Rambus Inc (RMBS) 2002 Q3 法說會逐字稿

Welcome to the quarterly conference call for Rambus Inc. This will cover the financial results of our fiscal third quarter. I am Bob Eulau, Senior Vice President of Finance and Chief Financial Officer, and Geoff Tate, our Chief Executive Office,r is joining me here today. I'll start with a summary and analysis of the financial results for the quarter, and then turn it over to Geoff for a discussion on the business progress we've made in the last quarter. We'll then open up the lines for questions and answers. We've e-mailed the financial release to everyone on our current list. If you have not received it, it's available on our website at www.rambus.com. A replay of this conference call will be available for the next week at 800-293-4252. In addition, we are simultaneously web casting this call, and it can be accessed on our website for one week beginning at 4:00pm pacific daylight time.

Before we begin, I need to state that our discussion will contain forward-looking statements regarding the company's financial prospects, development plans, anticipated product shipment dates, relations with licenses, and other third parties, and various other matters, and actual results may differ materially. Among the reasons which could cause actual results to differ materially is the possibility of inadequate shipments of RDRAM memory devices and controllers for the Sony Playstation 2 and the PC main memory market. The market response to these products, any deterioration in the DRAM market, including declining prices, any delay in the development of rambus-based products by licensees, any delay in the development and shipment of new rambus products, a greater than expected -- a greater than expected response of the market to competing technology, a lack of progress on price and cost reduction by RDRAM suppliers, inadequate progress on signing new contracts for RDRAM's Yellowstone SDRAM, or DDR memory devices and controlers, inadequate progress on finding new contracts for RaSer serial link, current licensees not fulfilling their contract obligations, current licensees terminating their contracts, adverse litigation decisions and other factors, which are available on our S.E.C. filings, including our 10-K and 10-Q.

This was our fourth quarter in a row where earnings per share were at least 6 cents. This was consistent with the guidance that we gave in April for this quarter. This was the first quarter in the last three in which revenue increased sequentially. Our operating profit was up 304% over the same period last year, and it was down 12% from the previous quarter. The dramatic improvement in operating profit from last year, is primarily driven by a significant reduction in litigation costs versus the third fiscal quarter last year.

Now a bit more detail on revenue. Total revenue for the quarter was $23.7 million, up 1% sequentially and up 2% over the same quarter last year. This is in line with the guidance I gave last quarter. Total royalties for the quarter were $22.2 million, which is up 2% relative to last quarter, and up 12% over the same period last year. Please remember that royalties we're reporting this quarter relate to our customer shipments that took place in the first calendar quarter for this year. RDRAM royalties were consistent with the seasonal patterns that we had expected from both the memory device and memory controller standpoint. RDRAM memory devices had sequential revenue decline of 15% and a sequential unit decline of 18%. For RDRAM chip sets, given our agreement with Intel, we only have unit information on chip sets supplied by other customers. RDRAM controllers, excluding the Intel chip sets, a sequential revenue decline of 24% and a sequential unit decline of 11%. SDRAM and DDR royalties were up sequentially about 21% for memory devices and controllers combined. This was also in line with our expectations since pricing was much stronger in the first calendar quarter than it was in the December quarter of last year.

One of the frequent questions that we have been asked is; what percentage of the SDRAM and DDR royalties comprised of our total royalties. The contracts were SDRAM and DDR memory and memory controller royalties comprise about one-third of our revenue for the quarter that ended in June.

Contract revenue in the third fiscal quarter of 2002, was $1.5 million, down about $100,000, which is consistent with our guidance. It is interesting to note that about 30% of our contract revenue now comes from the RaSer serial link business. We believe that contract revenue will stay at these levels until we've signed a Yellowstone licensee. Spending for the June quarter was, as expected, overall with litigation being slightly higher than we had anticipated. We spent $16 million, which was up 8% from last quarter and down 25% from the same quarter last year. Cost of litigation was up about $700,000 from last quarter, at $2.4 million, and down about $6.4 million from the same quarter last year. The litigation expense was higher than we anticipated at the beginning of the quarter due to expenses related to the SEC investigation and the complaint that was filed. Due to the additional work we did in Europe and due to some additional preparation for the Federal Circuit Appellate Argument on June 3rd.

