新思科技 (SNPS) 2004 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by and welcome to the Synopsys fourth-quarter and full-year fiscal 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). Today's call is scheduled to last one hour. Five minutes prior to the end of the call, I will alert the conference of the time remaining. As a reminder, today's call 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 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' quarterly report on Form 10-Q for its third quarter of fiscal 2004, which is on file with the Securities and Exchange Commission and in Synopsys' earnings release containing its fourth-quarter and full-year fiscal 2004 financial results, which is posted on its website.

In addition, 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 now available on the Company's website, in the earnings release containing its fourth quarter and full-year financial results, as well as in the financial supplement. The Web address for such information, including the earnings release, is www.Synopsys.com/corporate/invest/invest.html.

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

Thank you, operator, and thank you for joining us. This is Aart de Geus, and I have with me Steve Shevick, our CFO. As we have communicated through our fiscal year 2004, our strategy rests on three cornerstones -- building a predictable financial model, delivering a complete silicon-proven design solution and expanding our addressable market. In Q4, we made solid progress on all three fronts.

First, we executed on a significant change in our license mix, moving to an almost completely ratable model. Second, we saw the initial customer benefit of major technology advancements that we rolled out in our core design and verification platforms. And third, we made strategic investments that expand our reach in design for manufacturing and intellectual property.

This approach appears right on target, given the current business environment. The Semiconductor Industry Association has estimated that 2005 semiconductor revenues will be flat, which is consistent with our market assessment from the last quarter. Customers remain cautious, as they have a hard time accurately forecasting the coming year. With pricing schedule pressure coming increasingly from the consumer segments, our customers are consolidating to broader solutions from fewer vendors, in order to minimize both technical risk and economic cost. Our well-honed design solution is an excellent fit in this situation.

Against this backdrop, our fourth-quarter results were very solid, and well in line with our targets. Revenue of 230.6 million, non-GAAP earnings were 1 cent and book-to-bill was 1.3, not including a large renewal that we now expect to occur in Q1 '05. We increased the ratable portion of our business to 93 percent of our license orders for the quarter.

For the full fiscal year, revenue was 1.09 billion, non-GAAP earnings were $1.04 per share, and book-to-bill was 0.9. From a geographic perspective, revenues in all regions were within the historical contribution ranges in Q4 and FY '04. North America accounted for 56 percent of total revenue in Q4, Europe 16 percent, Japan 15 percent, Asia-Pac 13 percent. Revenue distribution for the full year was within 1 percent of these numbers.

In orders, North America contributed 50 percent for the full year, slightly below FY '03. Europe turned in a strong performance at 21 percent, much higher than last year. And for the first time, Asia-Pacific finished ahead of Japan at 15 percent versus 14 percent.

Looking at our products, our strategy is simple -- provide the most complete silicon-proven IC design environment. This strategy is based on our Galaxy platform for implementation and our Discovery platform for verification. These two major platforms are expanded down to silicon through our design for manufacturing solutions, and up to the systems level through our reusable IP.

Customers measure their success with our product in three dimensions -- quality of results, meaning area, timing, signal integrity, power and test; time to results, meaning the time it takes to go from concept to silicon; and cost of results, meaning primarily mask cost and yields. During the last two years, we have focused on completing our product offering and optimizing our solution against these criteria.

In implementation, the June '04 Galaxy release delivered full correlation in area, timing, power, signal integrity and test, as well as substantial run-time and capacity improvements. For example, Lucent reported a 25 to 47 percent run-time reduction in design compilers. Cisco reported "an awesome improvement," with 3 times faster speed. BAE Systems used the full Galaxy platform to design 24 radiation-hardened, space-qualified ASICs. The Galaxy platform and the use of the Milky Way database to support multiple design teams enabled BAE's teams to work together on 12 of the ASICs simultaneously.

With power becoming a primary concern in 90-nanometer designs, the power management features of Galaxy are proving very valuable. In more than 50 engagements in the last 12 months, Synopsys has not lost a single implementation benchmark where power was a critical factor. ST Microelectronics use our SoC disk test product to reduce the amount of test data by 97 percent, which improved test time and chip quality.

The power of the Galaxy platform increases in value as customers move to smaller geometries. With our signoff tools used on over 90 percent of all designs, we continue to have excellent visibility on advanced chip activities. The 90-nanometer tapeout count now stands at 119. 61 of them use Synopsys place and route exclusively, and 33 more use a mix of Synopsys', competitors' and internal products. The 65-nanometer tapeout count stands at 7. All use Synopsys back-end tools, five of them exclusively.

In addition to all the benefits that came from our June 2004 release, we have been investing simultaneously in the next generation, which is well on track to roll out in 2005. With smaller geometries, connecting the implementation tools to silicon becomes more important to ensure acceptable manufacturing yields. For this reason, as part of our effort to expand our addressable market, we have systematically invested internally and via acquisitions in design for manufacturing.

