新思科技 (SNPS) 2004 Q1 法說會逐字稿

Ladies and gentlemen, thank you for standing by, and welcome to the Synopsys first-quarter fiscal-year 2004 earnings conference call. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session, and instructions will be given at that time. (OPERATOR INSTRUCTIONS). As a reminder, today's conference is being recorded.

During the course of this conference call, Synopsys may make predictions, estimates and other forward-looking statements regarding the Company. While these statements represent the best current judgment about the Company's future performance, the Company's actual performance is subject to significant risks and uncertainties that could cause actual results to differ materially from the those that may be projected. In addition to any risks that may be highlighted during this conference call, important factors that could cause the Company's actual results to differ materially from those that may be projected in this conference call are described in Synopsys's annual report on Form 10-K, which is on file with the Securities and Exchange Commission.

In addition, during the course of this conference call, Synopsys and Monolithic System Technology, known as MoSys, will make forward-looking statements regarding the benefits of the proposed acquisition of MoSys by Synopsys. While these statements represent each Company's best current judgment about the expected benefits of the acquisition, actual performance is subject to uncertainties that could cause actual results to differ materially from those that may be projected, including the risks described in Synopsys's most recent annual report on Form 10-K and MoSys's quarterly report on Form 10-Q, which is on file with the Securities and Exchange Commission.

Finally, Synopsys would like to advise you that financial and other statistical information to be discussed on this conference call, as well as the reconciliation of certain non-GAAP financial measures discussed on this call to GAAP financial measures, is currently available on the Company's Website. The Web address for such information is http://www.Synopsys.com/corporate/invest/invest.html.

At this time, I would like to turn the conference over to Aart de Gues, Chairman and Chief Executive Officer. Please go ahead, sir.

Thank you, operator. This is Aart de Gues, and I have with me Steve Shevick, our CFO, and Fu-Chieh Hsu, Chairman and CEO of Monolithic System Technology, or MoSys for short. Thank you all for joining us.

Synopsys entered 2004 with two key long-term objectives -- to gain share in the core EDA market and to accelerate our business penetration in the emerging high-growth areas of IP and design-for-manufacturing. In Q1, we made excellent progress in our core tools, and today I have some very important announcements to share with you in the IP area.

But first, I will provide a brief overview of our results, then give you my thoughts on the current spending environment and share what I'm hearing from our top customer executives. After that, I will tell you about our announcements and give you updates on our core products.

So let me start by reviewing our financial results. In Q1, Synopsys achieved revenue of 285.3 million and pro forma earnings per share of 33 cents. Book to bill was between 0.5 and 0.6. Orders were within our range of 10 to 15 percent of our original annual orders guidance of 1.425 to 1.475 billion, largely reflecting typical Q1 seasonality.

Looking around us, consumer demand for electronic products worldwide is very healthy, and semiconductor production is nearing capacity. The semiconductor outlook remains strong, and I would not be surprised to see semiconductor revenues grow in excess of 20 percent in 2004.

Cost controls remain tight, but a number of execs have told me that their design plans will require an increase in spending in the second half of the year. Reflecting on my interaction with top customer executives over the last 18 months, I see a substantial change in their expectations from us. Instead of looking at us as just an EDA supplier, and perhaps having a dialogue on tool utilization features or bugs, today, CEOs and COOs are calling us with a complex set of economic and technical vision questions.

Coming out of the most massive semiconductor downturn on record, they are worried about three dimensions. First, their ability to differentiate in terms of chip speed, power consumption and functionality. Second, their total cost in terms of design complexity, test and manufacturing yields. And third, their schedule predictability, in terms of verification risks and IP reuse. Top-down, customers' decisions are driven primarily by economics, but bottom-up, they will move forward only if they can rely on their partners' technical credibility. This is music to our ears, as we provide the best solutions for both.

We have built Synopsys to become a key player in the broader silicon infrastructure market. In this broader market, concepts such as IP and yield are on par with the more traditional areas of design and verification. All these areas work together to meet our customers' major concerns about differentiation, costs and schedules. IP reuse directly impacts design costs and schedules, while our design-for-manufacturing is clearly aimed at fabrication costs and yields. From a Synopsys growth point of view, these areas are funded from outside the traditional EDA budget. Thus, they expand our potential market significantly.

Speaking of IP, our IP business has done very well in the last 18 months. Driven by demand from our customers for more IP this year, we have decided to increase our investment in this area. I would like to update you on two important acquisitions that we're announcing today -- Accelerant Networks and MoSys.

Let me give you the background. As long ago as the early '90s, Synopsys had a vision of design reuse as one of the fundamental ways to increase design productivity and reduce costs. Over the years, we built a very complete, high-quality collection of building blocks under the name DesignWare. This collection is available on a subscription basis to our customers, and is the most widely-used commercial IP in the world. As design complexity increases, so does the complexity of the building blocks. Standard interfaces such as PCI or USB become more and more sophisticated with each new generation. Customers need to outsource these building blocks so they can focus on their own differentiation, but to ensure that their product gets to market quickly with acceptable quality, they require rock-solid quality levels from their IP providers.

Our own internal verification tools and expertise have turned out to be great assets in achieving that objective. Our ability to provide the right building blocks, with quality they can trust, frees them up to let go of these portions of the SoC and focus on what they do best. As a result, our business has grown rapidly in the last 18 months, and customers increasingly want Synopsys to become their house supplier for the majority of their diverse IP needs. Exploiting this opportunity, we're taking big steps to expand our IP portfolio this quarter, by adding more standard space IP and embedded memory IP.