Operating expenses, excluding cost of litigation, were up 4% compared to last quarter, and they were up 10% when compared to the June quarter last year. The increase over last quarter's primarily related to research and development in our Yellowstone and RaSer technology areas.

From a profitability standpoint, this was another good quarter. Net income for the third fiscal quarter was $5.9 million, or about 25% of revenue. Compared to $6.7 million in the previous quarter and $3.7 million in the same period last year. This profitability led to strong cash flow for the quarter, such that even while investing at 7.9 million dollars in share repurchases, our cash, cash equivalents, and marketable securities during the third fiscal quarter, increased to $163 million from $162 million last quarter and $147 million as of June 2001. The balance sheet clearly demonstrates the strength of our financial model.

We announced last October our intent to repurchase up to 5 million shares of our common stock over an undefined period of time. As of the end of June, we've now repurchased about 2.5 million of the shares authorized. A key objective of the program is to reduce the dilutitive effect of employee stock options. in the quarter just ended, we repurchased just under 1.5 million shares at an average price of $5.33. Volumes of purchases will vary by quarter based on the opportunity. At the end of this fiscal quarter, we had average basic shares outstanding of 100 million and diluted shares outstanding of 102 million.

Now I'll give you some guidance on what to expect in the next quarter. This guidance obviously has a lot of uncertainty associated with it. This guidance reflects our best guess at this time and our actual results could differ materially from what I'm about to review. We do not expect revenue to differ significantly from the revenue that we recognized this quarter. We expect that spending will increase next quarter, and nonlitigation operating expenses, we expect that spending will be up by about a million dollars. This is primarily related to incremental investments in both the RaSer serial link business and our Yellowstone program. Litigation spending is always difficult to predict, because we do not control the time lines and requests from the courts, nor do we control the actions that our adversaries may take, which would cause us to incur more expense. Based on what we know today, we believe it's likely that litigation spending will again be below $3 million. This is very dependent on a schedule established by the FTC's Administrative Law Judge, which we won't know until the second half of the quarter.

Finally, we're expecting at our tax rate to remain at 35% for this year, and if you add up all of this guidance, the result is earnings per share of about 5 cents for the September quarter. I want to also offer you some guidance on our new technology, on revenue for our new technology areas, the RaSer serial link area and Yellowstone. If we are successful in executing our plans in these areas, including signing new contracts, we think you can reasonably expect that in the next 12 months, these two businesses would account for over 10% of our revenue. We have a lot of work to do to accomplish this goal, but we think this is a reasonable assumption at this stage.

Now, on the subject of litigation. We will quickly review the major litigation items -- the major items in our litigation time line. Let me start with an update in the U.S.. Obviously, we were disappointed by the FTC's decision to file an administrative complaint against us on June 19th, which contains allegations that arise from the very same facts and issues that are currently on appeal in the Court of Appeals for the Federal Circuit. It is important to note that if the FTC were successful, the relief would only relate to products sold in the United States. There really is no significant update on the FTC since our interim call on June 20th. We have not had the first meeting with the Administrative Law Judge, and we will be filing our response to the FTC's complaint on or by July 29th. This FTC filing followed what we and others believed was an excellent oral hearing at the Appeals Court on June 3rd. We are now waiting for the opinion to be issued from the court. It is our belief that the opinion is likely to be issued before the end of the year, perhaps even by late summer. The other patent litigation cases in the U.S. Are currently delayed or stayed pending the Federal Circuit decision. The shareholders suit in northern California is progressing on schedule. We have filed our motion to have the case dismissed and there's a hearing planned for September 13th. We mentioned a new suit last quarter from a company named Holiday Matinee, which alleges that DRAM pricing is too high as a result of Rambus royalties. Based on our understanding, we believe that it has no merit given the overall rapid decline in memory prices during the period Rambus was conducting its SDRAM and DDR licensing program. In addition, not all of the memory manufacturers paid Rambus royalties and many were selling products below costs.

In Europe, the major activity is in the Mannheim court in the case against [INUADIBLE]. The experts opinion was delievered to the court, and both parties in mid May. It is our view that the report is favorable to Rambus. Interestingly, in the midst of the United States' Federal Circuit Appellate Argument, [INAUDIBLE] counsel stated that the German report is favorable to them. Surprisingly, this comment was admitted by [INFINIANS] lawyer, Ken Starr, to be irrelevant to the issues before the Federal Circuit. It is also surprising, given that shortly thereafter, they asked the Mannheim court for a two-month extension in the deadline to file briefs on the opinion. The request for an extension was granted and the deadline for the briefs is now August 26th.

