Kathleen Sullivan, Quinn, Emanuel, Urquhart & Sullivan partner
Kathleen Sullivan, Quinn, Emanuel, Urquhart & Sullivan partner (Jason Doiy / The Recorder)

SAN FRANCISCO — Before Judge Lucy Koh held up two tablet computers and challenged attorneys to identify the designs, before Robert Van Nest wheeled a file cabinet into Judge William Alsup’s courtroom, Oracle v. SAP was the “it” trial of Silicon Valley.

An SAP A.G. subsidiary stood accused of hacking into Oracle Corp.’s computers and illegally downloading millions of files. Oracle CEO Larry Ellison testified that the theft cost his company $4 billion. Former SAP chief Leo Apotheker, by then Hewlett-Packard Co.’s CEO, played Where’s Waldo with process servers.

Heavyweights from Boies, Schiller & Flexner; Bingham McCutchen; and Jones Day duked it out for three weeks. And the trial ended in November 2010 with a whopping $1.3 billion copyright judgment.

But U.S. District Judge Phyllis Hamilton of the Northern District of California cooled things off by ordering a new trial while limiting Oracle’s theory of damages. The two companies struck a $426 million settlement that preserved Oracle’s right to get the original verdict reviewed by the Ninth Circuit. “Now Oracle has to put up or shut up on appeal,” Jones Day’s T. Gregory Lanier said at the time.

That time comes next month, and the case raises an issue that has never been squarely confronted by the U.S. Court of Appeals for the Ninth Circuit: whether a jury can award damages based on a hypothetical license when the plaintiff has no track record of granting comparable licenses.

Oracle says “yes” and is bringing reinforcements to make its case. Quinn Emanuel Urquhart & Sullivan partner Kathleen Sullivan will argue that the jury properly determined what Oracle would have charged Germany’s SAP for the software and support documents purloined by its subsidiary, TomorrowNow. “The value of a hypothetical license is a long accepted measure of actual damages under the Copyright Act where, as here, it is based on objective evidence,” Sullivan argues in her brief to the Ninth Circuit.

Jones Day’s Lanier will argue that Hamilton properly restricted damages to actual lost profits plus infringer’s profits, because Oracle never would have licensed its software to its No. 1 rival. “There was no licensing opportunity to lose, so no license fee (however measured) could properly be awarded as actual damages in this case,” Lanier wrote in his brief.

Despite all the copyright cases litigated in Silicon Valley and Hollywood over the decades, there’s no clear law in the Ninth Circuit for estimating a hypothetical license under these circumstances. “There is no case on point,” said University of California Hastings College of the Law copyright expert Ben Depoorter. “This case is unique because this is a business transaction that would never take place.”


Oracle acquired PeopleSoft Inc. for $11 billion in 2004, leaving SAP as its chief rival for business-management software. SAP responded by spending $10 million to acquire TomorrowNow, a Texas-based company that provided support services to a few hundred PeopleSoft customers. The plan was to offer “safe passage” to companies who did not want to migrate to Oracle or who found the transition disruptive.

But TomorrowNow was taking a huge, illegal shortcut. The company used its customers’ logins and passwords to access Oracle websites and systematically download thousands of PeopleSoft software updates and millions of customer-support documents. “This case is about corporate theft on a grand scale,” Oracle proclaimed in its 2007 complaint. SAP shut down TomorrowNow, which eventually pleaded guilty to criminal copyright charges.

Oracle alleged that SAP executives had been complicit in the scheme, and with trial looming in 2010, SAP defused the issue by accepting liability. The company acknowledged that “TomorrowNow made mistakes,” and agreed to pay Oracle for them. “That compensation must be based in reality and the law, however,” the company declared.

Oracle’s version of reality was that a license for all of that software would have been worth between $881 million and $2.69 billion, according to its expert witness. Ellison put the number at $4 billion, on the assumption SAP could have used TomorrowNow to peel off several thousand Oracle customers. In closing arguments Boies Schiller’s David Boies, who tried the case with Bingham McCutchen’s Geoffrey Howard, asked for $1.7 billion while also suggesting as much as $3 billion.