The acquisition of Accelerant strengthens our existing position in what is called connectivity IP. These are the devices that allow computers and popular consumer products to talk to each other. You may be familiar with some of the names, such as the USB or 1394 ports on your computer, but there are many others, such as PCI, serial ATA and so on. Synopsys's DesignWare cores are already the market-leading solution for productivity IP cores which get data in and out of SoC. To expand our IP into emerging connectivity standards further, we have acquired Accelerant Networks. Accelerant provides highly efficient, low-power, high-speed analog interface technology called serializer/deserializers, or SerDes for short. The SerDes technology will be the foundation for high-speed analog interfaces for such new connectivity standards as PCI Express and serial ATA. Accelerant brings a very talented set of designers and some excellent technology to Synopsys, just as the connectivity blocks are moving to the next level of speed and sophistication. In that sense, Accelerant will accelerate our IP connectivity solutions, and should especially benefit such markets as consumer products, storage and network infrastructure.

The major new expansion of our IP portfolio is in embedded memory IP. We're very excited to announce the acquisition of MoSys, a leader in innovative memory for SoC design. Most people do not realize how much of the area of a modern SoC is taken up by embedded memory. In 2002, it was already about 50 percent of the chip. The international technology roadmap for semiconductors sees this growing to over 70 percent by 2005. Consumer electronics with integrated voice, data and multimedia are driving this trend. This is why it is so timely for Synopsys to be expanding into embedded memory IP right now. As you can imagine, embedded memories have real impact on both the yields and the overall cost of chips. This is where MoSys's key product, the 1T-SRAM, makes all the difference. MoSys's 1T, meaning one-transistor memory, is unmatched in density, power, speed, low cost and high yield, and works with regular digital processes. 1T-SRAM technology offers all the benefits of traditional six-transistor SRAM, such as high speed, simple interface and ease of manufacturability, but there, all comparisons end. First, the 1T SRAM requires 50 to 70 percent less silicon for the same amount of memory. Manufacturers of cellular phones and video game consoles can thus either add more functionality or reduce the cost of their chips. Second, 1T memories consume less than one-quarter of the power for the same speed of operation. Third, for very large memories, the speed is equal to or greater than those of traditional SRAM. And last, it is scalable for current and future process geometries, and has already been ported to a wide variety of foundries including TSMC, UMC and Chartered. In fact, MoSys's licensees have shipped more than 50 million chips, using a wide range of silicon processes and applications.

As you can see, we're systematically advancing our IP portfolio to help customers deal with differentiation, costs and schedule concerns. In Q1, examples of customers that chose Synopsys IP were AMS (ph), Sony and TI.

Let me now say a few words about our more traditional design and verification markets. The move to smaller geometries continues unabated, and my semi equipment colleagues report strong business across all geographies. Interestingly, roughly half of all the new equipment is going toward 130-nm geometries of 200-mm wafers, with the other half going to 90-nm design of 300-mm wafers. If we assume that this equipment sees production usage in six to nine months, the timeframe jibes well with our tracking of 90-nm design starts, which are steadily growing, and with customer reports of increased success on advanced chips.

Speaking of advanced chips, as I mentioned earlier, our interaction with customers now about the impact of our tools on differentiation costs and schedules. In that context, I'm pleased to highlight our new agreement with ATI, which we announced in Q1. ATI designs very advanced, high-volume graphics cores, and our agreement provides them with long-term access to Synopsys's design and verification technologies. We expect to work closely with ATI on their most advanced chips. In Q1, we made excellent technical progress in both design and verification. Starting with verification, which consumes 60 to 70 percent of overall chip design times, we moved the bar up significantly. In Q1, we released a 10x faster version of VERA release 6.2, which is getting superior results in competitive benchmarks. We also released version 7.1 of VCS, which extends our lead in the built-in test bench capability, resulting in up to 5x faster overall verification speed. System Verilog customer momentum keeps growing, with over 1,500 engineers attending an industry-sponsored seminar series during the last four months.

Building on this momentum, we expanded our relationship with ARM to jointly develop a system Verilog reference methodology, which will be showcased at the design automation conference in June.

Moving to the intersection of digital and analog, in Q1, Synopsys introduced Discovery AMS, our analog mixed-signal verification platform. Since the launch, we have already won and deployed Discovery AMS at 69 logos worldwide.

A new area that is showing great promise is automatic analog optimization. After carefully evaluating all the opportunities, we decided to acquire the technology leader in this emerging field, Analog Design Automation, or ADA for short. ADA provides brand-new, state-of-the-art optimization technology that can immediately build on our strong analog simulation position. It allows customers to rapidly tune their circuits using our simulators HSPICE and NanoSim.

Now, looking at the design side, as I mentioned earlier, the move to 90-nm and below continues unabated. 90-nm designs are underway at 7 of the top 10 semiconductor companies, and we are already actively engaged in 65-nm design projects at 4 of the top 10. Looking at our own R&D output, I told you last quarter that our engineering teams were really cranking in the Galaxy design platform. This became clearly visible when we shipped a slew of big improvements and enhancements in December. For example, the new version of Design Compiler is 3x faster, and offers 25 percent more capacity than the prior year's release. This boost in speed and capacity makes a big difference to our customers, and is having a positive impact in competitive benchmarks. I'm also pleased to report that Design Compiler now supports System Verilog for design. Stay tuned for more improvements in DC in the coming months. DC Ultra continues to capture the high end of the synthesis market, adding 11 new logos in Q1 and making up 25 percent of all DC bookings in the quarter.

