776.) The problem with the expert’s analysis in Sargon was that he did not “base his lost profit estimates on a market share Sargon had ever actually achieved” but rather “he opined that Sargon’s market share would have increased spectacularly over time to levels far above anything it had ever reached.” (Ibid.) In contrast to the expert in Sargon, Orszag relied on actual market share data from a third-party upon which other companies, including Oracle, relied. Unlike the business in Sargon, HP is a long-established enterprise, and Orszag appropriately relied upon HP business information and data. Oracle does not dispute that HP is an established business or that lost profits could be quantified. Rather, Oracle argues that “an enterprise server business in a dynamic, competitive market requires more than an assumption that future market share over 81 almost a decade would exactly mirror the past.” In our view, whether other events—such as the introduction of Intel’s Xeon E7 microprocessor or HP-specific business issues— affected or undermined Orszag’s calculation of lost profits were factual matters for the jury. Oracle vigorously cross-examined Orszag on these and other topics and provided the testimony of a competing expert who challenged Orszag’s analysis. In sum, we conclude that the trial court acted within its discretion in admitting Orszag’s testimony and rejecting Oracle’s contention that Orszag’s expert opinion was impermissibly speculative.

D. HP’s Cross-Appeal

In its cross-appeal, HP contends the trial court abused its discretion in declining to award HP prejudgment interest under Civil Code section 3287, subdivision (b). HP maintains that it was entitled to prejudgment interest on the jury award for the period in which the phase 2 jury trial was delayed due to Oracle’s appeal from the denial of its anti-SLAPP motion. HP acknowledges the “high bar” to reverse the trial court’s discretionary decision but argues that where Oracle, through its delay, “extracted a three- year, ‘interest-free loan’ from HP in the amount that the jury determined Oracle properly owed,” fundamental fairness and sound judicial policy require a limited award of prejudgment interest to remedy the harm presented. “Prejudgment interest is awarded to compensate a party for the loss of the use of his or her property.” (Bullis v. Security Pac. Nat. Bank (1978) 21 Cal.3d 801, 815 (Bullis).) More specifically, it “provide[s] just compensation to the injured party for loss of use of the award during the prejudgment period—in other words, to make the plaintiff whole as of the date of the injury.” (Lakin v. Watkins Associated Industries (1993) 6 Cal.4th 644, 663 (Lakin).) Section 3287 authorizes the recovery of prejudgment interest on damage awards. The provision at issue here is section 3287, subdivision (b) (hereafter section 3287(b)). Section 3287(b) governs cases involving unliquidated contract claims and grants the court discretion to award prejudgment interest from a date no earlier than 82 the filing of the action.18 (North Oakland Medical Clinic v. Rogers (1998) 65 Cal.App.4th 824, 829.) We review the trial court’s ruling on prejudgment interest under section 3287(b) for abuse of discretion. (Faigin v. Signature Group Holdings, Inc. (2012) 211 Cal.App.4th 726, 752 (Faigin).) We will uphold the trial court’s exercise of discretion “if it is based on a ‘reasoned judgment’ and complies with the ‘legal principles and policies appropriate to the particular matter at issue.’ ” (Bullis, supra, 21 Cal.3d at p. 815.) HP requested prejudgment interest pursuant to section 3287(b) following the jury’s verdict in the phase 2 trial. HP sought prejudgment interest only on the past lost profits component of the $3.014 billion damages award, measured not from the start of the litigation in 2011, but from April 8, 2013, the date on which Oracle filed its appeal from the denial of its anti-SLAPP motion. The trial court denied HP’s request in a written order after a hearing. In its denial of HP’s request, the trial court cited several considerations. It recognized that Oracle filed its appeal of the anti-SLAPP order “just as Phase II of the trial was scheduled to commence,” which “had the effect of delaying Phase II for over two years.” It noted this court’s “strongly-worded opinion” affirming the denial of Oracle’s anti-SLAPP motion and deeming “ ‘the appeal, like the motion engendering it, . . . utterly without merit.’ ” (Quoting Hewlett-Packard, supra, 239 Cal.App.4th at p.1178.) And it weighed the “improper delay” caused by the appeal against several other factors. These factors included: (1) that HP’s damages “were not only unliquidated but