Our TCAD business had a good quarter. We are beginning to see TCAD tools expand from their traditional base in advanced process development to improving manufacturing yields, as well. KLA-Tencor and Toshiba recently announced a collaboration with Synopsys to develop new models that predict the impact of process variations on chip performance during real-time production. We are excited about this leading-edge area of TCAD for manufacturing, and have made a major commitment to this link (ph) to silicon by acquiring Integrated Systems Engineering just following the end of Q4.

Looking ahead, we believe that TCAD's predictive capabilities will complement our Galaxy design platform to help customers shorten the time to ramp up their production yields, there by reducing their manufacturing costs. In Q4, Proteus, our optical proximity correction product, turned in a strong quarter, with repeat orders from major foundries and adoption by NEC.

In the photomask area, we announced a first-of-its-kind collaboration with Photronics to reduce imaging costs and improve overall photomask quality for our joint customers.

In design rule checking, Hercules delivered speed improvements of 2X for single CPUs and 2X to 5X for multiple CPUs. We also established a 10X faster interface with our Star-RCXT extraction products. For the year, Hercules added 25 new customer logos, including many competitive displacements, and is now in production usage at TSMC.

Moving on to verification, our Discovery platform had a strong quarter. The benefits of 3X to 5X improved performance and VCS advanced bug-finding technology are paying off. 11 customers already use our new native test-bench VCS in production, and our assertion-based verification solutions are now deployed worldwide in over 100 project groups. NVIDIA, among others, is relying on our advanced formal verification to find very complex bugs. Our bug-finding technologies are helping us win head-to-head benchmarks, leading to competitive displacements. During the quarter, VCS added 18 new customer logos, including six displacements. We also added six new customer logos for Magellan, our hybrid RTL formal verification product. Our system Verilog capabilities are in over 35 (indiscernible), and are already deployed in seven live customer projects.

In analog mixed-signal verification, our Discovery AMS solution is performing well. HSPICE did very well, and NanoSim added eight new customer logos during the quarter. Atheros reported a 10X run-time improvement over competing simulators on their leading-edge wireless 802.11g chip.

Also in analog mixed-signal, we announced today an agreement to settle our litigation with Nassda through an acquisition. This settlement resolves the litigation between our two companies in a very positive way for Synopsys, and sends a strong message of our resolve to protect and preserve Synopsys' intellectual property.

We are finding that the worlds of verification and IP are moving closer together. While we continue to build our verification solution, we're also expanding our customers' ability to reuse IP blocks that have been thoroughly pre-verified.

In Q4, we solidified our leadership in the fast-growing PCI Express market. In addition to increasing our PCI Express business in the quarter, we taped out the physical layer interface, or PHY for short, with the smallest-area, lowest-power and most highly-differentiated diagnostics capabilities. We also announced the industry's first and only PCI Express core to pass the governing body's compliance tests.

In Q4, we acquired Cascade Semiconductor, immediately adding their route and switch port IP to complete our PCI Express portfolio. We now have over 25 PCI Express customers, more than double those of the nearest competitor. In USB, Samsung standardized on our solution. In addition, we announced the industry's first USB high-speed on-the-go PHY for 90-nanometer processes. To respond to the increased porting demands for geometries ranging all the way down to 90-nanometer, we acquired a team of more than 80 experienced digital and mixed-signal IP engineers from LEDA Design. Finally, four major customers standardized on Synopsys' DesignWare cores as their preferred IP solution, and committed to multi-year, multi-million-dollar orders for our portfolio.

During the year, our professional services executed over 30 tapeouts and actively participated in over 20 90-nanometer designs. In Q4, we also booked three flow optimization orders in Asia-Pac and Japan. In addition, we are actively collaborating in a unique, low-power initiative with ARM, national artisan and UMC.

In Q4, we strengthened our executive management team. Jay Greenberg has joined us as Senior Vice President of Marketing. With a background that includes TSMC, Jay will help us move towards more complete IC design solutions, and brings a wealth of marketing expertise to Synopsys.

In summary, we are executing well in providing the most comprehensive silicon-correlated design environment in the industry. In Q4, our technology investments of the last two years began to have a real competitive impact. We also moved to an almost entirely ratable business model, which provides stability and predictability while preserving the value of our technology. With a very solid Q4, we are well-positioned for major technology rollouts and a strong competitive performance in 2005.

With that, let me pass it on to Steve Shevick.

Thanks, Aart. I will first review our Q4 and FY '04 results, and then we will provide our guidance for Q1 and fiscal year 2005. My comments, including the numbers, will be posted on our website after the call.

During this call, I will discuss certain non-GAAP measures of our financial performance. Reconciliation of non-GAAP results to GAAP results appears in our press release, and in the financial supplement posted on our website.