PrimeTime, our flagship timing verifier, and V-Sign (ph), our product for the IC industry, also saw a 3x run-time improvement, and up to 3x capacity improvement. These enhancements, combined with leading signal integrity capabilities, enabled PrimeTime to handle the most advanced, 100-million-gate 90-nm design.

Power has been another key engineering worry for our customers. Whether they want to increase battery life in consumer products such as cellphones, or minimize heat dissipation in high-speed computers and routers, power is an enormous challenge at the smaller geometries. In Q1, we took a major step forward with our December release of Power Compiler, increasing its speed of power optimization by 10x.

Physical Compiler is improving, as well, with greater than 2x run-time improvement, 20 percent increase in capacity and new capabilities for physical power optimization.

Let me not forget to mention the most recent release of Astro. It delivers 2x faster run-time, up to 40 percent capacity increase, and improvements for 90-nm design including yield, which is of very high value to our customers. In fact, we're receiving first indications that Synopsys products are getting better yield results than those of our competitors. Astro added 13 new logos in the quarter.

And while we're on the subject of yield and manufacturing, design-for-manufacturing, or DFM for short, is of extremely high interest to our customers. It has the potential to profoundly impact their economic equations. Here again, we are making excellent progress. For example, we announced a substantial customer win when Toshiba licensed our PSM technology for their 65-nm technology node. This win is significant, as Toshiba will be one of the very first customers to put 65-nm technology into production in the first half of 2005.

To help customers develop even smaller geometries, we released Taurus Process Atomistic. This is TCAD software that mulls (ph) interactions at the atomic level and improves time to yield for 65-nm and below. Its software value has already been endorsed by leading companies, including Toshiba and Matsushita.

Optical proximity correction, or OPC for short, is final for mask-making. Here we have also made great progress. Our product Proteus demonstrated near-linear scalability when distributed on more than 1,000 Intel Xeon processors. As a result, Proteus can reduce the time to obtain OPC results from days to hours on advanced 90-nm and 65-nm chips. Synopsys's leading semiconductor customers have already taken out complex designs containing over 500 million transistors with Proteus, using thousands of compute servers.

In summary, against an economic backdrop that is showing promise for the second half of the year, we have forcefully moved the ball forward in our core technologies, as well as in our new growth areas. Our major business thrust, acquisitions and technical investment strategies are in solid alignment, and we are first in line with customers as their spending picks up. IP and DFM are resonating very well with customers, and seeing sales momentum today. While our incremental investments in these initiatives will have a modest impact on fiscal 2004, they should positively impact orders and profits in 2005.

With that, I will turn it over to Steve Shevick for further details regarding our financials.

Thank you, Aart, and good afternoon. As usual, during this call, I will discuss our financial performance on a pro forma basis. Our pro forma measures exclude amortization of intangible assets, in-process R&D and other merger-related charges, and follow-on charges relating to our workforce realignment implemented at the beginning of fiscal 2004, which was communicated in our Q4 earnings call in December. Reconciliation of pro forma results to GAAP results appears in our press release and financial supplement. For your convenience, a copy of our remarks and all the numbers I discuss will be posted on our Website.

Now, for the numbers. Revenue for the first quarter was 285 million, within our guidance range of 275 to 290 million. Approximately 21 percent of revenue came from up-front licenses, 60 percent from time-based licenses and 19 percent from maintenance and consulting. Revenue increased 6.4 percent from Q1 of FY '03, principally due to growth in time-based license revenue. Time-based license revenue increased by approximately 30 million to 171 million, up-front license revenue increased by approximately 5 million to 59 million, and services revenue decreased by 17 million to 55 million. Sequentially, revenue declined seasonally from Q4 of fiscal year 2003, due to a lower level of up-front licenses. Revenue from up-front licenses dropped by 28 million from Q4 to Q1. Revenue from time-based licenses increased by 3 million, and revenue from services declined by 6 million. One customer accounted for more than 10 percent of revenue for the quarter.

Geographically, Asia-Pacific contributed 13 percent of revenue, its highest contribution ever, reflecting the success of our field investment in that region. North America contributed 56 percent of revenue, Europe contributed 17 percent of revenue and Japan contributed 14 percent of revenue. Revenue from our core EDA products accounted for 72 percent of total revenue, and revenue from our emerging businesses 18 percent, with 7 percent coming from IP and 11 percent from DFM.

From a bookings perspective, as is typically the case in our first quarter, book-to-bill was well below 1, ending between 0.5 and 0.6. A number of sizable orders pushed from Q1 to Q2 on account of unfavorable terms. These orders closed early in Q2, which gets us off to a good start. Notwithstanding the pushed orders, Q1 orders were within our guided range of between 10 and 15 percent of our original orders target for the year. One customer accounted for more than 10 percent of orders in the quarter. Approximately 81 percent of software orders were booked as renewable licenses and 19 percent as perpetual licenses. We expect the percentage of perpetuals to trend downward over the course of the year, and for the year, perpetuals we expect to contribute 10 percent or less of license orders.

During the quarter, up-front licenses represented 57 percent of license orders, including between 5 and 10 million and license orders that will not turn to revenue until after Q1. This was well above our target for the quarter, and reflects both the acceptance by the customer base of term licenses and the high sensitivity of this percentage in a low-orders quarter. For the year, we expect up-front orders to be in a range around 25 percent of total license orders.