18 Section 3287(b) provides: “Every person who is entitled under any judgment to receive damages based upon a cause of action in contract where the claim was unliquidated, may also recover interest thereon from a date prior to the entry of judgment as the court may, in its discretion, fix, but in no event earlier than the date the action was filed.” (§ 3287(b).) While section 3287(b) authorizes the discretionary award of prejudgment interest on unliquidated damages, section 3287, subdivision (a) provides for the nondiscretionary award of prejudgment interest on liquidated damages, or damages certain. (See Espejo v. The Copley Press, Inc. (2017) 13 Cal.App.5th 329, 376.) 83 highly contested and uncertain;” (2) the uncertainty was amplified by “damages continu[ing] to accrue during the litigation while HP received at least partial performance under the contract;” and (3) the jury was aware of the delay caused by Oracle’s appeal and “may have considered the delay when it elected to award HP the full amount of the damages it requested without reduction.” The trial court found these factors “more significant” under the circumstances and declined to award HP prejudgment interest. Having considered the applicable law and reviewed the record on HP’s motion for prejudgment interest, including the trial court’s assessment of the issues at the hearing on the trial court’s then-tentative ruling, we cannot conclude the trial court abused its discretion. “ ‘An abuse of discretion occurs if, in light of the applicable law and considering all of the relevant circumstances, the court’s decision exceeds the bounds of reason and results in a miscarriage of justice. [Citations.] This standard of review affords considerable deference to the trial court provided that the court acted in accordance with the governing rules of law. We presume that the court properly applied the law and acted within its discretion unless the appellant affirmatively shows otherwise.’ ” (Faigin, supra, 211 Cal.App.4th at p. 752, quoting Mejia v. City of Los Angeles (2007) 156 Cal.App.4th 151, 158.) The trial court’s refusal to grant prejudgment interest on HP’s damages for past lost profits was neither irrational, arbitrary, nor contrary to applicable law and governing principles. HP contends that the trial court employed an incorrect understanding of the law because it conflated issues of uncertainty and delay. HP argues that while it may be unfair to require a defendant to pay full compensation if the reason for its delay was uncertainty about the amount owed—hence the discretionary component of section 3287(b)—uncertainty is not an independent basis for denying prejudgment interest under section 3287(b). We agree in principle with HP that uncertainty in the amount of damages is not alone determinative, because “[u]ncertainty is inherent in any award of prejudgment 84 interest under section 3287(b).” (Carmel Development Company, Inc. v. Anderson (2020) 48 Cal.App.5th 492, 525.) The original statutory scheme, prior to the enactment of section 3287(b), “provided for the recovery of prejudgment interest only where damages were ‘ . . . certain, or capable of being made certain by calculation.’ ” (A & M Produce Co. v. FMC Corp. (1982) 135 Cal.App.3d 473, 496 (A & M Produce), quoting section 3287, subdivision (a).) The 1967 amendment adding subdivision (b) to section 3287 “created a limited exception to the prevailing general rule that prejudgment interest is not allowed on unliquidated obligations.” (Lewis C. Nelson & Sons, Inc. v. Clovis Unified School Dist. (2001) 90 Cal.App.4th 64, 69 (Lewis C. Nelson).) The rationale for precluding prejudgment interest on unliquidated claims is “that it is unreasonable to expect a defendant to pay a debt before he or she becomes aware of it or is able to compute its amount.” (Lewis C. Nelson, supra, 90 Cal.App.4th at p. 69.) By allowing the trial court to consider awarding prejudgment interest on an unliquidated contractual claim within the limits prescribed by the statute, section 3287(b) aims “to balance the concern for fairness to the debtor against the concern for full compensation to the wronged party.” (Lewis C. Nelson, at p. 69.) The “discretion to the trial court to allow prejudgment interest even in cases where the amount of damages was ‘unliquidated’ . . . was designed to allow trial courts flexibility in circumstances . . . where the exact amount of damage is in dispute.” (A & M Produce, supra, 135 Cal.App.3d at p. 496.) Given the intent of section 3287(b) to enable trial courts to award prejudgment interest despite the uncertain amount owed on a contract claim for unliquidated damages, it would appear contrary to the statutory scheme for the trial court to refuse prejudgment interest for the sole reason that the amount of damages was highly uncertain. HP claims that the trial court did just that when it attributed more significance to the “highly contested and uncertain” damages than the “improper delay” caused by Oracle’s untimely

