There are few companies in the tech world as infamous as Rambus, an IP-only RAM development firm. For the better part of 10 years now they have been engaged in court cases with virtually every RAM and x86 chipset manufacturer around over the violation of their patents. Through a long series of events SDRAM did end up implementing RAMBUS technologies, and the American courts have generally upheld the view that in spite of everything that happened while Rambus was a member of the JEDEC trade group that their patents and claims against other manufacturers for infringement are legitimate. At this point most companies using SDRAM have settled with Rambus on these matters.

Since then Rambus has moved on to a new round of lawsuits, focusing on the aftermath of Rambus’s disastrous attempt to get RDRAM adopted as the standard RAM technology for computers. Rambus has long held that they did not fail for market reasons, but rather because of collusion and widespread price fixing by RAM manufacturers, who purposely wanted to drive Rambus out of the market in favor of their SDRAM businesses. The price fixing issue was investigated by the Department of Justice – who found the RAM manufacturers guilty – which in turn Rambus is using in their suits as further proof of collusion against Rambus.

The biggest of these suits was filed in 2004 against the quartet of Samsung, Infineon, Hynix, and Micron. In 2005 Infineon settled with Rambus for $150 million while in 2010 Samsung settled with Rambus for $900 million, leaving just Hynix and Micron to defend. That suit finally went to court in 2011, with Rambus claiming that collusion resulted in them losing $4 billion in sales, which as a result of California’s treble damage policy potentially put Hynix and Micron on the line for just shy of $12 billion in damages.

Today a verdict was finally announced in the case, and it was against Rambus. A 12 member jury found in a 9-3 vote that Hynix and Micron did not conspire against Rambus, effectively refuting the idea that RAM manufacturers were responsible for Rambus’s market failure. As Rambus’s business relies primarily on litigation – their own RAM designs bring in relatively little due to the limited use of RDRAM and XDR – this is a significant blow for the company.

Rambus can of course file an appeal, as they have done in the past when they’ve lost cases, but the consensus is that Rambus is extremely unlikely to win such an appeal. If that’s the case this could mean that this is the beginning of a significant shift in business practices for Rambus, as while they have other outstanding cases – most notably against NVIDIA – the anti-trust suit was the largest and most important of them. Not surprising their stock also took a heavy hit as a result, as it ended the day down 60%. Rambus winning the anti-trust suit had long been factored into the stock price, so the loss significantly reduced the perceived value of the company.

Source: The Wall Street Journal; Reuters; Businesweek

Comments Locked


View All Comments

  • Beenthere - Thursday, November 17, 2011 - link

    Definitely patent trools too. Reality doesn't change due to denial.
  • zzss - Friday, November 18, 2011 - link

    We were working on both SDRAM and RDRAM motherboard/system at that time and below are the facts.
    RDRAM technology works on the principle of narrow bus with high speed frequency to provide the bandwidth, hence compare to SDRAM (or DDR):

    1. Memory module design was more stringent due to high speed signal. Higher power vreg was required due to high speed RAM chips. Higher power dissipation led to the requirement of heat sink on the RAM chips and with more expensive RAM chips, the memory module cost was higher.
    2. The Chipset memory controller design was more tough due to the high speed bus and the need for a license fee.
    3. Motherboard design was a nightmare. The requirement was so tight that there were no room for error. The team spend a lot of hours tuning the high speed trace and yet the result was not as stable as SDRAM board. On top of that, it needed at least 6 layers PCB. So, higher development and material cost for the motherboard.
    4. All the RDRAM slots need to be filled up with active DIMM or dummy bypass DIMM.
    5. It was the "Pentium 4" of the memory world. Look good on paper but not practical in real world.

    All in all, we were looking at a more expensive solution for system manufacturer and consumer.
    I am glad the industries move to the right direction.
  • thorr2 - Saturday, November 19, 2011 - link

    Yikes, you just brought me back to the days when I was using Computer Shopper magazine to buy a 386 computer. There were some companies that had nice computers that would match any competitor's price. However, when I called them and said "Look at Computer Shopper page 147, I want that price" they would say "We have a 6 layer motherboard. They must have a 6 layer motherboard or we won't match the price." What a bunch of crap that was.
  • bennyg - Friday, November 18, 2011 - link

    *Samsung slaps forehead*

  • RaistlinZ - Saturday, November 19, 2011 - link

    Oh lord, this company needs to DIAF. Their technology was terrible, slow, and expensive. I can't wait until they have no one left to sue and go bankrupt.
  • JonnyDough - Saturday, November 19, 2011 - link

    "Rambus can of course file an appeal, as they have done in the past when they’ve lost cases, but the consensus is that Rambus is extremely unlikely to win such an appeal"

    Consensus of WHOM? This article seems to lack credibility without stating "who" is doing the deliberation. The court clerks who handle the paper work for this case? The people around your office? Some guy on the net? WHO?!!!
  • vitality - Sunday, November 20, 2011 - link

    Here is some Rambus humor:
  • vitality - Friday, May 4, 2012 - link

    More Rambus humor:
  • MRFS - Thursday, September 26, 2013 - link

    Sorry to jump in so late: one key FACT which is being entirely overlooked by most if not all observers of the Rambus saga is the abject criminality of all California attorneys: extensive research has proven that not one of the 200,000+ past or present "members" of The State Bar of California has ever had a valid license to practice law in that State. See Sec. 6067 of the State Bar Act, for starters. Moreover, all U.S. Attorneys stationed anywhere in California are obligated by the McDade Act to have executed the same certificate of oath. Conclusion: every attorney that appeared formally for any of those Proper Parties committed multiple acts of FELONY mail fraud (see PROOFS OF SERVICE) and multiple acts of FELONY wire fraud (e.g. every time they transmitted email or picked up the telephone). There were also comments above about "DOJ's Opinion": well, those Federal employees are now turning up in droves with fatally defective credentials, as proven by the obvious absence of OMB control numbers and other fatal defectives on their Office of Personnel Management SF-61 APPOINTMENT AFFIDAVITS required by 5 U.S.C. 2906 and 3331. Hey, folks, I didn't make these laws; Congress did! WAKE UP, California, because the global surveillance "state" is now in your face, and the ATTORNeys are still laughing all the way to the banks who "bankrolled" that R&D. p.s. Anand knows nothing about any of this, and I doubt that he even cares. Someone should invite him to post a reply here. I can't wait: one of his employees once accused me falsely of posting pornography on the Internet, but our homepage had been hacked with JAVA malware after a confidential password was stolen from my INBOX. Gloves are now OFF here: I LUV a good fight. Now you know why I moved to Seattle.

Log in

Don't have an account? Sign up now