The other important event in Europe is the validity hearing with the European patent office. The hearing is now set for September 10th and 11th. The EPO has issued a preliminary nonbinding report that is very favorable to Rambus. In fact, the EPO's preliminary report indicates that 16 of the 17 claims in the European patent are valid and enforceable. The only claim that had issues is, in our opinion, not very significant given the claims preliminariarily found to be valid. It is our best guess that the next oral hearing on infringement will be some time after the validity hearing.

That covers the key financial and legal issues for the third quarter of fiscal 2002, so now I'll turn it over to Geoff to discuss the current state of Rambus business.

Thanks, Bob. Today, I'm going to discuss the significant progress that we've been able to publicly demonstrate in the last quarter. As you know, we've been working for the past couple of years to diversify and increase our technology offerings. In october, at the U.S. Rambus developers forum last year, we laid out our technology road maps in three areas, and we're excited that we can now publicly demonstrate the beginnings of our delivery on these road maps. Last week was a big one for us, because in Japan, we held our annual Rambus developers forum in Tokyo. We were able to meet with over 900 current and potential customers who attended our forum, and we received very positive feedback on all of our innovations and developments.

I'll start with the RaSer serial link business first. To refresh your memory, the RaSer serial link technology is a family of analogue macro cells that can be integrated into A-6 and ASSP's to implement the physical layer of a wide range of communications and computer IO's. Today, most serial links use stand-alone chips, but the trend we see is towards integration, moving these links onto highly integrated logic devices which ties well into Rambus' business model of providing well-engineered circuits and systems support. Using the Rambus RaSer cell for integration allows system designers to save internal resources money, board space, and power consumption, plus speed time to market. The one gigahertz, and three gigabyte per second family, our first family, addresses wide area networks which are router backplanes. One gig and ten gig ether-net fiber channel, infina band, PCI express and serial ATA applications. In total, we expect the available market for these serial limps to reach over a billion links within a few years. Now, a single integrated circuit may integrate 1, 4, 8, or even more of these links. In an extreme example, crossbar switches, such as are found in the backplanes of switches and rueters may use well over 100 links a piece. The one gig, the three gig RaSer cells are application proven and available now for licensing. These cells lead the industry in their low power and small area. We've executed to date, ten RaSer cell licenses with the expectation of our first royalties on production shipments this year. Our publicly announced so far this year are Intel Microelectronics,Bander Com Internet Machines, and Trans Switch, having announced they've used our RaSer technology. Bander Com has publicly demonstrated its RaSer-based channel adapter and will soon announce a major semiconductor supplier who is licensed RaSer technology for integration into chips for the ethernet market. RaSer based chips are undergoing production qualification at a major OEM now. We have partners who have made the transition from design to product with four partners having sampled RaSer based working chips to their systems customers, many of them. We continue to demonstrate progress towards our serial link technology road map as well.

In Japan, two weeks ago, we announced the industry's first licenseable 6.4 gigabyte per second integratable serial link called RaSer B. It's development is a grounds up third-generation design that meets the demanding needs of the harsh backplane environment. In Japan, we publicly demonstrated RaSer B serial links, driving at 6.4 gigabytes per second over industy standard at [INAUDIBLE] backplane. This was viewed by hundreds of people who attended our conference, and we received feedback that we are far ahead of competition in this area. Jeremy Donovan, Vice President and Chief Analyst at Gardner Data Quest stated that he expects backplane market to increase substantially over the next three years. Factors that are driving the enterprise switch market include increasing bandwidth demands, introduction fo new storage and server blades, and traction of gigabyte ethernet technology in the desk top market. We expect to be leaders in delivering five and then 10 gigabytes per second serial link technology. We're executing it on the road map we announced last October. RaSer B links cells running at 6.4 gigabytes per second is the first step, and next step will be revolutionary and significant advancement with a 10-plus gigabyte per second serial link that we expect by the end of this year. Our solution is based on what we believe is well differentiated and protected intellectual property.