85 anti-SLAPP motion and subsequent appeal. We disagree with HP’s characterization of the trial court’s ruling. The trial court properly recognized that the unliquidated damages rendered prejudgment interest discretionary under section 3287(b), then articulated the factors it deemed most relevant to its decision. It declined to award prejudgment interest after weighing those factors, some of which it deemed “more significant considering all of the circumstances” at issue in the case. There is no authoritative list of criteria for courts to consider, and “[f]ew cases have discussed the standards by which a trial court’s exercise of discretion under section 3287, subdivision (b) are to be judged.” (A & M Produce, supra, 135 Cal.App.3d at p. 496.) We believe the trial court identified and properly considered the factors most relevant to the prejudgment interest question as it arose in this case, balancing as well as possible the permissible concerns of fairness and just compensation. (See Lewis C. Nelson, supra, 90 Cal.App.4th at p. 69; Lakin, supra, 6 Cal.4th at p. 663.) First, the trial court acknowledged not only this court’s condemnation of the dilatory impact of Oracle’s anti-SLAPP appeal (Hewlett-Packard, supra, 239 Cal.App.4th at p. 1178) but also noted that Oracle brought the motion and filed for appeal on the eve of the phase 2 trial. The trial court recognized that this factor supported HP’s request for prejudgment interest. Next, the trial court articulated its reasons for deciding that among the most important considerations under the circumstances of the case was the “highly contested and uncertain” damages. The court identified one aspect “[a]dding to this uncertainty” was that “damages continued to accrue during the litigation while HP received at least partial performance under the contract, when Oracle resumed or continued porting its software to HP’s servers following Phase I.” Although the trial court recognized that HP’s expert took Oracle’s partial performance into consideration in calculating damages,

86 it found that “Oracle’s partial performance mitigated HP’s damages during the appeal and added to the complexity and uncertainty of the damages.” The record thus demonstrates a nuanced assessment of these points and does not support the suggestion that the trial court “conflated” its consideration of delay with the uncertainty of damages. Nor are we aware of any authority under which the jury’s award—viewed in relation to the complexity or uncertainty affecting the unliquidated damage estimates—is an improper consideration. HP challenges the trial court’s supposition that the jury “may have” taken the delay caused by Oracle’s appeal into account when it awarded HP its full damages “without reduction.” It points out that the jury’s damages award cannot support a conclusion that the jury factored prejudgment interest into account when it was neither instructed nor authorized to do so, and when it awarded damages equal to the exact amount of lost profits calculated by HP’s expert, which did not include prejudgment interest or damages predicated on delay. HP moreover questions whether the trial court’s speculation on this point was proper since the decision to award prejudgment interest rests only with the court. (§ 3287(b) [authorizing trial court, in its discretion, to award prejudgment interest].) We perceive no error in the trial court’s consideration of the jury’s damages award. The record supports the trial court’s observations that the jury was “aware that there was a two and a half to three-year delay” and that it awarded HP damages “on the higher end of” the range requested.19 Case law, including Esgro Central, Inc. v. General

19 Indeed, HP’s closing argument in rebuttal raised the issue of the three-year delay due to Oracle’s appeal. In addressing the reduction in Orszag’s damages calculation from $4 billion in his 2012 report to $3 billion in his 2015 report, HP’s counsel reminded the jury that “we had a three-year delay because there was an appeal filed in this case by Oracle to cause a three-year delay. So Mr. Orszag had to go back in 2015 and look at his damages again and calculated them from 2012. [¶] He had three new years of data. He had actual hard data that was not available to him in 2012 when he

87 Ins. Co. (1971) 20 Cal.App.3d 1054 (Esgro)—upon which the trial court specifically relied in its order—supports the trial court’s approach. Esgro involved businesses that claimed prejudgment interest on a judgment entered on an insurance policy, where the extent of damage for losses covered was in dispute. (Esgro, supra, 20 Cal.App.3d at pp. 1062–1063.) The trial court denied interest under section 3287(b). (Id. at p. 1064.) On appeal, the court determined there was no abuse of discretion, despite a record that “undoubtedly” could have supported an exercise of discretion in favor of an award of prejudgment interest. (Id. at p. 1065.) The factors supporting prejudgment interest included that the respondent insurer’s refusal to honor the appellants’ original proof of loss “resulted in a delay of over four years in the payment of indemnity to appellants” and allowed the respondent “the benefit of an investment return for that period upon the sum first claimed by [the] appellants and eventually ordered paid to them by the trial court.” (Ibid.) But there were also factors supporting the trial court’s decision, specifically an inference drawn from the record suggesting that the trial court “denied prejudgment interest on appellants’ business interruption policy because, in view of counsel’s statement in closing argument, [the judge] was of the opinion that the jury had already considered that item in awarding damages.” (Ibid.) The court concluded that this provided “a basis in fact supporting the denial of prejudgment interest” and held that the trial court’s exercise of discretion under the circumstances was not “so unreasonable as to be an abuse.” (Ibid.)

