Caltech's $1.1 billion victory

As readers probably are well aware by now, in late January a jury in California returned a $1.1 billion verdict for the California Institute of Technology in its patent infringement suit against Apple and Broadcom.  For discussions in the media at the time, see, e.g, articles in Bloomberg, Law360, the Los Angeles Times, and Reuters.  More recently, Law360 has published a couple of articles on the parties' post-trial motions, here and here.

Aside from the amount of the judgment, one thing notable about the damages award is that the defendants chose not to put on a damages case at trial.  There was a good deal of skirmishing pretrial, however, about the admissibility of each sides' proposed damages experts' opinions.  Recently Professor Bernard Chao sent me a copy of the district court's redacted November 21, 2019 order on damages evidence.  Toward the order of the order, the court discusses the defendant's proposed methodology for apportioning the value of the technology in suit, which is interesting:

Defendants' experts conduct analyses where they count up certain things, like lines of code, and compare the total amount corresponding to the accused feature (e.g. all lines of code related to the allegedly infringing aspects of the product) to the overall total (e.g. all lines of code for the accused product). Specifically, Defendants' experts count: (1) lines of code; (2) physical die (circuit) space; (3) number of features; and (4) pages of the 802.11 standard as they correspond to the accused technology. See Docket No. 1280 at 2. Defendants' experts rely on these analyses as a basis for apportionment opinions. . . .

The Court agrees with Plaintiff that, whether or not other courts have accepted such counting methodologies as a reasonable input for determining proper apportionment, the opinions offered by Defendants' experts are not adequately tied to the circumstances of this case and, as presented, must be excluded. . . .

The Court particularly agrees with Plaintiff that source code line counting in this case is especially unsound given that different programming languages are involved for different code included in the overall decoder product.

The court left open the possibility that the use of methods (1), (2), and (3) might be permissible if better tied to the facts, but rejected method (4) outright.