Rambus, Antitrust & the Common Law

In the next few weeks, the FTC is expected to issue a final order in its five-year case against Rambus Inc., a California-based developer of memory technology. Rambus has proven to be the longest and possibly costliest litigation in FTC history. The FTC’s trial costs alone approached $3 million, with over $1 million going to “expert” witnesses and consultants.

The Rambus case started as a patent infringement dispute between the company and several memory manufacturers. Rambus doesn’t produce any memory itself; it develops and patents technologies and licenses them to manufacturers. During the mid-1990s, Rambus participated in a memory standard-setting group, JEDEC, and this is where the trouble began. The manufacturers claim Rambus misled JEDEC into incorporating Rambus patents into certain memory standards. Rambus said it was denied permission to present its technologies for standardization and that JEDEC members simply infringed Rambus’s patents.

In May 2001, a federal jury in Virginia ruled against Rambus in a counterclaim to a patent infringement lawsuit filed against Infineon Technologies AG. Infineon said Rambus violated JEDEC’s written rules regarding patent disclosure. Based on this finding, the FTC filed its own antitrust complaint in 2002, charging Rambus with “unfair” methods of competition in an attempt to monopolize several components of the computer memory market.

The Federal Circuit Court of Appeals overturned the Infineon verdict in 2003, stating that under the common law of fraud in Virginia, Rambus did not violate JEDEC rules. The FTC continued its case, however, maintaining that the antitrust laws imposed a higher disclosure duty than JEDEC’s written rules or the common law.

In February 2004, the FTC’s chief administrative law judge issued an initial decision dismissing the Commission’s complaint. It was only the second time in ten years that an ALJ dismissed a Commission complaint on the merits. Rambus’s victory was short-lived, however, as the full Commission reversed the initial decision on appeal and held Rambus liable for antitrust violations. A separate round of briefing and oral arguments to determine remedies concluded last month.

On top of all this, the memory manufacturers have dealt with their own antitrust troubles. The DOJ’s Antitrust Division eventually charged every major manufacturer except one—coincidentally, the only U.S.-based firm, Micron Technology—with participating in a price-fixing conspiracy that was designed, in part, to discourage the sale of memory products that incorporated Rambus patents. This has created the unusual scenario of the FTC prosecuting a case where its alleged victims and chief witnesses are themselves admitted antitrust violators.

Meanwhile, Rambus and the manufacturers have continued to duke it out in numerous patent infringement and civil antitrust cases throughout the country. The FTC’s proceedings have forced a delay in at least one case in San Jose, and the prospect of a Rambus appeal of the FTC’s final order means it could take another five years to sort out the underlying patent issues.

Regardless of one’s position in this fight—and there are strong advocates for both Rambus and the manufacturers—it’s become clear that the FTC’s intervention has only made the situation worse. Patent litigation is costly and complicated enough without having to run everything through an additional layer of antitrust litigation. But the bigger problem is that we have two judicial systems—the Article III courts and the FTC’s administrative process—working at loggerheads to resolve the same dispute.

The Federal Circuit rejected the fraud case because JEDEC’s rules were not drafted clearly on the critical issue of patent disclosures. (Actually, it was a dispute over whether Rambus had to disclose certain patent applications, not actual patents.) The FTC, conversely, said that the written rules were irrelevant, and that the Commission’s own subjective review of Rambus’s behavior was sufficient to determine legal liability.

The FTC’s decision to replace common law fraud with its own peculiar brand of antitrust liability is particularly harmful to the rule of law. Rather than deal with the rights and duties of the parties, the FTC wants to construct a parallel universe where economic outcomes conform to the Commission’s predetermined expectations. During the FTC’s hearing on remedies, which I attended, there were numerous references to the “but-for” universe—a fictional construct where the Commission imagines what life would have been like “but for” Rambus’s conduct within JEDEC. The regulators seek to “restore” market conditions that never existed in the real world. That’s not law enforcement; that’s central planning.

Additional links
FTC’s case information page

One Comment

  • Meanwhile, the market handled this one just fine: Rambus’s memory format, RDRAM, was never embraced by the market because it was expensive, had weird requirements (their memory sticks had to run in pairs, though a single stick could be run with a memory-less filler stick made to satisfy the pairing requirement). Rambus’s temporary marriage with Intel motherboard chipsets caused business to go to VIA, Intel’s major competitor. In short, the RDRAM from Rambus failed to compete with, much less displace the new DDRAM standard which is somewhere around its 3rd or 4th generation of tech these days.

    Amazing that someone sees evidence of price-fixing among RDRAM’s competitors– what did they do, collude to set their price much lower than their competitor’s? Heaven forbid!