did his first report. He could replace forecasts with actual data . . . and yes, some of it caused the damages to go up, and some of it caused it to go down. [¶] He reduced his damages from $4 billion to $3 billion because he said . . . [n]ow that I have this three years of data, that’s a fairer number.” The trial court’s observation that the jury chose “the higher end of” the range of damages is not inconsistent with Orszag’s presentation of three methodologies for calculating past damages (“kinetic” valued at $1.5 billion; “constant market share” valued at $1.7 billion; and “regression” at $1.8 billion), where he recommended—and the jury adopted—the “constant market share” approach. 88 HP submits that Esgro is inapplicable. HP points out that in Esgro, supra, 20 Cal.App.3d at page 1065, the record reflected an invitation by the appellants’ lawyers for the jury to consider the delay to judgment and revealed the trial judge’s assessment that the jury’s award was larger than would have been expected based on the evidence. Here, HP argues that its lawyers consistently told the jury that HP’s expert had lowered the calculated lost profits based on data obtained during the appeal, and the trial court never implied that the jury’s award was excessive. HP also emphasizes that in Esgro there was no claim the defendant’s actions were meritless or in bad faith, whereas Oracle’s decision to appeal the anti-SLAPP order was a bad faith delay tactic. (Hewlett-Packard, supra, 239 Cal.App.4th at p. 1178.) HP emphasizes that courts since Esgro have declined to recognize jury contemplation of prejudgment interest since they are not instructed to do so and because that discretion lies with the trial court. (See, e.g., George v. Double-D Foods, Inc. (1984) 155 Cal.App.3d 36, 48 (George) [holding trial court “erred in confiding the issue of whether prejudgment interest should be awarded to the discretion of the jury, rather than exercising its own discretion”]; Copart, Inc. v. Sparta Consulting, Inc. (E.D. Cal. 2018) 339 F.Supp.3d 959, 1004 [rejecting argument that “jury adequately compensated” the claimant based on Esgro and noting the jury instructions did “not contemplate an interest award or even some form of delay”].) Esgro is not factually identical to the circumstances here, but it is instructive as to the breadth of factors that may reasonably be considered in the court’s assessment of a request for prejudgment interest. Here, as in Esgro, the jury heard closing arguments that emphasized the delay in the plaintiff’s compensation for harm suffered. We note little difference between HP’s counsel’s reminder to the jury that “there was an appeal filed in this case by Oracle to cause a three-year delay” and the reference in Esgro to the passage of time since the businesses incurred damages. (Esgro, supra, 20 Cal.App.3d at p. 1065 [noting counsel’s argument that it had been “ ‘five years since the Watts riots occurred’ ” and “ ‘[w]e still haven’t gotten a nickel from this company’ ”].) Similar comparisons 89 may be drawn between the trial court’s observations about the damages verdict in this case and the trial court’s observations in Esgro. In Esgro, the trial court described the jury’s verdict as “ ‘a very, very substantial judgment . . . based upon the facts as [it] heard them’ ” (ibid.), whereas here the trial court noted that the jury awarded “the full amount of the damages [HP] requested without reduction.” We believe that taking account of the jury award was no less proper a factor for the trial court to consider in this case than it was in Esgro, where the appellate court credited it as a reasonable basis to deny the requested prejudgment interest. (Id. at p. 1063.) Nor do we interpret the trial court’s reference to the damages award in this case as an unsupported presumption that the jury included prejudgment interest in its calculations. In George, the trial court improperly placed the issue of prejudgment interest before the jury, rather than exercising the court’s own discretion whether to award it. The appellate court held the error was prejudicial, noting that shifting the decision about an interest award onto the jury “deprived [the] defendant of its right to judicial objectivity.” (George, supra, 155 Cal.App.3d at p. 48.) Here, unlike in George, the trial court acknowledged and exercised its discretion and articulated the factors that it viewed as supporting its assessment. The trial court’s observation that, despite the highly contested and uncertain nature of damages, the jury elected to adopt the expert’s preferred damages model at the “higher end” of the range presented, implies that the trial court found the jury’s assignment of damages fairly compensated HP for its losses, even when balanced against Oracle’s responsibility for the delay to the phase 2 trial. This was an appropriate exercise of judicial discretion, with a view toward effecting “the requirements of substantial justice.” (George, supra, 155 Cal.App.3d at p. 48.) HP further maintains that where the delay in payment is due to wrongful or vexatious conduct, justice and fairness require the defendant to compensate for the damage brought by the unjustified delay. To bolster its point, HP draws from a 1933 90 California Supreme Court decision, which says that “where delay in payment is due to vexatio[u]s conduct on the part of the defendant, ‘it is only just that he should repair the damage that has followed from the breach of his obligation’ although the balance due to the plaintiff is ‘in a certain sense unliquidated.’ ” (Hansen v. Covell (1933) 218 Cal. 622, 630 (Hansen).) HP urges that the same principle applies here. It argues that the trial court abused its discretion by excusing Oracle’s delay based on uncertainty of damages, when in fact the postponement of the time to judgment was attributable entirely to Oracle’s “vexatious conduct” and deliberate delay in filing what this court deemed was a meritless appeal. The above-quoted language, though evocative of HP’s position, does not address the issue before us. Hansen both predates the amendment to section 3287 authorizing prejudgment interest on unliquidated damages and involves a different issue—whether the trial court could properly award prejudgment interest on liquidated damages, where the defendant claimed the amount owed was not ascertainable by reason of an unliquidated offset. (Hansen, supra, 218 Cal. at pp. 630–631.) Hansen does not dictate reversal of the trial court’s order, as “ ‘ “cases are not authority for propositions not considered.” ’ ” (B.B. v. County of Los Angeles (2020) 10 Cal.5th 1, 11.) Even assuming the cited authority is instructive, however, as a general admonition that willful delay warrants compensation to the injured party, Hansen does not alter our evaluation of the trial court’s decision. The primary feature of section 3287(b) is that it confers discretion, allowing “the trial court the flexibility to determine whether an award of prejudgment interest is appropriate in light of the particular facts and circumstances in the case.” (Faigin, supra, 211 Cal.App.4th at p. 751.) The trial court here was aware of the delay caused by the anti-SLAPP appeal in this case and of this court’s indictment of the reasons for that delay. We conclude on this record, given the trial court’s balanced consideration of the delay against other mitigating factors—including Oracle’s partial performance under the contract during the period of delay and the jury’s award of the full 91 amount of damages requested “without reduction”—that HP has not demonstrated an abuse of discretion in the denial of its request for prejudgment interest.