Effective this quarter, we began a transition to an almost fully ratable license model. This has had and will continue to have a profound impact on the Company. Most obviously, and as communicated last quarter, this shift reduced Q4 revenue and earnings, as orders that once would have been taken to revenue immediately instead contributed to backlog. Our year-over-year revenue and earnings comparisons will be negative for a year, as quarters under the new model will compare unfavorably to quarters under the old, higher up-front model.

On an absolute basis, however, revenues should show sustained quarter-over-quarter growth, with occasional exceptions based on how the existing backlog turns to revenue, both as orders grow and as layers of subscription revenue build on top of each other.

From an operational point of view, the license mix shift is already making it easier for us to focus on growing the business for the long term, by delivering a stable and predictable revenue stream. With 90 percent or more of revenue coming from backlog, we have largely removed the quarter-end revenue pressure that can lead to bad business deals. This gives us greater flexibility to negotiate with our customers, and to close business when it is most advantageous to do so. It also should help us capture more value for the new technology we have coming out in 2005.

Both of these impacts were reflected in the fourth quarter. Revenue declined 27 percent from Q4 of FY '03, based principally on the change in license mix. Approximately 90 percent of our revenue came from beginning-of-quarter backlog. Backlog grew, and we hit our revenue target well in advance of the end of the quarter, avoiding the quarter-end squeeze that is common for companies more dependent on up-front license revenue.

Let me now turn to our results. Fourth-quarter revenue totaled 231 million, within our guidance range of 220 to 240 million. Full-year revenue was 1.09 billion, within our target range of 1.08 billion to 1.1 billion. One customer accounted for more than 10 percent of revenue for the quarter and the year.

Up-front license revenue in the fourth quarter was 18 million or 8 percent of total revenue, reflecting our shift to a more ratable license mix. Time-based license revenue was 165 million, or 72 percent of total revenue. Services revenue, which includes both maintenance and consulting, was 47 million or 20 percent of total revenue. 81 percent of revenue came from our core EDA products in design and verification, 15 percent came from our emerging businesses in IP and designed for manufacturing, and 4 percent came from services.

For the year, up-front revenue was 216 million or 20 percent of total revenue, time-based license revenue was 663 million or 61 percent of total revenue, and services revenue was 213 million or 19 percent of total revenue. The Q4 revenue mix is typical of what we expect for fiscal 2005.

Regarding bookings, the book-to-bill ratio for the quarter was approximately 1.3 to 1. 93 percent of product orders were booked as time-based licenses, and only 7 percent were booked as up-front licenses, versus the 63 percent to 37 percent time-based to up-front split for the first three quarters of the year. Approximately 5 million in up-front orders booked in Q4 will turn to revenue in fiscal year 2005. One customer accounted for more than 10 percent of orders in the fourth quarter.

For the year, total orders were 956 million, down from fiscal year 2003, due to the weaker-than-expected orders environment during the year, and to the timing of a major contract renewal. For the year, time-based licenses were 74 percent of license orders, and up-front licenses were 26 percent of license orders. No single customer accounted for more than 10 percent of orders for the year.

As indicated by the Q4 license mix, customers strongly prefer time-based licenses to up-front licenses. Time-based licenses also serve our goal of building backlog and revenue visibility. Going forward, we intend to offer up-front licenses only where the customer requests a perpetual license. Therefore, for 2005 we now expect that up-front licenses will account for no more than 7 percent of total product orders, and could account for less. This is reflected in our revenue guidance for FY '05. The average length of all of our renewable customer license commitments was approximately 2.9 years in Q4 and 3.2 years for all of FY '04.

At the end of the fiscal year, our aggregate non-cancelable backlog, which excludes consulting, was 1.48 billion, down from 1.61 billion at the end of fiscal year 2003, reflecting a book-to-bill below 1 for the year. Our aggregate backlog including consulting was approximately 1.53 billion versus 1.67 billion at the end of 2003. Aggregate non-GAAP operating expenses for the fourth quarter were 229 million, within our guidance range of 222 to 232 million, and were 853 million for the year, within our target range of 846 to 856 million. Non-GAAP operating expenses exclude amortization of intangible assets and deferred stock compensation, as well as merger-related and other special charges amounting to 36 million in Q4 and 150 million for the year.

The increase in non-GAAP operating expenses from Q3 to Q4 was expected, and was primarily due to an increase in sales and marketing expense, reflecting year-end sales commissions in Q4, and to savings in Q3 from a true-up of commission expense to reflect revisions to the orders forecast and a weeklong North American shutdown over the Fourth of July week. Q4 expenses also include 6.6 million of severance from actions taken during the quarter to control the growth in expenses.

As a result, non-GAAP operating margin was under 1 percent for the quarter. For the year, non-GAAP operating margin was 22 percent. Other income for Q4 was approximately 1.1 million, within our target range. Other income for the year was 2.3 million, also within our target range. Our non-GAAP tax rate remained 31 percent throughout the year.

Fully-diluted share count was 152 million for the quarter, within our target range of 150 to 158 million, and 160 million for the year, within our target range of 158 million to 162 million.

