FormFactor Inc (FORM) 2006 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the third-quarter 2006 FormFactor, Inc. earnings conference call. My name is Shaneek, and I will be your coordinator for today. (Operator Instructions). As a reminder, this conference is being recorded for replay purposes.

I would now like to turn the presentation over to your host for today's call, Ms. Annie Leschin. Please proceed.

Good afternoon and thank you for joining FormFactor's third-quarter 2006 earnings conference call. With me on today's call are Joe Bronson, President; and Ron Foster, Chief Financial Officer; and Igor Khandros, CEO. Joe will provide a summary of our third-quarter performance and then review our market segments and recent announcements and provide an update on our corporate priorities. Ron will then take us through the financials in more detail and provide guidance.

I would like to take a moment to mention that during the fourth quarter of 2006, the Company will be presenting at the Lehman Brothers Global Technology Conference, which is held December 5th through 7th in San Francisco. As details become available and other events occur, we will make additional announcements.

Finally, before I hand the call over to Joe, I will review the Safe Harbor statement. During the course of this conference call, we may make projections within the meaning of the federal securities laws, including statements regarding FormFactor's growth and financial performance as well as our strategic and operational plans. These forward-looking statements are based on current information, and expectations are inherently subject to change.

Actual results may differ materially and adversely to those in our forward-looking statements due to various factors, including but not limited to, the rate at which customers adopt the company's newly-released architectures, technologies and products; the extent to which chip manufacturers modify and announce capital equipment expenditures and device roadmaps; and the company's ability to efficiently execute on its capacity expansion plans.

Please refer to the company's recent filings on Form 10-Q and 10-K for more related discussions of the relevant risks and uncertainties. FormFactor undertakes no obligation to review or update any forward-looking statements or update the reasons actual results could differ materially from those anticipated in forward-looking statements.

Finally, a breakdown of revenues and bookings by market and geography and a GAAP to non GAAP reconciling income statement is available on our website. In conjunction with Ron's comments, the schedule that reconciles our GAAP financial -- our GAAP financial guidance and certain non-GAAP financial guidance with respect to FAS 123R stock compensation expensing will also be available on our website.

I would now like to turn the call over to Joe Bronson.

Good afternoon. FormFactor had another record quarter as revenue production and net income reached historical highs. Demand for advanced probe cards in the third quarter continued to reflect the ongoing trends in the market, including customers ramping production from memory and logic to transition to advanced nodes and the ongoing need for higher parallelism in wafer tests.

Revenues grew 5% sequentially to $96.8 million. Gross margins were 53.2%, and non-GAAP operating margins continued to perform in the target range of 25%. Factory capacity is ramping as production efficiencies continued with increasing yields in many segments of our manufacturing operations. During the quarter, we announced lead-time reduction commitments to our customers, which may impact incoming order patterns while ensuring FormFactor a considerable operating advantage as we grow the business.

The market indicators remain positive for FormFactor this quarter. The third quarter grew mainly due to record-high logic and mobile RAM business. Growth was moderated in DDR2 momentum due to strong DRAM pricing market dynamics. Our 300-millimeter business grew again as capacity additions continued.

Design activity for advanced products continued to accelerate throughout the quarter. We began to penetrate the NAND Flash market with initial shipments of our Harmony one-touchdown Flash product. We're making volume shipments for the fourth quarter, although less than previously anticipated as product development requirements and customer trials were completed later than anticipated with our early adopter customers.

In logic, we also shipped beta versions of our fine-pitch logic probe card to an early adopter customer. DRAM continues to be a key market with significant growth still ahead. Our business grew this quarter due to the ongoing transitions to advanced nodes, 90 nanometer and below, to move to next-generation architectures, KGD in the mobile RAM market and the continued investment by our customers in 300-millimeter factories.

This quarter, we strengthened and expanded our relationships with industry leaders and penetrated new customers as well. We signed a major agreement with Hynix, which was publicly announced in June, and began to deliver advanced probe cards this quarter, which is enabling Hynix to significantly reduce their cost to test.

Mobile RAM was strong this quarter, driven largely by consumer applications as well as increasing new probe card design activity as customers prepare for new mobile designs, seasonal consumer applications and future node transitions. We anticipate mobile RAM will have very strong bit growth in 2007 as the testing challenges associated with increasing electrical requirements on lower power devices continue to benefit from FormFactor solutions. Year-to-date 2006 has shown strong DDR2 bit growth due to the ramp to 90 nanometer, 512 megabit.

We continue to see Vista as a major event for FormFactor's 2007 business. As Vista is being adopted, PC configuration will move to higher memory densities and therefore drive strong bit growth in DDR2 and graphics DRAM.

We saw the continuation of a business trend we first highlighted in Q4 of last year, the coexistence of multiple architectures -- DDR, DDR2 and now DDR3 -- this quarter, driving the proliferation of designs and incremental tooling cycles which amplify our business. As we mentioned, strong DDR2 -- DDR pricing caused some companies to shrink to the next node of DDR, which equated to additional tooling cycles and increased probe card sales last year.