Now, the second exciting area of innovation that I want to discuss is our ultra fast, next-generation memory interface technology, code named Yellowstone. We first unveiled this technology last October, and since have had two major break throughs in this area. The first that we publicly demonstrated in Japan last week, is our interface signaling technology running at 3.2 gigahertz data rates. Yellowstone offer as quantum leap in performance in memory signaling, while lowering system costs through pin count reduction and the elimination of external termination resistors. To date, we've not seen any comparable competitionfor this technology, and we're proud to say that ours is real and it works. The second breakthrough is our Flexphase Technology that was also announced earlier this week. This is an important part of the unique intellectual property that makes the Yellowstone solution possible. Flexphase allows the chip interface to automatically correct the timing differences that are experienced in the real life application of very high-speed chip connections and represents a major departure from traditional circuit technologies. Yellowstone's signaling is initially a 3.2 gigahertz technology, but as we said in October, we have plans to scale it up over time to 6.4 gigahertz. We expect to have our first Yellowstone licensees signed by the end of this year. Yellowstone is targeted at consumer electronics, PC graphics and networking markets where the demand for high-performance from a small number of DRAM's is greatest. Now, I'll make some specific comments about our third business area, RDRAM.

RDRAM memory devices continued to gain design winnings in the consumer, networking, and computing system areas. In the consumer segment, the Sony Playstation 2 continues to sell well against all competitors. We also recently announced that the Texas instruments digital light processing, or DLP chip, is using RDRAM to deliver 1.6 gigabyte per second of bandwidth with one memory chip. This replaces four memory chips in TI's previous generation. RDRAM memory now is a key enabling factor for lightweight, high-quality business projectors using TI's chip product. In fact,RDRAM Bandwidth combined with TI's technology also provides such high video quality that the same chip solution will also be shipping in digital TV's by the end of the year.

In the networking segment, it is now publicly known that network products from companies likeSisco and Juniper are shipping with RDRAM memory solutions. We expect a number of networking systems that use RDRAM to increase over time. Intel disclosed that the Intel developer's forum in February that RDRAM memory has been selected for their next generation network processor. Intel joins with the test, meaning two of the largest network processor will connect to RDRAM memory.

In the computing segment, RDRAM memory continues to ship in PC's, servers and workstations. Intel's new chip set, the 850E began to ship this quarter. This chip set is the highest-performance chip set shipping out for the Pentium IV. The benchmarks of the RDRAM with the 850E have been superb. On June 3rd, at Computecks in Taiwan, silicon-integrated systems, or SIS, showed a static demo of their dual channels, 658 set, also uses RDRAM, and provides a full set of performance features such as USB 2.0 and ADP 8x.

In addition, we expect that HP will be shipping the next generation of alpha microprocessors that use RDRAM memory by the end of the year. The announcement of 2.5 gigahertz Intel Pentium IV processors last quarter is the important step to faster and higher performing processors. The faster 533 megahertz is now shipping like the Pentium IV to take greater advantage of the memory bandwidth that the 800 and now 1066 megahertz RDRAM devices have to offer. Independent third party benchmarking experts, such as PC Magazine and Tom's Hardware Guide, show that the performance advantage of RDRAM memory, in conjunction with chips such as the 850E, over DDR, increases as processors become faster. We're excited, therefore, about the continuing high-performance road map for Pentium IV processors since we think it increases the value of our solutions over time. RDRAM memory is well positioned as a safe and prudent technology to deliver superior performance today and into the future for performance desk top PC's. Now we recognize that the performance isn't the only feature that matters. The price of RDRAM memory versus DDR is also very important. For any purchaser, the system level price difference to upgrade to RDRAM today is not that significant. Now, prices fluctuate, but taking a recent example of the Dell Home Systems catalog, we see that the price for upgrading from 128 to 256 meg bytes of memory in their catalog is $70 for DDR and $80 for RDRAM. $10 is not much additional cost for the substantial increase in performance associated with the RDRAM and 850 chip sets. This is especially true in the context tha RDRAM memory delivers a higher percentage performance increase than the corresponding increase in system costs. RDRAM continues to be most cost effective means to provide improved system performance. Innovation in RDRAM continues with our team making excellent progress in the RDRAM road map that we announced last year. The 32-bit RIM 4200 shipped during the last quarter and has received excellent reviews. It's great that Azeustech, one of the largest motherboard companies, is already shipping a RIM 4200 motherboard. It offers dual channe performance, with single-channel convienence. The RDRAM road map is real and offers evolutionary performance, scales over time, enough to keep up with even the aggressive road map expected and announced for the Pentium IV. The price performance advantage of RDRAM over competitive offerings at both the device and system level have never been more evident than they are today. If you look at recent issues of PC Magazine and PC World, 800 and 1066 megahertz RDRAM systems are winning top rankings for both business and home systems. We expect RDRAM to continuing to play a leading role in the performance desktop space.