Non-GAAP earnings were 1 cent per share, within our guidance range of 0 to 4 cents per share. The severance costs referred to above reduced earnings by 3 cents per share. For the year, non-GAAP earnings per share were $1.04 per share, within our guidance range of $1.01 to $1.05 per share.

Turning to cash flow, cash from operations on a GAAP basis was 34 million in Q4. Non-GAAP free cash flow was 28 million, reflecting 10 million of CapEx spending and 4 million of merger-related disbursements during the quarter.

For the full year, cash from operations on a GAAP basis was 264 million. Free cash flow on a non-GAAP basis was 252 million, after subtracting CapEx of 45 million and adding back 19 million in merger-related disbursements and 14 million of FY '04 charges relating to our workforce realignment in FY '03. For 2005, we expect approximately 50 million in CapEx.

Cash and short-term investments were 579 million at the end of Q4, a decrease of 120 million from the end of Q3. The decrease was driven by 135 million in stock buybacks during the fourth quarter, plus the change in our license mix, which results in more customers receiving extended payment terms. Geographically, approximately 203 million of our cash is in accounts in the US and 376 million is in accounts overseas. Earlier this quarter, approximately 95 million of overseas cash net was used in connection with the closing of the ISC (ph) acquisition. We will not benefit from the recently enacted Homeland Investment Act, as our overseas earnings occurred in fiscal years not eligible for repatriation under that legislation.

During the quarter, we purchased 8.5 million shares of Synopsys stock, at an average price of $15.88 per share. For the year, we repurchased 16.9 million shares, for an aggregate purchase price of 423 million. As of the end of the year, 77 million remained in our repurchase program, and as of today, the Board has renewed the program and topped up the amount available to 500 million. We expect to repurchase at a slower pace in 2005, as the license model shift, combined with the level of orders received in 2004 and expected in 2005, will result in a lower level of free cash flow in 2005 than in 2004.

Q4 accounts receivable totaled 132 million, down 11 million from last quarter, and DSO was 52 days, up modestly from 46 days in Q3, due to lower revenue in Q4. Deferred revenue at the end of the quarter was 403 million, down 11 million from the end of Q3, reflecting the continuing shift toward quarterly payment terms associated with an increase in the proportion of ratable licenses. Aggregate backlog, which includes both deferred and off-balance-sheet deferred, and measures the aggregate commitment of customers to Synopsys, increased during the quarter.

Headcount totaled 4,447 employees at the end of fiscal year 2004, an increase of 59 from the end of Q3. This number reflects the elimination of 84 positions in connection with our fiscal year 2005 expense planning. The net increase is attributable to the addition of 58 people as part of the Monterey and Cascade transactions and regular hiring activities. The end-of-the-year total does not include 96 employees from the ISE transaction and 84 from the LEDA transaction, both of which closed in Q1.

I will now describe out expectations for the first quarter and for fiscal year 2005. In general, our view is shaped by the caution that we see in our customer base, which we think will impact the entire industry. More specifically, our outlook is shaped by our position at a relatively low point in the renewal cycle for our major customer contracts, as indicated in our last call, and the shift in our license mix. Let me remind you that all of our target numbers are estimates only, and are based on our judgment as of today. Actual results may be affected by many factors, including industry conditions, the mix of licenses actually sold, pricing payment terms and tax payments.

For the first quarter, our targets are as follows -- revenue between 233 million and 243 million; total non-GAAP expenses between 223 million and 233 million; other income and expense between negative 1.5 million and positive 1.5 million; a non-GAAP tax rate of 31 percent; outstanding shares between 144 million and 152 million; and non-GAAP earnings per share between 3 cents and 7 cents per share. We expect revenue from backlog to account for over 90 percent of revenue.

For the full year, our targets are as follows -- revenue between 930 million and 980 million; outstanding shares between 144 million and 152 million; a non-GAAP tax rate of 31 percent and non-GAAP earnings per share between 22 cents and 32 cents per share. We expect GAAP cash flow from operations of at least 150 million. These targets supersede all prior guidance for fiscal year 2005. The change in the non-GAAP earnings per share range reflects a reduction of approximately 10 cents per share from the notch-down in our expectation of up-front license revenue, and a reduction of about 2 cents per share in FY '05 dilution for lease and acquisitions. The dilution was offset somewhat by 2005 revenue improvement from Q4 subscription orders and by some expense savings.

We are not providing a target for orders. We will continue to provide the current level of information on completed quarters, but with our license mix, the actual timing of bookings matters far less than for companies dependent on current quarter turns revenue. We expect that orders will fluctuate, and will naturally move across quarter and fiscal year boundaries.

To provide greater visibility into our future revenue growth, we will provide estimates of revenue expected to roll off backlog over the next two years. Based on current backlog, we expect approximately 715 million to turn to revenue in fiscal year 2005, which provides a strong base on which to start the fiscal year. We expect approximately 465 (ph) million to turn to revenue in fiscal year 2006.