We continue to believe in this trend, and our analysis shows us that we are roughly halfway through the probe card tooling cycle to DDR2. We do however expect to see some volume in DDR but continue with some small tooling events even DDR3 emerges.

The larger phenomenon witnessed as we exited the quarter was the strength of DRAM pricing, which is delaying customer decisions to move to the next node -- 70 nanometer -- as profits remain high at current levels. Although customers are not ramping 1 gigabit, which is dependent on 70 nanometer, the relative profitability of DRAM will drive more capacity increases here, even though the spot market prices have already started their anticipated decline. As DRAM prices decrease, manufacturers must adjust by making node shrinks to drive down costs or diversify to niche DRAM applications.

In the near term, we expect manufacturers to continue to increase their DDR2 capacity to benefit from DRAM pricing and to ramp 80 nanometer to maximize the lower-cost benefit over 90 nanometer, evidenced by designs of which we are already seeing in Q4. We look for the more significant tooling event to be the move to 70 nanometer DDR2 in 2007.

We're seeing a confluence of factors, including the strength of DRAM prices and the process challenges related to the 70 nanometer conversion combined with variation in customer tooling events, that are occurring before the market moves to 70 nanometer. Therefore, Q4 will still be dominated by capacity buys for 90 nanometer and 80 nanometer nodes and some customer specific tooling of specialty memories before the 70 nanometer ramp.

NAND Flash continues to be a key area of focus, both for the market and FormFactor. We introduced our breakthrough Harmony OneTouch NAND Flash product this summer. Since then, we have begun the integration activities with multiple testers in the market with shipments of both our 200 millimeter and 300 millimeter one-touchdown products to early adopting customers.

The initial performance data has confirmed our view that FormFactor's product will redefine productivity in this area, as the architecture enables higher uptime, faster setup. The architecture is also based on a new NAND-specific spring technology designed for shorter manufacturing cycle time, low cost and low maintenance.

We have now shipped early production units to three customers, as highlighted on the last earnings call. We're making initial volume shipments of our Harmony OneTouch NAND Flash product in Q4. Although initial engagements with customers have validated the value proposition of our new product, we have a longer product development cycle than originally anticipated as we integrate this revolutionary technology with multiple test platforms and use by our customers. This in turn has resulted in a push-out of volume production by a few months.

In the NAND landscape, we see many trends that will drive increasing complexity of testing requirements, including node transitions, which are necessitating more electrical precision and higher degrees of complexity; the need for known good die; higher access speed, which is becoming a significant issue for many consumer applications; and proliferation of designs to address specific needs of end-user applications.

As a result, the NAND market will become more complex and will develop in a similar fashion to the DRAM market. FormFactor's scalable Harmony platform is designed to deal with this broad set of requirements, including power delivery, electrical speed, high pin count, accurate probe to pad alignment over a wide temperature range.

FormFactor's logic business continued its strong performance this quarter in our existing flip-chip market, driven by the data processing market and new customer engagements in game consoles and other applications. We continued to make progress in the development of our new fine-pitch logic product. We shipped additional probe cards for qualification with our early adopter customer.

The market for advanced probe cards in the logic area is just emerging. The transition from needle cantilevered probe cards continues to accelerate across all major device segments. The movement to advanced probe cards is driven by test cost reduction, enabled by higher test parallelism; new requirements for small feature sizes; pitch; higher I/O counts; migration to flip-chip packaging; and the need for higher device reliability, especially in automotive and KGD applications. This migration is impacting the computer, automotive, communications and consumer device markets. We believe that logic is a major growth driver for FormFactor as our existing parallel probing products and new product offerings in fine-pitch logic and area array will change the landscape of the logic test market in 2007 and beyond.

Demand for known good die continued in the third quarter with the acquisition of three new customers for our upstream wafer-level burn-in product, which is now being used by almost a dozen semiconductor manufacturers to enable KGD applications. Customers are seeking to reduce their back-end cost of test, and KGD is an important contribution to that effort by enabling testing on wafer rather than in the package.

Today's devices require more advanced burn-in and wafer-level final testing to offset the higher cost of failure. As such, we expect continuous adoption of KGD applications for mobile RAM, logic, NAND and NOR. Our customers are also beginning to see the value of moving known good die beyond the mobile market into gaming and server applications, among others.

Our factory continued to make strides this quarter as capacity ramped and we made significant productivity and efficiency improvements. We improved factory yields in many of our operations, enabling notable improvement in manufacturing capacity which has reached approximately 110 million per quarter as anticipated. We also reduced lead-time commitments to our customers, enabling shorter cycle times in their manufacturing and further improving their productivity.