The average length of our renewable customer license commitments in Q1 was 2.7 years, roughly 10 percent shorter than the average length of such commitments in fiscal 2003. Aggregate pro forma operating expenses for the first quarter were 205 million, at the bottom of our target range of 205 to 211 million, principally due to lower sales commissions associated with the quarter's level of orders and lower-than-expected litigation and depreciation expenses. Expenses for the quarter also reflected a savings of 3.4 million from the shutdown of our North American facilities between Christmas and New Year's and the headcount reduction, each of which was described on our Q4 earnings call and contemplated in our original guidance.

The decline of the dollar versus the Euro cost us 1 million in expenses versus plan during the quarter, as backlog in Europe is denominated in dollars and our European expenses are primarily in Euros. Beginning this quarter, we are hedging our Euro expenses from beginning to end of quarter, which will give us stability in our current-quarter forecasting of Euro expenses. Over the rest of the year, however, the strength of the Euro could negatively impact earnings by approximately 3 to 4 cents. Our pro forma operating margin for the quarter was 28 percent.

Other income for Q1 was an expense of 1.1 million, at the bottom of our target range, and included 1.4 million in investment write-downs. Our pro forma tax rate was 31 percent, down from 32.5 percent last year. As I indicated last quarter, this reduction was made possible by a refinement of our international subsidiary structure. Weighted average shares outstanding for the quarter was 166 million shares, near the high end of our target range of 159 to 167 million shares. During the quarter, 4.4 million options were exercised, contributing 1.7 million shares to weighted average shares. Our overhang currently stands at 24 percent, approximately 12 percentage points lower than one year ago.

Pro forma earnings was 33 cents per share, 1 cent above the top of our guidance range of 28 to 32 cents per share, and 2 cents above the First Call consensus. Cash and short-term investments were 618 million at the end of Q1, down approximately 80 million from the end of fiscal 2003, largely reflecting the net effect of share repurchases, offset by option-exercise proceeds.

During the quarter, we purchased approximately 4.6 million shares of Synopsys stock, at an average price of 35.13 per share. And between the end of the quarter and February 13th, we repurchased an additional 1.6 million shares under our 10b5-1 plan, at an average price of $35.48 per share. As of today, therefore, approximately 284 million remains in our repurchase program. As a result of the MoSys transaction, we will be out of the market until completion of the transaction.

Cash from operations on a GAAP basis was 11 million, and reflected the payment of FY 2003 bonuses and cash disbursements for taxes of 34 million in Q1 of FY 2004. CapEx was 15 million for the quarter, and expenses and disbursements relating to our workforce realignment were 10 million, leaving pro forma free cash flow of approximately 6 million. Q4 accounts receivable totaled 194 million, down approximately 3 percent from last quarter, and DSO was 62 days, up from 58 days at the end of Q4 -- in both cases, minor movements reflecting solid collections and the impact of improved payment terms in the quarter. Deferred revenue at the end of the quarter was 434 million, up 14 million from the end of Q4. Payment terms improved during the quarter, because of the high proportion of up-front licenses and the fact that, unlike others in the industry, we do not permit extended payment terms on up-front licenses.

Aggregate backlog was seasonally down from the October quarter, and should grow sequentially throughout remainder of the year. For the second quarter, we expect that over 70 percent of our target revenue will come from backlog.

Headcount totaled 4,206 employees at the end of the quarter, down from 4,408 at the end of Q4, reflecting the headcount reduction implemented at the beginning of the quarter. As a result of the MoSys, Accelerant and ADA transactions, and our planned incremental investment in promising growth areas of the business, we expect to exit the year with headcount of approximately 4,500 to 4,550.

Before I provide our targets for Q2 and FY '04, let me first briefly summarize the terms of the transactions announced today. The purchase price of MoSys is $13.50 per share, or 432 million, including the value of assumed options, and approximately 349 million net of cash. The acquisition will be paid half in cash and half in Synopsys common stock, although we have an option to offer all cash at our discretion. The acquisition is structured as an exchange offer, which we expect to launch within two weeks. The transaction is subject to customary closing conditions and receipt of regulatory approvals. The target for closing is the end of April. Accelerant was acquired for 22.5 million in cash, and closed last Friday.

I will now describe our expectations for Q2 and fiscal 2004, which reflect the MoSys and Accelerant transactions, the ADA transaction announced last month, which we expect to close before the end of February, and the acquisition of certain memory test technology from iRoC Technologies.

Our second-quarter targets are as follows -- revenue between 285 million and 300 million. Revenue is impacted by a lower level of targeted up-front licenses and a slightly more conservative view on first-half orders in our traditional EDA business. Total pro forma expenses between 214 million and 221 million, reflecting the ADA, Accelerant and iRoC transactions, appreciation of the Euro and incremental investments. Other income and expense, between an expense of 1 million and income of 2 million. Pro forma tax rate of 31 percent. Outstanding shares between 160 million and 168 million. And pro forma earnings between 31 and 35 cents per share. We expect TBLs to account for 70 percent of license orders and up-front license orders to account for 30 percent of license orders, in each case plus or minus 5 percentage points. And we expect TBLs to contribute 53 to 58 percent of total revenue, up-front licenses to contribute 25 to 30 percent of total revenue and services to contribute 16 to 21 percent of total revenue.