SAP argued those numbers were based on pure guesswork because Oracle had never licensed its software to competitors. Such a license would be “unprecedented,” Oracle copresident Safra Catz acknowledged on the witness stand. Former copresident Charles Phillips called it “unthinkable.”

The better measure of damages, SAP argued, was Oracle’s actual lost profit because of the illegal copying, plus any profit SAP gained from infringing. Given that TomorrowNow never grew its customer base beyond 358 companies, that amount was small. Jones Day’s Robert Mittelstaedt pegged it at $28 million in closings.

Hamilton let the jury choose the measuring stick, leading to the $1.3 billion award. Afterward she agreed with SAP that Oracle had not presented enough concrete evidence about the value of a license—to SAP or any other company. Without it, “any such award would be based on a subjective, not an objective, analysis of fair market value,” she concluded.


Quinn Emanuel’s Sullivan will argue to the Ninth Circuit on May 13 that Oracle can’t be required to present evidence of actual comparable licenses. Doing so would unfairly penalize plaintiffs such as Oracle who choose not to license their intellectual property to competitors, she argues.

Instead, the jury was entitled to look at the massive nature of TomorrowNow’s illegal copying, the fact Oracle had just paid $11 billion to obtain the software via its PeopleSoft acquisition, and SAP’s own internal projections about customer acquisition through TomorrowNow.

Analogizing to patent law, Sullivan argues that kind of evidence can inform a reasonable royalty based on Oracle’s expected losses and SAP’s anticipated profit at the time of the hypothetical negotiation in 2005—not in hindsight after TomorrowNow’s failure.

Given those factors, “a rational jury easily could have found the total value of the required licenses to be $1.3 billion,” Sullivan wrote.

Nor can she resist one more dig at SAP. “Both parties relied on that data in making extraordinarily important business decisions—Oracle in spending $11 billion to purchase PeopleSoft; SAP in undertaking a plainly illegal course of conduct and exposing itself to substantial liability,” she wrote.


Jones Day’s Lanier will argue that because Oracle did not seek statutory damages, it’s limited by the Copyright Act to “actual damages” caused by the infringement. The phrase “actual damages” appears six times in the first three paragraphs of his brief.

To Lanier, that means a hypothetical license is not available unless there’s concrete proof of lost license fees or opportunities to license.

“Because Oracle is not in the business of licensing its copyrighted works to third-party support providers,” he argues, “it cannot argue that SAP’s infringement diminished the licensing value of the infringed works.”

The Ninth Circuit has permitted the hypothetical-license theory only in “limited circumstances,” Lanier wrote, such as when the plaintiff had previously licensed the material to the defendant or others for the same use. That creates an objective benchmark for determining the market value.

Instead of presenting such evidence, Oracle pointed to “self-interested opinions of its executives” and the cost of acquiring PeopleSoft, which included not only intellectual property but real estate, office equipment and other assets. All the while Oracle poured on “days of prejudicial and inflammatory liability evidence, even though liability had been conceded,” Lanier wrote.

At bottom, Oracle is trying to import a theory of reasonable royalties from patent law so it can argue for “outrageous multibillion-dollar license fees,” Lanier argues. But Congress opted to set statutory damages, not reasonable royalties, as the floor for copyright infringement. “This court should reject Oracle’s attempt to rewrite the Copyright Act to create a new floor for damages,” Lanier wrote.


Objectively measuring a business transaction that never would have taken place in the real world is a challenge, Hastings professor Depoorter said, but it’s not impossible.

“It’s a very fascinating case,” he said. “It’s one where both parties take very extreme positions and the truth is probably somewhere in between.”

He doesn’t see Oracle’s cost of acquiring PeopleSoft as especially relevant. “It’s not sufficient to say, ‘We bought this company for X number of dollars.’ This isn’t about a sale,” he said. “That’s not what a license is.”

On the other hand, if damages are limited in the way SAP is proposing, that might not do enough as a policy matter to discourage illegal copying, he said.

Ultimately, the Ninth Circuit may have to decide how high a burden to place on a copyright owner to prove damages, and how willing courts will be to let an infringer off the hook when there are no benchmark licenses.

“This is cutting edge,” he said. “I haven’t seen a case like it.”

Contact the reporter at sgraham@alm.com.