We have been discussing the need to globalize the company's operations in order to be closer to customers. Our approach is to establish benchmark factory layout, processes and manufacturing techniques in Livermore and then transfer the capability to a geographic region in Asia. The first phase of this plan will include assembly and test for probe cards, applications engineering, some development engineering, design, and customer support activities.

We have identified Singapore as the location for our first non-US factory, where we will add additional capacity for assembly and final test to service customers in Asia. We believe that Singapore offers the company an excellent location for this investment due to the availability of skilled personnel, the ability to protect intellectual property and the opportunity to reduce the company's costs including taxes. This is the first phase of our program, and we intend to have our Singapore facility functional by early 2008.

In summary, we were pleased with the company's performance this quarter. We achieved target model operating performance, while increasing our manufacturing capacity and reduced lead-times to our customers. 2006 continues to be another remarkable year for FormFactor with 63% growth year-to-date.

The decline in bookings we experienced this quarter was due to two factors -- strong DRAM pricing, which is deferring the DRAM transition to 70 nanometer into early 2007 for our major customers, and second, the delay we experienced reaching volume production of our Harmony OneTouch NAND product. We're currently starting to ship volume production and will begin market penetration in the fourth quarter.

As the transition to 70 nanometer begins in 2007 and we begin to penetrate the NAND Flash market with increased volume shipments in Q1 of our new NAND Flash Harmony product, we expect a resumption in FormFactor's strong growth rate. As customers' need for test cost reduction becomes increasingly imperative, the demand for our products will continue to grow. Our solutions deliver the lowest cost of test through our innovative technology and industry-leading manufacturing.

All of the long-term fundamental growth drivers of FormFactor's business remain intact -- strong NAND and DRAM bit growth. We expect DRAM at over 55% bit growth in 2007 -- second significant design activity driven by product proliferation in tooling transitions in DRAM and Flash, including increasing incidents of mid-node transitions, the transition to advanced probe cards in logic and Flash and finally the migration of final test to the wafer.

We continue to expect to grow at an annualized rate of 30% compounded over the next several years. We believe that FormFactor is well positioned to take advantage of these opportunities and transform the markets that we enter.

I would like to thank our employees for their hard work and effort to make this another successful quarter for FormFactor. I will now turn the call over to our CFO, Ron Foster, who will elaborate on our financial performance for the third quarter and provide guidance for the fourth quarter. Ron?

Thanks, Joe. Revenues once again hit record levels this quarter, climbing for the seventh consecutive quarter, reaching $96.8 million. This represents an increase of $4.3 million or 5% over the second quarter and $34.4 million or 55% versus Q3 of 2005. Year-to-date revenues are up 63% compared to the same period last year.

Revenues grew in several regions, particularly in Japan and Korea. Revenues in Japan grew as a result of strong mobile DRAM business, while revenues in Korea increased significantly as a result of higher DRAM orders and following our press release in June announcing a new business relationship with Hynix, which became a new 10% revenue customer in the quarter.

Revenues grew in North America in both logic and Flash customer business. Taiwan revenues decreased in Q3 from the record levels of Q2, due principally to the completion of the significant 90 nanometer tooling cycle in Q2 of one of our customers.

DRAM revenues grew 4% sequentially on a sharp increase in mobile RAM business. DDR2 revenues remained strong in Q3, although down from the historic high in Q2. Continuing the transition to smaller nodes, 86% of the DRAM business was generated by nodes of 90 nanometer or smaller compared to 68% of the business occurring in such nodes in Q2.

As projected, Flash revenues decreased in the quarter about 5% as [NF150] business declined, and Harmony NAND shipments will commence in volume beginning in Q4. NOR revenues increased as key customers executed their 90 nanometer tooling transitions. Today, most NOR-related products go into multi-chip packages and are already beginning to benefit from FormFactor wafer-level burn-in and high frequency test at probe, HFTAP, solution.

Logic revenue climbed 28% versus the second quarter on increased microprocessor business and importantly growth in our customer base as we build relationships for future business opportunities.

Known good die continued to be a growth driver for the company. KGD revenues, which consist of wafer-level burn-in HFTAP, grew 22% sequentially to approximately $13.3 million in the third quarter. HFTAP was up more than 100% from last quarter as we continued to grow our base of customers with this new technology.

We now have high frequency wafer test business in at least eight customers in DRAM, logic and Flash applications. And we have doubled the number of wafer-level burn-in customers since last year. A breakdown of revenues by market and geography is available on our website.

Bookings were $90.5 million, down 6% or $6.1 million compared to the second quarter and an increase of $27.5 million or 44% over the same period last year. Sequential decreases occurred in DRAM and Flash, while logic bookings increased compared to the second quarter. All segments increased year-over-year.

DRAM bookings declined sequentially 8% with reduced orders for 90 nanometer DDR2 and mobile RAM designs as customers are completing their tooling cycles for such products. And, as Joe mentioned, customer market economics, such as high DRAM prices, are delaying volume transition to 70 nanometer nodes.