Our targets for the full year are as follows. Orders, excluding acquisitions, will be in excess of 1.4 billion, a modification of our previous range of 1.425 billion to 1.475 billion, reflecting a slightly more conservative view of spending on core EDA tools in the first half of the calendar year, with a quarterly distribution of orders for the rest of the year as follows. 20 to 25 percent of the year's orders in the second quarter, versus our original expectation of 22 to 27 percent, 20 to 25 percent in the third quarter, and 42 to 47 percent in the fourth quarter, versus our original expectation of 40 to 45 percent. We expect revenue between 1.2 billion and 1.25 billion, with a quarterly distribution of 23 to 25 percent of the year's revenue in the second quarter, versus our original expectation of 24 to 29 percent, 24 to 29 percent in the third quarter and 25 to 30 percent in the fourth quarter, versus our original expectation of 24 to 29 percent.

With the impacts of investment in M&A and R&D, we expect pro forma operating margin between 27 and 29 percent versus our original expectation of 28 to 30 percent. Pro forma tax rate of 31 percent and pro forma earnings per share between $1.30 and $1.40 per share, versus 1.50 to 1.60 in our original guidance. The change in our guidance range for earnings reflect adjustments of 3 to 4 cents for the impact of exchange rates, 3 to 4 cents for incremental investments in selected businesses, and 10 to 12 cents relating to the ADA, Accelerant, iRoC and MoSys transactions, including both the expense impact and the share count impact. Pro forma free cash flow will remain in excess of $1.80 per share, assuming payment terms and collections consistent with the fourth quarter of 2003 and our orders outlook. Higher cash expenses versus plan, due to M&A and R&D investments. Lower net tax impact, due to more efficient use of tax NOLs and credits, along with the tax benefits related to higher option exercises, and CapEx of 40 to 45 million, an increase from our original estimate of 30 million for the year, with a large piece of the increase attributable to integration of acquired companies. We expect PBLs to account for 75 percent of license orders, and up-front licenses to account for 25 percent of license orders, in each case plus or minus five percentage points. We expect PBLs to contribute 54 to 59 percent of total revenue versus our original expectation of 57 to 62 percent, up-front license revenue to contribute 23 to 28 percent of total revenue versus our original expectation of 21 to 26, and services to contribute 15 to 20 percent of total revenue.

Although we are maintaining our original comments on fiscal 2005, we believe that our incremental investments and our growth initiatives should have a positive impact on our results. As a reminder, we said to expect fiscal 2005 revenue of approximately 1.4 billion, and pro forma earnings over $1.90 per share.

With that, we will now open for questions.

Garo Toomajanian, RBC Capital.

The first question, I guess, that comes to mind is why are you paying such a big premium for MoSys?

Well, Garo, if you look at IP valuations, they are obviously quite rich, in general. And frankly speaking, I think some of that hopefully will transfer to Synopsys, as we clearly show that we're building a very strong IP position. The most important aspect, though, is that the value of the IP is very high for us. We look at this acquisition in terms of a strong return going forward, also a top line growth and, most importantly, adding these pieces will really allow us to come out of the box as the house supplier for IP for many of our customers, and we think that that is a great opportunity.

How does the acquisition of MoSys technology impact your relationship with other IP vendors like Artisan and Virage? Do you start running into more of a competitive situation?

Well, some of those are obviously very tight in the library domain, and are strong partners of ours, and we expect that to continue. Maybe one of the most important IP vendors is really ARM, where we have had already very positive feedback. And so what we are looking at really is providing all the major blocks that are semi-differentiated, where our customers can use us as the provider of a complete solution.

Can you go over again what the terms were of the Accelerant acquisition?

22.5 million, all cash, closed last Friday.

Steve, did you say that 57 percent of orders were up-front in Q1?

57 percent of product orders in Q1 were up-front, owing to a couple of things -- acceptance of the term license model and the relatively low level of overall orders for the quarter, which makes the percentage very sensitive.

Was that one 10-percent deal an up-front order?

It was.

And probably a question for Aart -- actually, maybe, and/or Steve. It looks like 2004 is shaping up to be a pretty back-end weighted year for most EDA vendors. And everyone seems to be talking about a lot of renewal activity slated for the end of the year. Do you think that could result, I guess, in heightened devaluation activity? And does that hurt or help you? And also, if you could answer specifically relative to Magma, what is your specific sales strategy there, when you see them knocking on doors of your customers?

Well, in general, the timeframe really helps us because, now that we have the integration of Avanti behind us, you may probably have gotten the gist from my scripts that we are really cranking on the R&D side. And so, technology progress is rolling out at a very rapid speed. And so, in general, the customers that work with us have been working with us for a long time, and therefore know well how to bank on our technology progress going forward. And so I actually look forward to a strong second half.

Okay. But is there anything, maybe specifically roadmap-wise, that you can tell your customers that might dissuade them from evaluating Magma?

Well, there are a number of capabilities. First, it is very important that on the sophisticated chips that they come back from fabrication actually working, and working well. And I alluded to the fact that our emphasis on yield is starting to pay off handsomely. And so I expect that going forward, we will gradually be able to put some cost numbers on that that will show very quickly that there is enormous value in using our tools. And I think, at this point, then, you should stay tuned for our announcements coming up about that.

Raj Seth, S.G. Cowen.

A couple of questions. Steve, just so I am clear, you previously had guided revenue at 1.2 to 1.25. You are making a couple of acquisitions here, and if I'm not mistaken, I think people were assuming that MoSys would do in the low $30 million range of revenues. Is the delta between what we would have expected, based on your guidance last quarter, and what we are now expecting -- is that simply because you're looking for a more back-end loaded year, in the way that revenue flows through, or is there something else going on? I guess I'm curious why revenues, relative to your previous guidance, given these acquisitions, are not higher.