In addition to developing great technology, we're working hard to get the word out. The Rambus developers forum in Japan was a huge success, as I mentioned with record attendance, and we're planning similar Rambus developer forums in Taipei, Taiwan, July 31st, and San Jose, California, October 28th and 29th. These venues will provide an excellent opportunity for our partners, as well as Rambus, to get the message out on the key advantages that we offer as the leading high-speed chip-to-chip company. We also plan to have an analysts meeting at the beginning of the developers forum in San Jose that we will webcast to all those who are interested. We'll be providing more details on the San Jose event in coming weeks.

That's the end of the prepared remarks. We'll now open the lines to questions.

At this time, we're ready to session. Begin the question and answer if would you like to ask a question, please press star 1 on your touch tone phone. You will be announced prior to asking your question. To withdraw your question, you may press star 2. Once again, if you like to ask a question, please press star 1 now. Our first question comes from Mr. Michael Crawford. Please state your company name and you may ask your question.

This is Ian [INAUDIBLE] for Mike Crawford at B. Riley and Company. A couple of questions, with DRAM croping up in products from Juniper Routers to Disco switches, HDTV products, could you break down RDRAM units by segment, say,PC, networking, communication, gaming, and then other?

Yeah, hi, Ian. This is Bob. Actually, the way we get our royalty reports in, it's not a question that we can -- it's not a question we can answer, because we'll get a report in from a DRAM manufacturer, but we don't know where they actually have sold the DRAM.

That's a good point. However, from the data that we can get in our market research, you know, we believe, as we've said for some time, that the largest unit consumption of our DRAM's is in PC's. The second largest unit consumption is game consoles, and the other applications in aggregate would be in third place. But we definitely have been seeing a lot of new systems in that third category and see that it should be growing well in the future.

Okay. And the Flexphase technology. Is that applicable to any other parts of the business besides Yellowstone, the serializer products?

No, not in the serializer products. The Flexphase is useful where you have wide parallel buses, because it greatly simplifies the layout, the board area required. In our serial link business, by the nature, it's one or two links at a time, and so it's not an applicable technology. But it -- Flexphase could have value for us and other product offering in the future.

Okay. And finally, why what do you think we're hearing from our sources in Samsung that they're increasing RDRAM production?

I'm sorry, did you say why?

Yes.

Well, since you talked with the sources, I couldn't comment directly. But we think that if Samsung is increasing their production, it wouldn't be inconsistent with the fact that the RDRAM-based PC's today are receiving, you know, popular acclaim and interest. So we're hopeful it's because, if it's actually happening, that consumers are reading the magazines and buying the best systems available.

Okay, thanks, guys.

Thank, Ian.

Our next question comes from John Gruff. You may ask your question. Pleat state your company name.

It's Mark Edlestone, and I had a couple of questions if I could, the first one just related to your comments on Yellowstone, and, Geoff, you said you expected the first licensee by year end. Was that fiscal year end or calendar year end?

Calendar.

And then, could you give us a general sense as to what we might see the recognition like on the license revenues for that, and when would you expect royalties for Yellowstone product to begin.

The royalty question would be two to three years for significant production royalties. Yellowstone, you could look at the examples of the kinds of products we've done in the past with RDRAM. They take some time for the conception, some time for the implementation and some time for the associated software and other manufacturing things, and so, two to three-year time frame, kind of like major RDRAM design wins. Bob, maybe can you comment on the revenue recognition, license fees?

Yeah, of course, the final revenue recognition will be a function of what the contract is. I think, again, a reasonable gauge might be the rdram contracts that we did in the past, and in that case, we were recognizing the contract revenue over a period of three years.