Average selling prices are driving healthy profits for our customers' 90 nanometer devices, thereby delaying transitions to smaller nodes. Though, we did have a number of design wins and orders as well for 70 nanometer from multiple customers this quarter.

Flash bookings declined sequentially as we completed a 90 nanometer NOR Flash tooling cycle at a key customer and as we anticipate the first volume Harmony NAND Flash orders in Q4. Logic bookings increased due to strong tooling in microprocessors and chip sets at a key customer and ongoing transitions to smaller nodes. KGD bookings declined compared to Q2 in line with a seasonal decline in consumer products where mobile RAM is the largest early adopter of KGD.

Backlog at the end of Q3 was again below our historical range of six to eight weeks. Shorter lead-times are contributing to the lower backlog levels as we enter Q4. Notably, bookings that turn in the quarter have increased each of the last three quarters to the mid 40% range in Q3.

As our factory productivity translates to shorter lead-times in response to customer preferences, quarter-end bookings will become a less meaningful indicator of future business performance. Over time, we expect more of our business to book and turn in the same quarter and the beginning quarter backlog may be less than half of any one quarter's revenue. This can limit our visibility into the next quarter, though it improves responsiveness to our customers, enhances our competitive position and strengthens our long-term growth prospects.

As with last quarter, I will cover some non-GAAP numbers to supplement understanding of our GAAP financials. A schedule is available in the investor portion of our website that provides the GAAP to non-GAAP reconciliations.

GAAP gross margin for the quarter was 51.9% and 53.2% on a non-GAAP basis, which compares to 52.7 and 53.7 respectively in Q2. The 0.5 point reduction in gross margin resulted in part from onetime asset write-offs related to factory equipment obsolescence. Inventory reserves also increased; although, the variation was not outside of typical ranges.

We continued to make gains in factory productivity, yields and on-time delivery that position us for continued revenue growth and profitability. This improved efficiency has contributed to current revenue capacity of about $110 million per quarter. We have acquired additional space in our current Livermore facility to provide for capacity expansion as needed to reach about $150 million per quarter.

As Joe mentioned, we have identified Singapore as the location for our global expansion into Asia. We're now developing specific plans for this first Asian expansion phase. This new site will initially perform our back-end processes of probe head and probe card assembly and test. We currently anticipate that this capacity expansion will be in place in the first half of 2008.

Singapore should be in a position to handle back-end growth beyond approximately $150 million per quarter. We are in the early stages of evaluating the next phase of front-end wafer fab capacity for Singapore once we exceed $200 million of revenue per quarter.

The move to Singapore will provide a new base in Asia that affords solid IP protection, lower resource costs and significant tax incentives. Activities are being initiated to acquire resources in Singapore. The plan includes capital investments, human resource acquisition and facility actions in 2007 in preparation for a 2008 ramp-up. These activities with assistance from the Singapore government are planned to be achieved within the parameters of our company financial model of 25% non-GAAP operating profit and CapEx at 10 to 12% of revenue. Once full manufacturing capabilities are in place in Singapore, net income should be additionally improved through a significantly lower tax rate.

Operating expenses for the quarter were $31.3 million on a GAAP basis and $26.8 million non-GAAP, up from $29.6 million and $26.2 million respectively in Q2. As a percent of revenue, operating expenses were 32.3% on a GAAP basis as FAS 123R option expenses added $4.5 million or 4.7%. On a non-GAAP basis, operating expenses for the third quarter were 27.7%, down from the 28.4% last quarter.

R&D expenses, excluding FAS 123R, increased slightly over Q2 but declined as a percent of revenue to 10.9% from 11.4% as we continued to invest in the development of our Harmony architecture, fine-pitch logic and other development projects.

SG&A expenses, excluding FAS 123R option expensing, increased $600,000 to $16 million or 16.5% of revenues for Q3 compared to 16.6% of revenues for Q2. Higher litigation and compliance expenses in the quarter of $2.1 million, up from $1.5 million in Q2 accounted for the increase in SG&A expenses.

We continue to take actions to protect our intellectual property in Korea, Taiwan, and the US as part of our long-term strategy. There are no new updates in the Phicom lawsuit. The proceedings continue in both Korea and the US. We intend to protect our investments, past and future, so that our customers can enjoy the long-term benefits they provide.

On a GAAP basis, operating income for the third quarter was $19 million or 19.6% of revenue compared to $19.1 million and 20.7% for Q2. On a non-GAAP basis, operating income for Q3 was $24.7 million or 25.5% of revenue, up from $23.4 million and 25.3% of revenue in Q2. FAS 123R stock option expense increased to 5.9% of revenue in the quarter as new hire and existing employee option grants were provided in the quarter and employee participation increased in the employee stock purchase plan.