Yes. I think you could look at it that the impact of revenue on the year is within the guidance range, number one. Number two, MoSys we expect to close late April or sometime in May, so we'll only have it, really, for two quarters of the year. The others are really technology acquisitions which, as they ramp up, in our license model, produce orders but not much revenue. I think we would be looking for a stronger orders contribution than we would for a revenue contribution.

Do you have to change -- I don't know MoSys's business model well, but do you change the way revenue is recognized there, and move it to more of a subscription model, if it wasn't --?

Well, the MoSys model is a license fee that gets recognized on a percent-of-completion method, which generally spreads over a year with some up-front and then the rest in some increments over the year, and then a royalty that may not kick in for another 12 to 18 months after the chip goes into production. We probably won't change that, but the revenue does stretch over the year. The others will translate probably into license revenue, which will go into the subscription model.

For Aart, I guess, underneath all this commentary, you have said that you now expect a more back-end loaded year and a somewhat weaker first half. I guess part of it we have already seen in Q1. What is happening? What is driving the change between what you expected a quarter ago and now, especially given all the positive things you said about what is happening in your customer base, et cetera?

I think that many of the customers remain very controlling on their expenditures, partially because they have been burned so badly in the last 2.5 years, and partially because they now see the expenditures go up, in terms of manufacturing equipment. And so, in many ways, they are sort of very happy on one hand, at the same time still a little careful in projecting what is going to happen going forward. And what I hear from them is that they are looking at increasing their spending in the design environment really in the second half, assuming that things continue on the actual manufacturing side. And so far, that's looking pretty good.

So is your expectation of a pickup in the second half driven primarily off large enterprise renewals or your expectation, given these conversations, that people are just going to start spending again?

It is actually primarily the second, although we do have some renewals coming up, as well, in the second half. But my earlier answer was really strictly the second, meaning that we do see the market becoming a little bit more solid, based on continued growth.

Last quick question was -- the MoSys transaction, were there other bidders in there?

Well, we normally don't comment about the transactions. By the time we file all the papers, you can read the gory details.

Sumit Dhanda, Banc of America Securities.

The first question I have is sort of a follow-up on what Raj just asked you about the fact that you are expecting now a pickup from your customer base in the second half of the year. My understanding at the analyst day was that this was more -- the big second half was more driven by the timing of the renewals, more than anything else. What has changed, in terms of the timing of the renewals, if you could just expand on that?

Sure. Nothing has changed in terms of the timing of the renewals. Those are still the same. I think it is all the sort of -- the smaller deals, the deals around it, where I think our better way way of looking at it is by virtue of sensing the market. For the large deals, we know pretty well exactly when they come up, and what our plans are, so that has not changed at all. So it is really more an environmental description.

The other question I had was, is this the first tangible sign that you're seeing of a big pickup, or at least a better attitude from your customers in terms of spending plans? Because at this point, we're sort of at least midway through the recovery in semiconductors, and I am just trying to get a sense of is this the first sign you're seeing of something good? Or is this grounded more in hope than reality?

Well, I think I am on record of having forecasted that semiconductors in general would pick up already about six quarters ago. And indeed, we see about six quarters of gradual return. And it's very gradual, steady. And now I think the return is accelerating a little bit, largely because of the capacity utilization being arrived at. Now, that drives up the pricing; that drives up also the positive feeling of the customers. The other thing I hear from the customers is not only pricing, it's actually the volume going up, as well. And so I would think that at this point in time, many customers are looking at how quickly can they come out with the next version of their chips, the next derivatives? And that invariably solely (ph) starts bringing back in the concept of schedules. And once people are worried about their schedules, invariably they are willing to spend some money if it helps.

Any commentary on the improvement in design starts, hiring, which -- of engineers, which often seems to be a key leading indicator of EDA demand?

You know, we hear quite a bit of anecdotal information about that. We asked ourselves the question, in preparation for this call. So what do we really know in terms of numbers? And it is all anecdotes, and so that's not really solid. I do see, though, that there is a marked pickup in demand for services, that invariably is also a better reflection of people accelerating the design or looking at adding more manpower. In that context, I think we are well positioned, because our service is very much specialized on special skills, and those are right now in high demand on all of the advanced chips.

Steve, a couple of housekeeping questions for you. What did you say the average length of the contracts were? And if you could break out the impact of EPS that you had indicated, due to additional line items?

Yes, the average length was 2.7 years, and the impact on EPS I gave was 3 to 4 cents for FX, 3 to 4 cents for incremental investments in existing businesses and the high-growth businesses, and then 10 to 12 cents for the acquisitions.

And then what were the orders that were rolled over from last quarter to this quarter? What was the dollar amount?

It was about 15.

15 million.

And from this quarter to next quarter is between 5 and 10.

Okay, and from this quarter to next quarter is -- and then the last question. What is the breakout again, in terms of different products, services and gradable (ph) for this quarter?

It's on the supplement that should be on the Website, or I can give it to you off-line. It's in the script and, rather than take a minute, why don't we just get it to you later.

Sounds good.

Bill Frerichs, D.A. Davidson & Co.

Steve, you mentioned -- and I think this is the same business that you just mentioned -- but you said the deal slipped because of unfavorable terms. I take it that your price was not acceptable, and it slipped over a quarter boundary, and the customer came in anyway? Is that what you're trying to tell us?

Yes. I think that there were things we could have done to bring those deals in in Q1, which we declined to do. And then, instead of caving or giving up some things, they came in in Q2 on the deal that was on the table.

Was it mostly pricing or payment terms, or some combination thereof?