III. DISPOSITION

The judgment is affirmed. Each party shall bear its own costs on appeal.

92 ______________________________________ Danner, J.

WE CONCUR:

____________________________________ Greenwood, P.J.

____________________________________ Elia, J.

H044371

Hewlett-Packard Company v. Oracle Corporation Trial Court: Santa Clara County Superior Court No. 2011-1-CV-203163 Trial Judge: Hon. Peter H. Kirwan

Counsel for Plaintiffs/Appellants Theodore J. Boutrous, Jr. Hewlett-Packard Company: Jeffrey T. Thomas Samuel G. Liversidge Rodney J. Stone Brandon J. Stoker Gibson, Dunn & Crutcher

Robert S. Frank, Jr. Choate, Hall & Stewart LLP

Counsel for Defendants/Appellants Margaret M. Crignon Oracle Corporation: Anne M. Grignon Grignon Law Firm

Dorian Daley Deborah K. Miller Peggy Bruggman Oracle Corporation

Fred Norton Bree Hann The Norton Law Firm PC

Sadik Huseny Sarah M. Ray Latham & Watkins LLP

Steven A. Ellenberg Ellenberg & Hull

William A. Isaacson Karen L. Dunn Boies, Schiller & Flexner LLP

H044371

Hewlett-Packard Company v. Oracle Corporation