Interest and other income for the third quarter increased to $4.5 million from $4.2 million in Q2. The increase was attributable to higher levels of investments and higher interest rates. The yield on our cash investments in the quarter was approximately 4.2%.

The tax rate for Q3 was 32.7% compared to 34.6 in the second quarter. A onetime benefit from a reserve adjustment provided the rate improvement. The tax rate for the remainder of the year is projected to be 36%.

Net income for the quarter was a record $15.8 million or $0.33 per fully diluted share on a GAAP basis and $19.7 million or $0.41 per fully diluted share on a non-GAAP basis. This compares favorably to $15.3 million or $0.32 per fully diluted share on a GAAP basis and $18.1 million or $0.38 on a non-GAAP basis in the second quarter and $9.8 million or $0.23 per fully diluted share in the same period of 2005.

The impact of FAS 123R was $0.08 per fully diluted share in Q3 compared to $0.06 per fully diluted share in Q2. Cash and marketable securities were $460 million in the third quarter, up $27 million from the second quarter. Operating activities generated $17.5 million of cash in Q3 compared to $33.8 million generated from operating activities in Q2. The decline was a result of the funding of certain employee benefit plans and nonlinearity of revenues in the quarter.

During the third quarter, we spent $8.3 million in capital expenditures compared to $8.9 million in the second quarter. The majority was spent on factory capacity expansions.

Day sales outstanding declined to 42 days in the third quarter compared to 46 days in the second quarter due to lower average accounts receivable balance throughout the quarter. The accounts receivable balance increased $9.2 million from the end of the second quarter as a result of nonlinearity of revenues in the quarter.

Net inventories increased by $2.3 million during the quarter to $26.3 million. Inventory turns declined from 8.6 to 8.4 quarter to quarter as the company increased WIP balances for early Q4 shipments. Headcount increased to 1052 from 1013 in the second quarter as product development, customer support and other infrastructure headcount investments exceeded manufacturing headcount growth for the second consecutive quarter.

Going forward, we anticipate key probe card tooling events at 70 nanometer with multiple customers and return of seasonally-strong consumer applications with expectations that they will ramp in early 2007. As we enter Q4, we're on the trailing edge of 90 nanometer DRAM tooling, anticipating seasonally lower consumer business and are just beginning Harmony NAND Flash volume shipments. Consequently, we expect Q4 revenue to be in the range of $93 to $97 million.

GAAP operating income is projected to be in the range of 19% to 20%, including about 6 points of incremental FAS 123R stock comp expense. Excluding FAS 123R, non-GAAP operating income guidance is 25 to 26%. We are continuing to invest in capacity in Q4 and beyond to meet the strong demand picture we foresee in 2007. We intend to continue funding this growth within the 25% target operating profit model.

As our growth rate this year exceeds 50% and is higher than our long-term annualized growth model of 30% combined with the need to invest for expected strong growth in 2007, we will continue to focus the company on balancing capacity and expense management in the business.

We target GAAP earnings per share of $0.30 to $0.33 per fully diluted share, assuming the higher tax rate in Q4. FAS 123R expenses will have an additional $0.08 EPS impact related to stock-based comp in the GAAP results. This would translate to non-GAAP EPS guidance of $0.38 to $0.41 for the quarter.

Now, with that, let us open the call for questions. Operator?

(Operator Instructions). Mehdi Hosseini, Friedman, Billings, Ramsey.

Yes, I have two questions. If you could help me understand the mix of DRAM booking by DDR2 and mobile DRAM. And if you cannot quantify it, just if you could talk about it qualitatively.

And the second question has to do with the one-touchdown probe market. Do you see that as more competitive compared to DDR2 when DDR2 market took off a couple of quarters ago?

This is Ron. I will address your first question. I think your question was about DRAM bookings, DDR2 and mobile. We see some trend in terms of both DDR2 with the tooling cycle move from 90 nanometer moving down somewhat and also the seasonally lower cycle on consumer, which is the mobile DRAM marketplace. The bookings are somewhat lower as we enter the fourth quarter in both those areas.

Right. My question has to do with if DRAM booking is down 7% in Q3, would DDR2 be lower than 7%?

Would be lower than 7% of what?

Would DDR2-related bookings be lower than 7% Q3 over Q2?

I think they are about flat if I understand your question correctly.

So the 7% decline in DRAM bookings must be driven by DRAM -- mobile DRAM?

Yes, DDR2 is close to flat because it is tending to grow more as DDR tails off.

Would DDR2 account for about maybe half of DRAM?

Close to half, yes.

Yes, it varies a lot depending upon the other plays, including the seasonality of our consumer products. But yes, it's in that range.

Then moving on to my second question regarding competitive landscape for one-touchdown probe card, would that be fair to assume that it's more of a competitive market compared to DDR2?

This is Igor. The competitive situation as we move to one touchdown in DRAM -- I think that's the question you're asking.

Yes.