It's a variety of things.

But you are hanging tough on pricing, versus perhaps discounting by competitors? Is that the picture you're trying to draw?

We are hanging tough on pricing.

Jay Vleeschhouwer, Merrill Lynch.

A few questions for you, Steve, around bookings. First, in terms of 2004, let's talk first about the composition of bookings by product area. Can you say whether Astro bookings are likely to be up or down in the aggregate for the year versus 2003? You have previously acknowledged some rough percentage of 2003 bookings that came from Astro, which gives us a dollar amount. So would you expect that number to be higher or lower in 2004?

Jay, we don't tend to give out product line bookings forecasts. We think overall, in terms of Galaxy bookings or implementation bookings, which we do expect to grow from '03 to '04.

When the renewals come up at the end of the year, or two of the largest ones in Q4, what are you thinking about -- or the likelihood that customers will expand capacity and/or implied revenue per user upon the renewals? Or will the largest users that are coming up later in the year, supposedly Intel and others, that use DC, PC, PrimeTime and so forth -- will they necessarily increase their capacity or the variety of products that they include in the contracts?

Well, Jay, for the last 15 years, the answer to that question has been yes, and I expect it to continue to be yes.

Yes to capacity and to value?

Correct.

Steve, in your comments about the MoSys acquisition, you mentioned that part of your business model entails royalties. Numerical had tried that, and you said you were going to continue that, subsequent to the Numerical acquisition, as not being a viable model. And Aart, last week at DATE, you were quite firm in saying that customers typically do not accept vendor royalty models. So isn't it likely that you'll end up having to dispense the royalty part of the MoSys model?

You know, Jay, Numerical had been all over the map in what they charged for a license, and MoSys has been extremely consistent in getting the royalty in every license agreement they have entered into. Also, the value proposition offered by the MoSys technology is so tangible, and so close to the wafer, that it is an area where I think we can continue to get some royalties out of it. So we don't really intend to change the MoSys policy, at all.

It's a somewhat subjective question, perhaps, but at DAC, this past year, you talked about customers being able to increasingly see the value of the integration of the Galaxy platform. Is that something you are, in fact, able to tangibly measure or discuss with us, that customers are in fact paying you for this perceived value of what you're doing on the integration level with Galaxy?

So the answer to that is absolutely yes. As you can imagine, bringing together Avanti and Synopsys was a fairly major task, because many of the products initially we developed fully independently. We have made just fantastic progress over the last year in having a very, very strong alignment on all the key things that matter, such as the timing model, the constraints, the libraries and so on. So that is of great, great comfort to the customers, and we have actually gotten explicit positive feedback about these things. And so, from that perspective, I think we have now moved into the next phase, which is really to create new capabilities, new value. And the integration has already demonstrated much of its value.

Aart, a couple of last things. The acquisitions -- MoSys, in particular, since that is the largest one today -- driven by just general market trends, or any particular customer influences? Some of your prior acquisition seems to have been -- for example, last week, ST gave a design roadmap at the conference in Paris, and much of what they discussed in terms of interfaces, power management, analog mixed signal and the like sounds very much like your rationale for the MoSys acquisition.

I didn't see the presentation in Paris, but I'm not surprised at all, given that companies such as ST are close to us, in terms of giving us sometimes very strong feedback as to what they want. And so, these acquisitions are absolutely both very much timed by both the market need and strong input from customers. And the specific input from customers that we have heard now over and over again is they like the IP that we have, they like the commitment and the delivery of the quality that we have. And because of that, they would like to see us deliver more of the IP and I use the term be the house brand of IP for them, and hopefully, it's the trusted house brand.

And so, in that context, specifically memories -- we have started to look at this already quite a while ago, and it was really the opportunity of finding MoSys with a brand-new, very strong capability that put us over the top of saying, hey, let's do this, because the timing is absolutely right.

And lastly, for Steve, did you in any way change the low end of the dollar value of your estimated bookings for the second half of this year, for product?

Did we change the low end of the dollar --

Right. In other words --

No, we raised it.

You raised the low end of your bookings forecast for the second half of the year?

Yes.

Rich Valera, Needham & Co.

Steve, could you comment -- you mentioned that there's 3 to 4 cents of incremental investment, I guess, in new areas. Can you talk about what those are, and sort of what has changed since you last gave guidance that caused you guys to want to spend in those areas?

Certainly. Well, the first area, of course, is the very area that we have talked about a lot today, which is IP. There's a lot of value in focusing some efforts there. But most importantly, the investments are in all these areas where there is immediate customer impact over the next 9 to 12 months, let's say. And specifically, we see support opportunities; we see some technology developments that customers have requested that would allow us to harvest some of the new value that we have created. So that's where this is going. It's fairly short-term oriented.

And, Steve, on the 10 to 12 cents from the acquisitions -- can you give us a sense of what the split is between share count and expenses for that 10 to 12 cent hit?

Yes, it's probably one-third share count -- one-third to one-half share count, and then the rest is expenses.

And just to make sure I understand, for the second quarter, the sort of light revenue there -- is that mainly a function of the low time-based bookings in the first quarter, or are you also expecting lower up-front bookings in the second quarter then you might have previously?

Yes, we are taking down the up-front bookings forecast for the second quarter, as well.

(OPERATOR INSTRUCTIONS). Erach Desai, American Technology Research.

Three questions, guys. The first question is, Steve, again, you may question my math. But when you look at Synopsys/Avanti combination, and for the last five years, the fiscal first-quarter revenues overall have been down maybe 1 percent sequentially. You were down 10 percent. So, even in this strong semi environment, can you or Aart allude or reflect on anything else that might be going on, on a macro level?

