The Federal Circuit’s recent decision in the litigation between Columbia University and Gen Digital is notable not only for its treatment of software patent eligibility, but also for what it says about potential expansions in the geographic limits of patent damages, especially in the context of software patents.
As discussed in our prior blog, the court remanded the § 101 issue for further analysis of whether the asserted software claims, though directed to an abstract idea, contain an inventive concept under Alice, step two. But even if its patents survive § 101 on remand, Columbia lost the $94 million of royalty damages attributable to Norton’s sales to customers located outside the U.S. Applying the principle from the 2006 Microsoft v. AT&T decision that software code reproduced overseas from a master copy sent from the U.S. does not contain any U.S.-made components, the Federal Circuit held that Columbia was not entitled to damages for Norton’s foreign software sales because, similarly, that software was not “made in or distributed from” the U.S.
In Microsoft, the Supreme Court held that software in the abstract (i.e., not yet encoded on a specific physical medium) is simply a “blueprint” or “template,” and does not become a tangible product until it is installed on a particular computer. The implication for digital distribution is significant: when software is transmitted from a server in the U.S. to a foreign customer, and that foreign customer’s computer downloads and encodes the software onto its own hard drive, the potentially infringing copy is made abroad; it doesn’t matter that the original code resided on a server in the U.S.
The Federal Circuit applied this logic from Microsoft to all four of the asserted claim types in Columbia’s patents. For example, there is no system as claimed until software is installed on a computer, which occurred outside the U.S. Similarly, the claimed methods are not performed until the software is run, which also happens outside the U.S. With respect to the computer-readable medium claim, which requires software encoded on a non-transitory medium, there was no medium as claimed until foreign computers wrote the software to their own hard drives, again, outside the U.S. In each instance, the copies came into existence overseas and thus were not made or distributed from the U.S.
The Federal Circuit recognized, however, that there was a viable legal theory under which Columbia might have recovered foreign damages: the causation theory articulated in Brumfield, Trustee for Ascent Trust v. IBG LLC, No. 2022-1630 (Fed. Cir. 2024). Under this theory, if a patent owner proves domestic infringement under 35 U.S.C. §271(a) based on some act of making, using, offering to sell or selling of the patented invention within the U.S. (e.g., maintenance and use of a master copy), it can recover damages for foreign actions that are proximately caused by that domestic infringement (e.g., based on foreign downloads and installation of that master copy).
The 2024 Brumfield decision “superseded” the often-cited rule set forth in the court’s 2013 Power Integrations, Inc. v. Fairchild Semiconductor International, Inc. decision that foreign damages are not available, which reflects the general rule that products made and sold outside the U.S. cannot infringe a U.S. patent. But there is an exception to this general rule, set forth in 35 U.S.C. 271(f), which makes the distribution of components of a patented invention from the U.S., for assembly abroad, an act of infringement. The Supreme Court’s 2018 WesternGeco LLC v. ION Geophysical Corp. decision provided a framework for determining lost profits damages under this special statute. The Brumfield decision expanded the damages framework set forth in WesternGeco, applying it to reasonable royalty damages and to assessments of standard acts of U.S. infringement as set forth in 35 U.S.C. 271(a).
Here, Norton’s creation of master copies on domestic servers might have qualified as a domestic act of infringement that triggered (proximally caused) the foreign sales. But Columbia never requested that the jury be instructed on this Brumfield theory. Instead, Columbia requested that the jury be instructed with a different theory: that foreign sales were recoverable simply because the software was “made in or distributed from the United States.” The Federal Circuit held that this jury instruction was legally insufficient in view of Microsoft, since distribution of software in the abstract, which occurred in the U.S., is not the same as making or using the infringing copy, which took place abroad. Having already committed to this theory at trial, and having failed to present the Brumfield causation theory to the jury, Columbia could not (and the Court would not) replace Columbia’s invalid damages theory with the potentially valid Brumfield causation theory.
In patent cases involving software products distributed to foreign customers and, more broadly, in any patent case where acts of infringement in the U.S. serve as a causal precursor to foreign sales, a plaintiff should identify, disclose, and preserve every viable damages theory from the outset―including those that are now plausible under Brumfield. A plaintiff should carefully develop and disclose those theories during fact and expert discovery, ensure that those theories are presented to the jury at trial and supported by a valid jury instruction, and create a verdict form capturing the factual findings needed to sustain the award.
Columbia’s failure to submit the Brumfield damages theory to the jury was costly, leading to waiver of an opportunity to preserve the approximately $94 million in damages award tied to foreign sales when its other theories failed. While Brumfield provides new and expansive opportunities to obtain damages for foreign sales, the lesson from Columbia is clear: Any patent plaintiff who wishes to recover damages based on foreign sales under the Brumfield causation theory must properly develop the record and provide clear instructions on its requirements to the jury, rather than conjuring it post-verdict to replace a failed damages theory on appeal.