Number one, it's something that the trend that will probably unfold next year. Number two, we believe that we will have very strong competitive position in it. We do not see a significant change there, the way we see it right now.

Edward White, Lehman Brothers.

On the NAND opportunity, can you give us a sense -- now that you've made some milestones and achieved some milestones there, can you give a sense as you look out to the next year what the -- what's sort of the rough scale of the magnitude of the opportunity would be compared to the DRAM side and the logic side?

DRAM is clearly going to be very big because it's got some drivers as well. But, if you look at NAND Flash, do you see that -- do you see 2007 as being the big year for it? How will it fit in relative to your other businesses?

This is Igor again, Ed. Well, NAND, as you know, we started shipments this quarter. We are introducing a product that we believe will be a very, very compelling product. It's going to be a product that is one touchdown at 300 millimeter. In some cases, it will be two touchdown, depending on the customer and their tester and test configuration.

When you have this high a parallelism and you have this much capacity riding on one product, we took extra care to make sure that we've thought through the product as a system that we've thought through the productivity, impact on customers, uptime, fast setup, temperature response and all the other things that make reparability -- that could make it a really compelling high productivity product. So, we believe with that, we will be making significant gains.

NAND Flash market is an important market as I said in the past. For example, known good die testing market is probably -- over time is a larger opportunity for us. But NAND is an opportunity where we plan to make gains, and it's an important target for us starting this quarter but really next year.

Jim Covello, Goldman Sachs.

A couple of quick questions. Ron, can you walk through the issue around the gross margin this quarter and then the difference between the GAAP and the pro forma again? You said it was a onetime item related to asset write-offs, is that right?

Yes, just quickly walking through -- if you look at our non-GAAP gross margin, we had factory asset write-offs. We're transitioning into our new factory and exiting the old one that had an effect in the quarter. We have anticipated for some time that there would be some effects in our factory as time went on. We had some obsolescence right off this quarter and that was one event.

We also had an increase in inventory reserves this quarter; although, it's not outside of the normal range of variation we've seen say over the last couple of years in inventory reserves. But there's sort of an almost reverse phenomenon that goes on that I probably should characterize here. That is that as a result of strong factory productivity, which I emphasized in that context, sometimes in a custom product marketplace, strong yields will outrun the planning yields we used to purchase materials. And when that happens, you can have higher inventory write-offs.

The good news is the factory is performing very well. And as we adjust the planning yields back in line with that, we will get our inventory write-offs improved and we will get the full benefit through the whole factory process of better yields. So that's part of the phenomenon that's going on there that we observed this quarter. It's just that --

Can you help us -- what would the GAAP number have been excluding these onetime items? In other words, if we just strip out the [ESO] expense that was flowing through gross margin, what would the gap GAAP gross margin have been, excluding the one-times?

I don't have a specific itemization for it. But the asset write-offs were close to the entire decline that we saw in the quarter. Inventory reserves were even a larger number than that in terms of the delta quarter to quarter. And as I mentioned, the inventory reserves will tend to move around on us in part as we align planning yields with our factory output.

Then just around the customer delays for the timing of the DRAM and the NAND issues, how comfortable are you on the timing for them as we get into 2007? There's always the risk that once they slip a little bit, they could slip a little bit more.

Well, where we sit here today, we believe that as witnessed by some of the orders we have in-house that 70 nanometer is beginning to happen. So that's what we see.

As far as Flash, as you know, this -- it just took somewhat longer to work through some of the issues in NAND Flash products. And the first results are extremely positive in terms of what the performance of the products that we see and we're beginning volume shipments this quarter. So on both items, we have confidence that we're going to be making progress.

Timothy Arcuri, Citigroup.

Ron, can you help me understand if I took a normalized tax rate for the quarter -- so if I used 35%, which is what you had guided to, you would've actually missed your earnings guidance. Is that right?

The tax rate effect versus last quarter was about 1.5 points if that is your question, Tim. Actually, when I provided the guidance, I actually factored that into my EPS guidance if that is the question you're asking because we had a statute of limitations on that tax effect and fully anticipated it would come into play.

Then I guess the second question, bookings at 91. It sounds like backlog has kind of fixed it below the kind of normal six to eight weeks. So, guiding revenue to a mid-range of 95 in December, you have to be counting on a pretty big uptick in bookings to get to where the consensus number is for March, given that backlog has come down so significantly.

Is that the right way to think about it? Or is there -- do you think there was some double ordering, or does it have nothing to do with double ordering?

This is Ron again. Yes, backlog we expect in general -- although there will be great variation quarter-to-quarter -- is going to tend to run lower than what we historically have. That was part of my message. And in no small part that we are reducing lead-times with our customers, our on-time delivery substantially improved much to the pleasure of our customers.

And so, we're booking and turning more in the quarter. As I commented, we have increased our book and turn rates each of the last three quarters to the mid-40s range in the third quarter. And we expect that trend in general to continue. To answer your question specifically, yes, you have a logical deduction that we would need to have a good bookings number and a reasonable turns number to meet the guidance that we gave you.

