On June 20, 2014, Judge Rogers of the Northern District of California issued a Daubert opinion in MediaTek Inc. v. Freescale Semiconductor, Inc., Case No. 11-cv-5341 YGR, addressing numerous grounds raised by the Defendant. The discussion on royalty rate and market share were very short and not particularly noteworthy, but the royalty base discussion addressed an issue that is of considerable importance and relevance to other cases.
The products at issue were computer chips. The defendant alleged that the plaintiff’s expert had not properly apportioned the royalty base to the smallest saleable patent-practicing unit, i.e., that the patents did not drive sales for the chips, and thus the chips could not properly be the base under Laserdynamics and other authority. The plaintiff countered by stating that the chips were in fact the smallest saleable unit, and there was “no way, conceptually or practically, to separate the patent-practicing features of the accused chips from the remainder of the chip.” The expert did opine, however, that the amount of profit to attribute to the invention would of course depend on the balance between infringing and non-infringing features. [It is not clear whether this was done as part of the base analysis or the rate analysis, but from context it appears to be as part of the rate.]
The Court noted that both sides had credible positions, and that this appeared to be a battle of the experts. Accordingly, the Court held that any disputes were ones that would go to weight, rather than reliability and admissibility, and denied the Daubert challenge.
This opinion sheds light on an issue that has been hotly contested since Laserdynamics, which is whether there is an expectation or requirement that a plaintiff’s expert “further” apportion the royalty base to something smaller than the smallest saleable unit. At this point, there does not appear to be much doubt that except in the instance of where the entire market value rule is satisfied, using a base larger than the smallest saleable unit is destined to fail (e.g., using a larger device such as a computer as the base when the patented feature is on a chip that is merely part of that larger device). There have been several district court opinions going opposite ways, however, on whether Laserdynamics and its progeny require an expert to apportion the base when there is so smaller saleable unit within the base that the expert has identified (e.g., using the same example, whether an expert must apportion the base to some fraction of the chip). Here the Court said no, but other district courts have held yes, creating considerable confusion for litigants (and for Judges being presented with conflicting authority). Federal Circuit input on this issue would (presumably) clear up this issue.
As a practical matter as to litigants, the value as to a certain feature should not change simply because of how “partition-able” the end-product is. Intellectually, if the base is 10x larger because it is not partition-able into 10 discrete parts, the rate should drop by a factor of 10 to come to the same result. Practically, however, this never seems to happen. This in turn creates market inefficiencies, because Companies are incentivized to create artificial partitions in products as a way of fending off exposure. With the law in its current state, a company selling a massive software or hardware product would be better served to split into nearly infinite parts, attribute a fractional cost to each of those parts, and then put all the parts together, rather than build a single integrated product with a single cost. From an engineering standpoint, however, this may not make any sense. Again, rational economic guidance from either Congress or the higher courts would assist all involved.