We have taken steps to control the growth of expenses, and continue to look at ways in which we can streamline the business and focus our resources on projects with the highest returns. Our strategy for returning the Company to growth and increasing profitability depends upon delivering key projects in 2005 on schedule. And therefore, we will not make major budget moves that would compromise our ability to deliver on these projects.

Let me briefly touch on the impact of acquisitions. Since our last earnings call, we have closed four transactions -- ISE, Cascade, Monterey and LEDA -- that have added strategic technology, deep customer relationships and R&D leadership to the Company. The aggregate net purchase price of these transactions was 120 million. An additional 31 million may be paid out over the next three years, based on performance.

In aggregate, in fiscal year 2005, we expect the acquired companies and assets to contribute approximately 35 million in orders and 20 million in revenue. The revenue number is heavily impacted by a transition to our license model. As a result, as mentioned above, we expect that the transactions will be dilutive by 2 cents per share in fiscal year 2005. We expect them to be accretive beginning in fiscal year 2006.

Recapping the terms of the Nassda transaction, the purchase price of the transaction is $7 per share or 192 million. The actual cost to Synopsys will be reduced by the 100 million of cash expected to be on hand at closing, and by approximately 62 million in settlement payments the individual defendants will make to Synopsys at closing. The net price is therefore expected to be less than 1 times Nassda's FY '04 revenue. Although Nassda currently sells mostly time-based licenses, the transition to our model will have an impact on revenue post-acquisition. We will provide additional information on the impact of the transaction on our financials on the first earnings call following the closing of the transaction.

With that we will now open for questions.

(OPERATOR INSTRUCTIONS). Rich Valera, Needham & Co.

It looks like you guys had a pretty solid bookings quarter, excluding the renewal, which you mentioned didn't close. Anything you could attribute to the bookings level? And was it within your expectations range, or was it maybe a little better than you expected for the quarter?

I think in general, it's absolutely true; we had a very solid bookings quarter, and we were not banking on any of the large deals that could have come in that quarter. I think partially it was the result of seeing the benefit of the technology we had rolled out in the June timeframe, which we have talked about. And as the benchmarking immediately solidified, and as the deals that were looking at technology (multiple speakers) -- I'm sorry; there is somebody else talking on the line here.

The technology really drove a number of the deals over the edge, and that was very positive for us, because we feel that we actually captured a few of the deals that some of our competitors had seen slip in their quarter. And that gives us a very positive sense.

And, Steve, it sounds like you are not going to provide any update to the 900 million bookings number from last quarter. Is that correct?

That is correct.

And with respect to the Nassda purchase, you have obviously got now two products that are very similar with NanoSim and HSIM. Anything you could say about how you plan to maybe eventually rationalize those products?

Well, the first observation is the products are actually aimed at very different market segments. The Nassda product is very much aimed at the memory and regular structure verification and simulation market. We have been focused so much more on the digital logic side. So from that perspective, we feel it is very complementary. What we do over time -- we'll see what we can do after we have closure.

Just a final question for you, Steve. The 150 million, I think, of GAAP cash flow -- is that correct? Free cash flow?

Yes.

What would that be? What would the pro forma free cash flow be?

Well, all I said was that we expected 50 million of CapEx, and then the merger-related charges we can't really predict at this point.

And are you assuming any change in working capital in that cash flow number?

Probably a level of detail I don't want to go into at this point.

Garo Toomajanian.

First, Steve, just wanted to confirm that your guidance did not include anything from Nassda, right?

Correct.

And also, did you say in your comments that you would expect from the license model shift that revenue would be down year over year, for a year?

What I said was that I would expect that the quarter-to-quarter compare -- say Q1 '05 versus Q1 '04, et cetera -- would be down for at least -- well, this quarter and then three more.

Okay. So not including Q4 of next year?

Not including Q4 of next year.

Also, you referred to a number of acquisitions recently. I wonder if you could articulate maybe what your acquisition strategy is right now. Are there still holes that you're looking to fill? Was that really the driver behind some of these, or what else are you thinking about here?

Well, as you know, generally we don't comment much about our forward strategies in M&A. But what we can say is, if you look at the acquisitions that we have done, they have primarily been in domains where we see expanding market opportunities. And so the two main buckets are clearly down toward the silicon, also sometimes refer to the design-for-manufacturing. And there, the set of problems and the complexity of the connection between those problems is growing rapidly, and so we have built a very strong portfolio all the way from TCADs to the maskmaking to the verification implementation.

The other area, of course, is the IP domain, which is really the opposite side of the picture, which is up toward systems. And there, the portfolio has really strengthen significantly around all the connectivity blocks that are used broadly by many, many chips and that can greatly benefit both from standardization and from a high degree of pre-verification. So that's a very good business for us to be in. So those have been sort of our main angles for the acquisitions.