And then I guess, Ron, just looking at March, the consensus is at 102 for March. So, looking out into March, you would have to have -- would you expect roughly half of that March number to be booked in December? Is that kind of the right way to think about it?

We don't run the company to consensus numbers. We basically analyze the business. As Ron mentioned before, our lead-times are coming down. The ultimate goal for the company is to be shipping probe cards that are built to order that intercept customers' first silicon. So we evaluate the business and we base our guidance on that, not on consensus numbers.

Chris Blansett, JPMorgan.

I just want to get kind of a status check on where your Harmony product gross margins are relative to the corporate average.

As we said in the past, when we introduce a new product -- although different products, they vary in complexity and they vary in the kind of value proposition that they give to customers -- but when we introduce a new product, we plan to introduce the product within the range of margins that is commensurate with our financial model that we communicated to you.

Of course, when new products ramp, you have some period of inefficiency. But normally when we ramp a product and when we ship it to customer, you should see similar profitability range and margin range as for all the other products.

Chris, this is Ron. The way to think about this is that has been commented, we have completely redesigned this product from the ground up to accomplish the objectives of the NAND Flash market. That includes a fast cycle time because that's what customers need here; higher uptime; higher productivity, which is very important in the NAND Flash world. And very notably to your question in addition to those value propositions, it's designed for low-cost and specifically oriented towards the NAND market. So, all those things are helping us with our margin objectives.

So, right now, I mean I know you're just beginning but are you already at the gross margin targets? Or is there -- it would be understandable if you are slightly below because you don't have a lot of volume yet. I'm just trying to get an idea --

That's right. We are just beginning to ship. So within a quarter or two, we may discuss it in more detail. But again, there is a pretty detailed analysis here that goes on in there. And when we introduce a new product -- and we've done it now on many occasions -- our intent is to introduce it within the gross margin range that would support the financial model that we have communicated.

And, as Ron mentioned, this product is very, very different from DRAM product. I mean it is specifically from ground up is designed for the NAND Flash requirements. And it is a compelling product -- very compelling.

So right now, are you able to command a bit of a pricing premium over the competitive products?

We normally sell based on what we say is the total cost of ownership advantage. That's how we price our product. Yet in most cases, that equation comes out with better price than competitors because their products are not as compelling.

And then last, I'm trying to get a little grasp on how big the NAND Flash probe card market is. And I wasn't sure if you could provide some color on either number of probe cards needed for a certain number of wafer starts or relative to the DRAM market if you had some idea.

I think it is a difficult question and the reason this is a difficult question -- when FormFactor enters a new market segment, it is not the current size of this market segment that matters. Really, what it can become when we begin adding the value proposition that we had, we'll reduce cost of test. And basically, that's how again we price our products.

If you look historically, we've amplified every market segment we have entered. So, again, we are just entering in this quarter and we will be reporting on it as we go. But the products do look very, very compelling.

Last question, for that market then, how much of it do you think is serviced by advanced probe cards already versus traditional probe cards as you're entering?

They are all in the one-touchdown NAND Flash. We believe that majority are -- I would say probably it's half and half today as we define advanced probe cards; that would be my guess. I don't have exact numbers in front of me.

But what's important here is that it's not just -- it's not just having a one-touchdown on contact. It's really figuring through the entire productivity equation, and these equations are pretty demanding. For example, you have so much capacity. If you have one-touchdown 300 millimeter, which is really truly an emerging market right now -- and if you look at how much capacity is riding on one probe card, the impact of damaging a probe card and having it sitting there for a day or the impact of longer setup time is really significant in terms of impact on overall customer capacity.

We have several components to this as a whole total product that make that an extremely compelling product to customers. And that's where we see gaining an advantage. So far, the response has been very good from initial adopters.

Mark Bachman, Pacific Crest Securities.

Ron, you mentioned 30% CAGR. When are you starting -- what is your starting time frame for this?

Starting time frame, we have generally characterized 2005 through 2009. Those are industry estimates, and we certainly intend to participate at that level. And we built our model construct on greater than 25% kind of growth range and have contemplated that we can accommodate a 30% compound annual growth rate with a financial model that we have put out there.

So, growing at 55% here or so in 2006, what -- in order to grow at a 30% CAGR, when do you figure you are going to have slower growth?

Well, if we knew that, I would be doing something different.

You must have some (multiple speakers) you must have some model out there in 2009, right?

I'm not sure that we are managing it that way. I can tell you that the fact that if we take the middle of the guidance that we will have grown well in excess of 50% in a market that we believe does grow to about 30% overall market shows that FormFactor is outgrowing the market.

But it doesn't mean that when we look at 2007, we are saying we can take a breather. What it means is and as we communicated, we're preparing to grow in 2007 -- a significant growth. As Ron explained and Joe explained, we are actually investing in additional capacity to satisfy that growth.