As to Q1, trends in Q1?

The quarter you just reported, your revenues were down 10 percent sequentially. They may be in guidance, but with respect to the last five years of Company history, which has been more like down 1 percent, I'm trying to understand what might have changed. Or it's just a bigger Company?

You know, Erach, I think in general, there's too much weight on being able to exactly forecast quarter by quarter. The reality is we run our businesses so much more year by year, because many of the large transactions come wherever they come. So it's really hard to make any statements based on Q1 only. I would say in general, our sense is that the merger is going very well or, actually, it's even strange to talk about the merger; that's a long time done, and we are now looking really at building the next basis of the Company, which is aside of the usual implementation of verification, VIP and the design for manufacturing. So we are not really reflecting any more on where Avanti used to be.

Steve, it may sound like a nitpick. I'm just trying to understand the dynamics of your expenses. It looks like maintenance and services were down about 24 percent year over year. That would be an expected trend. But the dollars spent on service were actually up year over year, and up like 5 million sequentially. Anything going on there? Or am I reading this wrong, somehow, because of the goodwill adjustments?

Yes. I think there are a couple of things that go on in that, Erach. One is that the way we do cost of goods sold has to do with allocation percentages, and those might have been adjusted from year to year. I think also what you see here is, as I had described previously, we have a consulting pipeline that has been working itself off, while another one is now in the middle of -- the beginning stages of picking up. And so I would expect that we would have expenses ahead of revenue on the consulting piece, until revenue begins to catch up as the consulting gets delivered. I think it's a timing and allocation question, but I don't think there's anything particularly funny going on there.

Rohit Pandey, Credit Suisse First Boston.

I'm trying to understand the synergies between the Synopsys and MoSys acquisition. What difference does it make to me if I buy the memory from Synopsys or I buy it from MoSys? What more are you bringing to the table?

Well, many things. First, a relationship with many, many customers. We are a very broad channel, that at this point in time is well experienced in how to deal with IP. And by the way, many customers have complained no end about the fact that they have to close legal and business deals with a lot of small IP vendors that are very complex, and therefore very inefficient. So customers really would like to have things from us. Secondly, we have a complete IP solution that has all of the major blocks, and therefore can put a high degree of emphasis on the interoperability. We know already today, for example, that the memory blocks from MoSys will be very compatible with our AMBA bus implementation, and therefore of direct help to customers.

Lastly, our relationships with the foundries are very good, indeed. And as a matter of fact, I think MoSys adds something to Synopsys, in that it allows us to get one more angle on the importance of yield, and on the ability to read process technology really well. So the combination of all those are, I think, very exciting, as a matter of fact.

So what kind of growth are you expecting from MoSys within Synopsys?

Well, we don't disclose individual product lines. I think I am on record as having said that our IP cores business is expected to grow over 30 percent, and that is on an organic basis. So the acquisitions are on top of that, and I would expect MoSys and Accelerant to certainly fall into a similar bucket of opportunity.

Do you have data on what's the yield for the MoSys chip, and at what technology?

I left that question to Fu-Chieh. I'm not sure if we disclosed that data, though.

This is Fu-Chieh, for MoSys. I am sorry that we cannot disclose the yield data from any customer-specific projects, although that we did over the last several years, is when we deploy our technology successfully, many customers saw it in production. They have generally fed back that the yield has been better than what they have been expecting. And the reason for that is that the overall cost reduction is not just because the 1T memory is smaller, but because the simpler construction is actually less sensitive to the process defect (ph), as well as just (indiscernible) density per unit area. So we are pretty excited about the opportunity, really, to have the right partner at the right time, to really position our product as a much broader deployment to the general SoC users.

Do you have memory blocks at 90-nm?

Yes, we do.

And do you have something at 65-nm you're working on, or something?

I think that would be what we would be working on.

On what fabs? TSMC?

I'm sorry --

We won't comment about future deliveries here. Obviously, many of the silicon vendors are very, very sensitive as to what we do in this domain.

Then you made a comment on the DC, increasing the DC capacity. You said that capacity went up by 25 percent. So what is the gate capacity for DC right now?

I don't have that number handy with me here. What we have known is that for a long time, the ability to increase capacity and runtime is of continuous value. Now, the 64-bit releases on the Optron (ph) platform specifically have immediately added a lot of additional capacity. What we are talking about here is on 32-bit platforms.

And the competitive tools that are offering synthesis without wire loads -- do you think that kind of a technology has some inherent speed advantages to the traditional synthesis?

Well, there are many, many different ways of calculating timing. And so, depending on what model you use, you can be much faster, and more or less accurate. And that is the essence of both synthesis and, as a matter of fact, also place-in-route (ph), which is, how do you trade off accuracy for runtime? And so the answer to that has to be one needs to be able to support a multiplicity of models. And as one gets closer and closer to the silicon, the accuracy and therefore the complexity of the model, by necessity, does increase. The key here is, can one predict what is going to come out in the silicon? And I think there, clearly, we have proven that we are very good at that.

Mr. Shevick, Mr. de Gues, I'll go ahead and turn the conference back to you for any closing comments.

Well, thank you very much for attending our conference call. As usual, Steve and I will be available for more detailed questions afterwards. And we appreciate the time spent with us. Bye-bye.

Ladies and gentlemen, that does conclude your conference for today. Thank you for your participation, and for using AT&T Executive Teleconference. You may now disconnect.