And you mentioned a large deal that had split from Q4 to Q1. Can you give us a sense, or do you have a sense of whether the size of that transaction has changed at all in the time since that deal has been ongoing? And were there are other deals that had flipped in the quarter? What's kind of been the pattern with customers there?

The problem with the fact that there has been on a lot of speculation on essentially one deal. Therefore, it's probably better to not comment about the specifics. It's never good to reveal anything about specific customers. I can tell you, though, that there were no other deals that slipped. And we have communicated well ahead of time to you guys that the large opportunity was going to be either in Q4 or Q1, and we are reasonably confident that it will be done well in Q1.

Jay Vleeschouwer, Merrill Lynch.

(Technical difficulty) call such as this that you stated your expectation that for renewals -- that for most if not all of them, there would be an expansion in the value or inflation of the contract size, driven perhaps by capacity or new tools or product mix that the customers would need. Is that still your expectation, that in effect, the average annual run rate for most if not all of the deals you are looking at for '05 and perhaps even into '06 will necessarily have that expanded value? Or are you in fact hearing from customers, as we are, that there is really very little intent of improving average revenue run rates?

With the caveat that I think the first part of your question was cut off a little bit, let me start with the back end of the question, which is there is no question that the customers are very sensitive to costs right now, because they see a high degree of pressure coming primarily from the consumer side of the market, which sees high-volume but very price-sensitive consumers. And so, in that context, virtually all of our customers are trying to see how they can rein in their costs. And part of the way in which they can rein in their costs is actually to consolidate the number of suppliers that they do business with. And by the way, I would argue that that also reins in their technical risk. And so, from that perspective, we think that we are pretty well-placed.

In terms of the deals that we're doing, our objectives are still to grow each one of the deals as they come up for renewal, and not only because we have a very solid customer base that tends to come back to us, because we have delivered, but also because we have a set of new product capabilities and new products that are coming online. And so these products allow us to displace existing players or broaden the market opportunity. And so, be it (ph) both in the existing core areas of implementation and verification, where, starting with the 04-06 release, I think we have gained a lot of strength, and with the new areas, we do think that there is an opportunity for growth.

As a follow-up to that, your prepared remarks implied that the next generation of your design platform was on track for next year. So you are clearly counting on Galileo being enough of a functionality release to engender incremental bookings, as you got with Astro last year. And I guess the question is, even if it is a sufficiently strong functionality release, is the spending or competitive environment such that you can in fact derive more value for it? Or do you, in effect, end up getting held flat in first-run release?

Well, my first comment is I don't recall that we announced any products such as Galileo. On the other hand, it's pretty clear to everybody that we've been investing systematically beyond the integration that we have done the last two years. And indeed, there are some key capabilities that will come online in '05. The proof, of course, is in the pudding. Is there enough value there to actually derive value in return? Our belief is yes, and our early indications are that actually the technology is quite strong. And so, in that situation, it gives further impetus for people that are looking at either switching or that are looking at consolidating their solution to move toward somebody who can really deliver technology on a steady-state basis. And I think the Synopsys R&D engine has certainly been cranking very well. So I do believe that, yes, this will help us grow our business.

One technology follow-up on that. There was a fair amount of attention in the press this week to a new chip by IBM, Toshiba and Sony that they jointly developed, a new high-performance chip for multiple applications, including consumer. The question for you is, presumably somewhere in there, Synopsys tools are being used. But with that design as an example, how well-positioned to you think you really are for the long term, in terms of hardware/software co-design and co-verification requirements, as clearly, a chip like that represents a whole new class of performance and design requirements?

Sure. Now you are alluding to one of the areas that traditionally have been maybe connected but fairly adjacent to the EDA market, which is the hardware/software co-verification. And as the market has always looked very alluring, the return on that market has always looked pretty horrible, because a lot of people have used essentially free software, C compilers, et cetera, to do most of their work. I think, over time, the area that will be of interest to us in this boundary towards the software is really everything that touches the IP, because with the IP increasingly there is embedded software, and that is an area that in the future may be of interest to us. I don't want to overplay that for right now. Our main focus is on the four areas that I mentioned. We want to first deliver well there, and the whole hardware/software co-design has been a little bit of a Holy Grail that has not materialized in the last 10 years.

Rohit Pandey, Credit Suisse First Boston.

Last time, when you gave the 900 million guidance, that did not include the large deal. So, why are you not commenting on the bookings for next year? Are you implying that that number will go up when the deal closes in Q1 this year, your fiscal year?

Yes. I think that 2005 orders will be higher than 2004 orders. Beyond that, we are not giving a number, but the movement of the large deal from '04 to '05 certainly has a pretty major impact on '05.

But this is already accounted for in the revenues?

There is some assumption of the deal closing and of a certain size in our revenue guidance for the year, yes.

So it is above 900, but it's already included in the revenues?

Correct.

And then you said you had about 3 cents of severance costs this quarter; is that right?

Yes.

And the book-to-bill was about 1.3?