But from even competitive situation of FormFactor, if you look at our growth rate and look at the fact that we are probably in order of magnitude larger than the next MEMS probe card competitor, we are pretty pleased with what we've done this year. Again, it doesn't mean that we're saying, boy, we can now grow only this much next year and everybody will be happy. We will just plan to grow.

But, all in all, when Ron communicated financial model, we said that we can support a financial model when we grow 25% or over 25%.

I think, Mark, you have to listen carefully to what Igor said on the previous question. As you enter a market, we will tend to amplify that market. So there are markets that we're just getting into, which will enable us to grow. For example, the logic market for advanced probe card is just beginning. So, we will be able to amplify that market once we get in there with the product pipeline that we have.

And, we will do the same in NAND Flash. We're just getting started. We also have a very healthy pipeline of DRAM derivative products that will come from Harmony that will drive future node transitions from 70 and downward.

So you have to -- it gets into the law of numbers, right, in terms of -- after -- you just have to see what those markets can do. And those are our estimates about what the growth of the total market is. But we should at least participate at that level for sure.

So Joe, with that amplification that you just talked about, do you guys assume that you are the only ones that can capture that amplification?

No, you have to assume there's going to be competition. And that's why you don't think you're going to get it all. But, you have to design the leading-edge product that will get the kind of performance to win the market and also to achieve the target financial performance.

Customers are really driving cost reduction in test. This is -- they've harvested a great deal of their other cost structures through 300 millimeter, etc. And the cost of test is a major area for them. Known good die modules, we're just talking -- we're just starting to talk about those things. Those are going to be major customer implementations down the road.

But, let me add. In terms of amplification, I do believe that FormFactor through our ability to develop compelling products, through our ability to redefine probing systems, through our ability to redefine markets that we should continue being for foreseeable future the major beneficiary of any amplification.

So, to summarize, Mark, the market growth rate projected by VLSI is 30% compound annual growth rate. We believe that we can at least participate in that level of growth. We believe we have outperformed it in 2006 with our greater than 50% growth rate. And we're continuing to invest in capacity to ensure we can participate at the level that our customers need our products.

I want to make one other little comment. Last year, we were capacity constrained and we grew at a little bit greater than 30%. So, there is probably some factor in there as we got their capacity online that we picked up on demand and really grew faster than we might not -- we might have normally if we weren't capacity constrained in '05.

Two quick questions on capacity. When you brought the new factory online, you thought you could get a 35% reduction in lead-times. How much of that has been captured so far?

Basically, with respect to lead-times, we've focused on a number of different things. We focused on first of all, the manufacturing cycle time and the yields and wafer fab and then strengthening our capability in the assembly and test area.

After we have accomplished a bunch of those things, we then turned to design cycle time. Because then design cycle time is also a key driver of our entire cycle time. So, we have made these lead-time commitments to customers. And we have dropped the time pretty significantly in 2005, and I think we've -- we're not done yet. There is certainly more to do, but I think we made great progress towards the goal.

Now, we will take our last question, Operator.

Mark Fitzgerald, Banc of America Securities.

On the NAND delay here, is this a problem that is on your side of the ledger or is it an integration issue?

I think it's both. I think on our side of the issue, we had to really -- as we were learning more and more about the market, it's really we do not believe we have served adequately. We had to think through this as a total product. And what I mean by that there is not just shipping of probe card that works but making sure with the amount of capacity that rides in one probe card, making sure that customers will indeed see significant productivity gains.

Second issue that was I think mentioned by Joe was that you have now to make it work for multiple tester interfaces. Tester interfaces are all different. So, it is an electromechanical issue. Part of it is not our issue, but big part of it is also our issue. At the end of the day, we need to make sure that our probe card integrates with interfaces. And that did result in a couple of months of delay. But, these problems have been worked through.

So those are behind you -- those mechanical issues because--?

That's correct.

Yes, behind us.

Then, can you just give me some sense on the mix here of this NAND in terms of what you're shipping? Is it 200 millimeter? Are you shipping 300 millimeter this quarter?

We have shipped both now.

Both versions.

We have shipped 200 and 300 millimeter.

Than just one final quick question. The lead-time is coming down. You have historically not participated in an R&D opportunity because of long lead-times. Are they short enough now that you can go after that part of the business?

Not yet. But, that is the ultimate goal is indeed to be able with the most complex product to intercept first silicon and we are working towards the goal. But, lead-times are coming down.

Also, on the NAND Flash product, for example, it was designed very differently. So, starting from the MEMS spring architecture and the probe head architecture in a way that it is a lower lead-time product.

Thank you all again for joining us today. We certainly look forward to seeing you at upcoming conferences and on the next earnings conference call. Goodbye.

Thank you for your participation in today's conference. This concludes the presentation. You may now disconnect. Good day.