Commission Affirms NVIDIA Violated Rambus Patents

The International Trade Commission has announced its findings in the NVIDIA/Rambus patent infringement lawsuit and it's not the sort of ruling Team Green would've preferred. The commission found NVIDIA to be in violation of three Rambus patents. The trade panel also granted an injunction Rambus had requested, which theoretically prevents NVIDIA and the various companies attached to the lawsuit (Asus, HP, Palit, and MSI among others) from selling products that contain the infringing IP. This last bit sounds more ominous than it actually is; there's a 60-day window before the injunction takes effect and NVIDIA believes it has other options.

"There will be no impact on our customers or our business as a result of this ruling," NVIDIA spokesperson Hector Marinez said. "We intend to take advantage of the mandatory European Commission License that is available. This will allow us and our partners to continue our business under the terms of that license and prevent the enforcement of any exclusion order." The impacted companies might disagree with this a bit—HP, Asus, and the rest can continue to sell products during the 60-day review process, but must post a 2.65 percent bond on every product sold.

The commission's decision this week affirms a January ruling that saw NVIDIA found in violation of three Rambus patents while dismissing two additional claims of infringement Rambus made. The two companies, predictably, are spinning the decision in two completely different ways. Rambus senior VP and general counsel, Thomas Lavelle, stated that "The ITC's decision is another demonstration of the value of our continued commitment to innovation...We are extremely pleased with the ITC's decision...The value of our patented inventions has been recognized by our current licensees, and we will continue our efforts to license others."

Like this, but with RAM chips.

Rambus has a certain history of patent trolling/aggressive defense (pick your poison) but it's impossible to ignore the general judicial findings of the past ten years. As the dust has settled on each successive wave of litigation, Rambus has consistently demonstrated that it possesses valid patents that cover certain core technologies at the heart of both DDR-RAM and memory controllers. The memory controller angle is part of what makes the NV/Rambus case unique—historically, most of the companies' lawsuits have dealt with memory companies themselves, not memory controllers.

NVIDIA's position is a bit strange on this one. They claim that the additional licensing fees from Rambus will take less than one percent off their gross margin, which raises the question of whether it might not have been smarter to just buy a license in the first place. This information hits at a bad time for NVIDIA—while analysts don't believe the ITC decision will impact the company's performance, many are concerned about Tegra and Fermi. Analyst Doug Freedman of Gleacher & Co has written the following:
Today we are lowering our estimates to reflect our conviction...that the company faces an uphill battle in the near- to-medium-term to win the favor of gamers and investors alike.Nvidia has been adversely impacted by: 1) sluggish traction for Tegra (roughly 1 year behind revenue plan on lack of infrastructure in software creating an air pocket before the Tegra 2 & 3 wins ramp); 2) less than stellar reviews on the launch of Fermi (particularly the GTX480 and 470, though reviews for the GTX460 are more favorable); and 3) rising operating expenses to support newer initiatives (operating expenses up 19 percent year-over-year, up 9 percent quarter-over-quarter).
According to EETimes, the question marks don't stop there. Craig Berger, an analyst with FBR, has a laundry list of concerns, including NVIDIA's decision to exit the chipset market, a recent spike in accounts receivables and inventories that might indicate channel-stuffing, notebook GPU losses to AMD, questions regarding the degree of hype around Tegra, Tesla, and Fermi, and some limited worries that channel stuffing in Q2 could backfire in Q3, thus disrupting seasonal trends.

Such fears could well be groundless, make no mistake. Inventory spikes could be explained by TSMC actually figuring out how to build 40nm products, the chipset segment would've mostly vanished whether NVIDIA began pulling out or not, and there's no serious evidence that NVIDIA is overloading ODMs. Even the Tegra delays could prove nothing more than growing pains in the long term. The ITC's decision doesn't necessarily hurt NVIDIA in any material way, but it does spread additional FUD at a time when the company doesn't have much certainty to spare.