That is correct.

Harlan Sur, Morgan Stanley.

Aart, you mentioned several large OEMs in your call, I think specifically Lucent and Cisco. I was wondering what percentage of your business is made up of the systems OEMs versus the pure-play semiconductor companies. And more specifically, are you seeing any differences in tools demand from these two segments of your customer base?

You know, this is a question that has actually always been fairly difficult to answer, because there are so many people that are a little bit of both. But the vast majority, really, of our customers are very much in the semiconductor area, either having fab capacity or doing chip design. What I have noticed recently is that some of the system companies are coming back a little bit towards wanting to do a bit more of their own chip design, largely because they feel they can control better the negotiations on cost downwards. And so, be it the Ericksons or the Sonys or the Ciscos, those are what we would call the system companies. And they go a little bit back and forth on their investments in chip design, but right now I see a little bit of an uptick there.

When you say that they are bringing more of the design in-house, is that more COT or more ASIC work?

It is really more COT because otherwise, they can't really negotiate on pricing with their end supplier.

And you also mentioned on the call the number of tapeouts at the 90-nanometer and 65-nanometer nodes, but you have also been tracking advanced chip design projects, and providing this data to us on a quarterly basis. I was wondering if you have this information for the October quarter.

I have the information. I don't have it here with me right now. With also one other caveat -- as the numbers start to grow, on the overall number of design starts, if I recall well, it was about 340 or something, around those lines for the 90-nanometer design starts. My guess is that by next quarter or the quarter after, we will no longer be able to track that, because now it's only -- there is just many, right? We have been very diligent at literally tracking every design by name, so to speak. And so I just gave you the most advanced aspect and the most solid ones, which is really the tapeouts. So just to give -- for a general sense, about 340 to 350 was the design starts at 90-nanometer.

Okay. But that's up pretty substantially from the last quarter, where I think it was 227.

That is correct. It was up substantially. And what I am seeing, just having talked to a couple of customers literally in the last three days, is they are saying, well, on the one hand, I hate to go to 90-nanometer because it's hard to do, and so on. On the other hand, the degree of convergence that I can have in putting multiple functions of my product on a single chip is quite strong. For those people that have high volume, and the consumer market is a high-volume market, that is definitely of high value. And so it's a risk-reward type game, and I think that the 90-nanometer has sort of passed the infancy phase, and is now becoming where design is at.

I would assume that most of these 90-nanometer projects are by the IDMs. Are you seeing a pickup in 90-nanometer designs by the fabless guys?

Oh, yes. I think your statement already was not quite correct when the wins were 227 or so. The advanced guys, fabless guys, have no choice; they need to be at the leading edge of semiconductor technology. And so I would say it's already a few quarters now that the people beyond the IDMs do design of 90-nanometer.

And then just one last question, if I may. Regarding your high-speed serial interface IP cores like PCI Express and Serial ATA, maybe you can tell us at which fabs you have qualified your cores. And you also mentioned the traction you are getting in PCI Express. I'm just wondering what the demand has been like for your serial ATA cores.

You know what? That's a question I would rather like our GM of VIP business to come back to you, because I can tell you that we qualify virtually every one of our cores on TSMC just for basic business reasons. That's where most of the customers are. But increasingly, there are a number of players that ask us, can we qualify on their process, as well, because it's part of their differentiation. And so I am at a loss to be able to tell you exactly which ones those are for PCI. So if you don't mind, we'll follow up with you.

Bill Frerichs, D.A. Davidson.

To go back to this 3 cents of severance costs, is that in non-GAAP?

Yes, it is in non-GAAP.

So without the severance it would have been a 4 cent number?

That's right.

Sumit Dhanda, Banc of America.

A couple of questions. First, Steve, it seems like from what you had to say about the large deal now being scheduled for closure in the current quarter, your bookings did come up like you'd indicated, better than you had expected. Where was the incremental bookings upside coming from, in terms of perhaps the type of account or geographically, if you could shed some light on that?

It was kind of across the board. I think it's hard to pinpoint any one area. Some of the deals that were in Q3 that didn't close as upfront deals actually came back as subscription deals in Q4, so clearly, their changing the model allowed us to present some options to the customer that they found more acceptable. I think just in general, we were working a pretty large pipeline down through the end of the quarter, and more things were ripe and ready to come in than originally forecast. But to give you a geographic or any type of account, I think, would be difficult.

And I know you're not commenting on full-year bookings guidance. But, given the fact you are closing the "large deal" in the first quarter next year, is it fair to assume that seasonality won't be quite what it usually is, given that Q1 is usually such a weak quarter for bookings?

Right. From a bookings point of view, because of that deal, seasonality won't be pronounced at all, and Q1 will be fairly strong from the bookings point of view.

And then, as you know, about a couple of months ago, you had commented that, as it relates to '05, you are banking more on your "territorial accounts" and the global accounts, whereas the business opportunities don't look that encouraging. Has that view changed at all, as you have got one more quarter or so under your belt?

