On February 5 the U.S. Federal Trade Commission (FTC) issued its final opinion in the long-running, and often convoluted, legal proceeding against Rambus, Inc., a computer technology company. The Commission began its case back in 2002, charging that Rambus had violated Section 5 of the FTC Act through the firmâ€™s deceptive practices at the standard setting body JEDEC (Joint Electron Device Engineering Council).
JEDEC promulgates standards for the memory chip market, DRAM in particular. Rambus was a member of JEDEC and participated in DRAM working groups for nearly five years. Even though JEDEC requires its members to disclose intellectual property that might be relevant for any of its standards, during its time with JEDEC Rambus disclosed only one of its many patents and patent applications relevant for the DRAM standards. The firm then left JEDEC right before the final vote on key DRAM standards. Once the standards were introduced, Rambus reemerged to charge JEDEC members with patent infringement.
The FTC found this behavior anticompetitive, and on February 5, issued its Opinion of the Commission on Remedy and Final Order. Under the terms of this order, Rambus is barred from making â€œmisrepresentations or omissionsâ€ to any standard setting organization. More significantly, the FTC set the maximum royalty rates that Rambus can charge firms implementing two DRAM standards. These rates are not to exceed 0.25 percent to 0.5 percent, depending on the standard, for three years, after which the rates drop to zero. While the FTC recognized that â€œRoyalty rates unquestionably are better set in the marketplace,â€ it argued that â€œRambusâ€™s deceptive conduct has made that impossible. Although we do not relish imposing a compulsory licensing remedy, the facts presented make that relief appropriate and indeed necessary to restore competition.â€
The FTC ruling sends some very clear messages. First, failing to disclose relevant intellectual property to a standard setting body is now a very risky strategy. The ruling implies that firms must â€œdisclose it or lose itâ€. It also makes clear that quitting a standard setting organization does not provide safe harbor. If a firm participates in a standard setting organization, it is bound by that organizationâ€™s rules whether or not it is still a member when the standard is commercialized. While these implications likely go farther than any standard bodyâ€™s intentions, they do clarify the rules of the game and make hold-out/hold-up tactics like Rambusâ€™s more difficult.
But the ruling could also set some perilous precedents. In particular, the royalty rates set by the order implicitly suggest a patent counting scheme where two patents are necessarily twice as valuable as one patent. Along these lines, the Commission granted one DRAM standard a royalty cap of 0.5 percent and another 0.25 percent because the latter had half as many patents. This assumption, while convenient, flies in the face of decades of scholarly research that shows patents are not all equal. One patentâ€™s value may far exceed anotherâ€™s.
Moreover, the Commission has assumed that royalty rates would be expected to decline over time. Unlike patent value, there is little scholarly work on the dynamics of licensing terms. One can easily imagine realistic scenarios in which royalty rates would actually increase over time â€“ say if a new technology were offered at introductory rates until its value was established in the marketplace.
We can only hope that these less attractive elements of the Rambus opinion do not become precedents.