Rambus CEO eager to move beyond the courtroom

Rambus' CEO wants to win the company's legal battles but also wants to move forward with new memory technologies.

Rambus Chief Executive Officer (CEO) Harold Hughes hopes his company's name won't always remind the memory industry of black-robed judges and endless pages of court filings.

It's been five years since Rambus and the memory industry kicked off their legal journey back and forth through various courtrooms across the U.S. In 2000, Rambus sued vendors such as Infineon Technologies, Hynix Semiconductor, and Micron Technology, alleging the memory companies were using Rambus' patented memory technology in their chips without a license. Those memory companies countersued, claiming Rambus' patents were invalid because the company should have disclosed them during the development of the SDRAM (synchronous dynamic RAM) standard.

The fiery rhetoric from both sides has cooled of late, after an unfavorable ruling from Judge Robert Payne of the U.S. District Court for the Eastern District of Virginia sent Rambus to the settlement table with Infineon in March. However, its patent claims against Hynix were cleared to proceed to trial on Monday. And an antitrust case as well as patent infringement suits over the emerging DDR2 (double data rate) standard will keep Rambus' Senior Vice President and General Counsel John Danforth, or as Hughes refers to him, "my rock star," busy for months to come.

Hughes, who became the Los Altos, California, company's CEO in January, sat down with IDG News Service Tuesday to discuss the latest developments in the company's legal drama as well as its participation in the development of the Cell processor with its memory and I/O bus designs. The following is an edited transcript.

Why, after so long, did [Rambus] choose that particular time to settle with Infineon? What led up to that settlement decision?

Foremost in my mind, I didn't want to put the company back into a three-year cycle of having Judge Payne's actions reviewed, and it's very hard to run a company when you have as many unknowns as we have. And that struck me at a time when Infineon saw their position as strong, and we were willing to come up with what we thought was a reasonable deal to avoid that three-year hiatus and all the unknowns.

We're trying to be an engineering company supplying services to people, and not a litigation mill.

Some observers have characterized the settlement as favorable for Infineon. It was a lump sum, not a per-unit volume deal, and seemed far smaller than some of the numbers that were bandied about in the early days of this legal dispute in terms of what these royalties would mean to Rambus.

As we've pointed out, we lost. But if you capitalize on that number, you get to a relatively attractive income level.

Do you see this Infineon case as a one-off, and not necessarily related to the other cases?

Yes. Rambus is a relatively young company, and was asked relatively early in its life to solve extremely hard problems. Solving a big problem for Intel and working with big DRAM companies is a difficult thing for a small startup to do. And as a consequence, not everything went perfectly. But things rarely go perfectly in business life, and the issue is, what do you learn from your mistakes and how do you restructure going forward?

So what have you learned?

Basically, how to make customers out of potential litigants. We're working very hard to do that; we're not in legal battles with all of the DRAM world. At the end of the day, we're coming down to just a few. We're getting to the point where we might have a higher volume in licensed than not licensed.

Are we seeing the end of days with all [the litigation]?

I put myself in the position of the CEO of Hynix, who like myself is an ex-finance guy, and Micron, who is not, and I, quite frankly, don't understand their thinking. What problem are they solving by continuing to litigate these issues?

If you're a business person, and you've got a problem you want to solve, you would initially, I would think, look first at where the market is going and how I can intercept it. We have a large number of gates or hurdles that people would have to jump in order to win these legal battles. It makes no sense to me.

But I don't want to talk about the legal stuff.

I understand. But rightly or wrongly, when people think of Rambus, they think of that.

Exactly. There's two ways to resolve that. One is by resolving the litigation, and number two is by making better strides in business applications for the business world.

I'd like to be known as someone who can work with the most successful companies in the world. If you step back a second and look at what we did, and it starts basically with our heritage, we solved the DRAM bandwidth problem relatively far in advance of where the DRAM market was going. We did that, as it turned out, in line with where Intel wanted the market to go.

But that sort of became our heritage, to create technology before the market wants it. Because that's where you can attract really, really good people who love to do semi-whacko things.

For us to then be successful in that, obviously, has some element of licensing as people grow into that. But it also gives us the opportunity to work with leading edge companies like Sony, the first to want to use things like that. It's an incredibly competitive world out there.

With regard to some of the memory and I/O technologies you've developed, what do you see as the primary application for some of those things in the next year or two? You've developed some extremely high-bandwidth products, but do you look at the PC market and wonder who's going to need that?

I couldn't agree more. I'd like to work more closely with Intel and some of the major OEMs to figure out how we can transition the PC into this world.

Obviously, a lot of that exists today in the Sony project, the games world. The Cell processor has eight CPUs sitting there, that puppy's looking for a lot of data.

Like everything, there's going to have to be some pretty good compiler work done in order for software to turn on all those circuits. The graphics world parallelizes very easily, so that's logically where it's going first. The games people are incredibly graphics intensive, but these things are going to have Internet connections where people are going to play games, and popping an e-mail program on there or a slot for a keyboard isn't going to be that hard.

So obviously Sony has big plans for this. You don't build a fab unless you're thinking volume, because small volume and fab is a recipe for total disaster.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about HISHynixHynix SemiconductorInfineonInfineon TechnologiesIntelMicronRambusSony

Show Comments