I don't think it's changed materially.

One more question. In terms of what incremental revenues you expect from your acquisitions next year, I think you indicated 20 million. Could you give what the equivalent number is in terms of orders? And again, for full year, does that not include Nassda, or is that only as it relates to the first quarter you are not including Nassda in your guidance?

I am not including Nassda in the number, and the full-year number I gave for orders estimates was 35 million, and 20 million for revenue.

And so Nassda is not there for full-year fiscal '05?

Well, we expect the thing to close, the deal to close in about three months. We don't have it in our published guidance now. Once it closes, we will then include it and give you the revised guidance.

Raj Seth, S.G. Cowen.

Steve, in '05, do you expect the average duration of the deals to change at all, or should we assume that it it's consistent with what we saw in '04? And with regard to '06, last conference call, you suggested -- and let's take out the impact that Intel in Q1 might have on the baseline. You suggested that '06, 25 percent or so orders growth would be the way to think about '06. Is that still the right way to think about that, again excluding the impact of the Q4-Q1 slip in the big contract?

Well, I am not giving '05 orders guidance. I am not going to repeat anything I said about '06. Clearly, even since last quarter, not just the big renewal, things have moved around a bit. I was trying to give a picture of what we saw as of last quarter, and so at this point, I think we will not repeat the comment on '06, and as '05 rolls out, then obviously things will change and develop on '06, and we will give guidance on that at the right time.

And how about the duration question?

Duration, probably close to 3.

And, Aart, one for you, if I might. Can you talk a little bit about what you are seeing with design start trends? You have talked about a lot of caution in the customer base, et cetera. What is your view on what is happening in semis? Is this a couple-of-quarter, two-three quarter inventory correction-driven cycle, largely, or something broader than that? And with regard to the installed base of software that is out there, how much of it today is capable of enabling 90-nanometer design? In particular, last time, you talked about some inventory or the notion of inventory, EDA inventory out there. And I guess I'm trying to get around where we are in the replacement cycle for tools to enable 90.

Starting with the environment around us, I use the words carefully of caution, lack of visibility. And actually, I think they are reasonably accurate. And it doesn't mean that people are looking at '05 particularly negatively nor positively. I think they just really don't know. And I think they don't know because there are too many moving parts all at the same time. We are coming off of, all in all, a fairly strong semiconductor year. We have just crossed back to the top level of semiconductor revenue that was the case in 2000, and people are just a little worried, are they going to get spooked? And then there are a set of macro phenomena that are all occurring at the same time. Globalization is moving very rapidly, opening up fairly substantial end markets in terms of consumer numbers. But consumers that obviously have less expandable revenue in their pockets, and therefore costs of the products will be under scrutiny.

Simultaneously, people are projecting fairly large amounts of silicon capacity to come online in the second half of '05, and that typically can have impacts on the pricing. And so you add this all up, and literally talking to dozens of semiconductor executives, they just don't know. And so they are just saying, well, we will be careful to not be in an '01/'02/'03 situation, and just take it a quarter at a time. And I think that's the caution that you see. It's not necessarily a negative caution; it's just a caution of we don't see very much.

In terms of the technology that is in place, as people move to 90-nanometer, they can perfectly well use the EDA that is there. However, they immediately need to amend that with sharper ability to deal with signal integrity, so the timing capabilities are really important. The power issues are very blatant already at 90-nanometer, and I highlighted earlier that it's an area that we have put a lot of emphasis on, because leakage currents are a big issue. And I would say that, for the first time, people are now consciously starting to say, what can we do about yield, because they know that by 65 nanometer, they need to be more than conscious; they need to be in full action. And so those are the things that are coming in, and that really layers technology opportunities on top of what we have today.

And lastly, regarding inventory, we always have a bit of a hard time to know how high the utilization rates are, but our sense is that, especially with going to a ratable model, we can wait a little bit until sort of the transactions are ripe, and therefore get a bit better value for the technology we do deliver.

The incremental functionality that you talked about increasingly necessary at 90 -- if I have a design seat that has worked for 130, and I am adding that incremental functionality, how much as a percentage of what I invested, assuming no renewals, do I have to spend extra, on average, to get that extra capability if I am already capable at 130?

Raj, It's a good question. I don't have a really good answer. I need to think about that. It's a good question I need to think a little bit about. My guess is that typically, you'll see the increase as a function of -- as the renewals happen, anyway, because we invariably layer in technology as it comes. Let me think about this question a little bit more before I give you an answer, if you don't mind. It's a good question. It's absolutely clear that there are some technologies that are needed that you did not need at 130 but you do need at 90, and we have many of those. So I'll go back to our GMs to check it out a little further.

I think our operator was trying to cut in to let us know that we have arrived at 3:00. Thank you very much for attending the conference call. As usual, Steve and I will be available for some follow-up commentary, and we appreciate the time you spent with us.

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