Okay, great. And then, when the RaSer royalties begin, you said production here is going to start for some of the licensees. Can you give us some general gauge as to what kind of royalty stream that ought to be able to generate here over the next several quarters?

Well, I gave a general indication on the combination of yellowstone and raser, and, of course, it was for all revenue. In the next 12 months, we're assuming that that's going to ultimately generate over 10% of our revenue.

Okay, I'm guessing, Bob, that the Yellowstone would be a fairly large percentage, though, right?

It's actually split. You know, if you believe our plan, which is, you know, our crystal ball is only so good, but it's split fairly evenly between the two.

I think our hesitancy here on the RaSer is that until the systems actually go into production, there's certainly chances of slippages, but also until they go into production, we can't really have a very good level of confidence as to, you know, how much -- how many units will ship. So once things actually start into production, we'll be able to, you know, be a little more definitive, I think, as to the kind of run rates we're seeing.

Okay. And your comment, Bob, on the revenues from these activities was -- would be more than 10% of revenue?

Right. A little bit more.

Okay, great. Just a final question. I know this one's tough to answer. But you gave us some guidance for legal expenses here in the fourth quarter. As you have to deal with some of the upcoming events, as you go into fiscal 2003, you have a general expectation as to what the trend might look like there going through the year?

You know, we've been hesitant to try and forecast out beyond one quarter. I'm comfortable saying that our internal plan is the litigation will increase a bit from where it's been lately. Given the fact that once we're through the -- once we get the opinion back on the appeal, we may be back in court in Virginia, and we may be starting up the other trials that we have going on here in the U.S..

I guess, is it possible, Bob, that we get back to the kind of run rate that we would have had a year ago?

I don't -- you know, anything's possible, I suppose. But I'd be surprised if we hit the kind of peak spending we hit last year.

Okay, thanks a lot, guys.

Thanks, Mark.

Our next question comes from Chang Chu, you may ask your question.

Hi my name is Chang Chu with [FLORON] Technology Research

Hi.

Yes, nice results. Listen, I want to understand the full litigation. The difference is only to the royalties for the DDR and also for the SDRAM product, right?

I didn't --

-- not any of your, say, RDRAM royalties, right? I mean, what would be --

Okay.

What would be the impact?

Let me restate the question. I think what you're asking is what impact is the litigation having on our royalties?

Right, on which stream of the royalties.

Right. Okay. Great question. One of the things we want to be real clear on is that the litigation does not have anything to do with the RDRAM business, nor does it have anything to do with RaSer serial link business or Yellowstone. So the litigation is only related to the SDRAM and DDR royalties that wer are receiving from the 8 companies that we currently have lisence agreements for.

So if we kind of imagine going forward, the out of standard is getting phased out. And if that is the case then it is going to get dragged down and become less and less important going forward. Right? I mean, any royalty, your customer or to pay to you, you will not pay back, right?

Well, yeah, our understanding is that given that the FTC filed an administrative complaint that they would not be going backwards and trying to, called discourage profits, the answer to your question is, yes, that's true, they wouldn't go back.

Mm-hmm. Okay, thanks.

Sure.

Once again, if question, please press star 1 at this time. You will be announced prior to asking your question. Our next question comes from Mr. Mike[INAUDIBLE] . Please state your company name. You may ask your question.

Solomon Smith Barney. How you guys doing today?

Good. Thank you.

Real quick question. Can you kind of characterize, even with the ongoing litigation, some of your relationships with the other DRAM manufacturers, and can you also kind of discuss the length of the contracts and what the duration as far as the RDRAM licensing goes?

The RDRAM contracts are for the life of the patents involved with our DRAM. We have license agreements with a number of DRAM companies, some are shipping, some are not, and some are shipping in low quantities. We have very good relationships with the companies that are in production in RDRAM, because they're enjoying a very good business, and it puts them in a differentiated position, shipping a product that's associated with the highest performancePC's and game consoles.

Just a follow-up. Do theRDRAM licensing and contracts you have with those companies that are not currently producing, do they have an expiration date of any kind?

No, there isn't a date of expiration.

Okay.

On any of the rdram license contracts.

Okay, thank you.

Use up our time, Bob? Yeah, do we have any more questions?

Operator: I'm currently showing no further questions.

We'd like to thank everybody for joining us today. And we'll be talking to you again next quarter. Thank you.