Inphi Corp
Q3 2014 Earnings Call Transcript

Published: 29 Oct 2014

[Operator:]

Good day, ladies and gentlemen, and welcome to the Inphi Corporation Q3 2014 Earnings Call. My name is Sarah, and I will be your operator for today. [Operator Instructions] And as a reminder, this conference is being recorded for replay purposes. I will now like to turn the presentation over to Mr. John Edmunds, Chief Financial Officer. Please proceed, sir.

[John S. Edmunds:]

Good afternoon, everyone. Thank you for joining us today for our quarterly earnings call to discuss our Q3 2014 financial results and business outlook. I'm John Edmunds, Chief Financial Officer. And with me today is Ford Tamer, our Chief Executive Officer. I will begin with the Safe Harbor and then Ford will give you an overview of our business. After that, I will provide a financial summary and the outlook for the fourth quarter of 2014, and then we'll be happy to take your questions. Please note that during the course of this conference call, we may make projections or other forward-looking statements. These forward-looking statements and all other statements made on this call, which are not historical facts, are subject to a number of risks and uncertainties that may cause actual results to differ materially. These forward-looking statements speak only as of today's call. We do not undertake any obligation to provide updates after this conference call. For further information regarding risk factors for our business, please refer to our registration statements on Form S-1, as well as our most recent annual and quarterly reports on Forms 10-K and 10-Q, all filed with the Securities and Exchange Commission, accessible at www.sec.gov. Please refer, in particular, to the sections entitled Risk Factors. We encourage you to read these documents. Also during the course of this conference call, we may make reference to non-GAAP financial information. A reconciliation of this information is included in the press release and on our website, which is available at www.inphi.com. This information is not a substitute for GAAP and should only be used to evaluate the company's results in conjunction with corresponding GAAP measures. Now to begin our review of the quarter, let me turn the call over to our Chief Executive Officer, Ford Tamer. Ford?

[Ford G. Tamer:]

Thank you, John. Good afternoon, everyone. Thank you for joining us for our third quarter 2014 earnings call. This was a very good quarter for Inphi, and 2014 is turning into a transformative year for Inphi. We have grown our company organically by nearly 40% year-over-year, and are continuing to drive to higher levels of profitability in the third quarter by achieving 14% non-GAAP operating margin. By closing the Cortina acquisition on October 3, we have now added a very profitable $80 million business with more than 110 engineers, bringing our growing communication products to approximately 75% of our ongoing revenue. We are truly excited about this increase in scale, allowing us to further capitalize more the high growth expected in the 100G and 400G service provider markets, as well as the next-generation data centers over the next several years. Now on to the Q3 highlights. As you know, we are always focused on delivering consistent and predictable financial results. Although semiconductors aren't Warren Buffett's cup of tea, we can still benefit from his often quoted line, "The predictable and proven results lead to stronger companies." We are very pleased to report that Inphi pre-Cortina revenue for the third quarter of 2014 was $36.3 million, a 7% sequential increase and a 36% year-over-year increase. This represents the sixth consecutive quarter of over 7% sequential revenue growth. In the third quarter, our earnings per share were 12% per share, right above our $0.11 target. Looking ahead, we expect more of the same high-speed performance from our core Inphi business in Q4, reflecting continued healthy adoption of our products and solutions. These consistent results demonstrate the value of our team in delivering to our customers and our predictable performance to our investors. I'm proud to say that since coming to Inphi 11 quarters ago, we have made all our projections as predicted. We made a strategic decision to focus on growing our communication business, and we're now proud to have evidence that our investments are paying off. Specifically, demand for 100G drivers, amplifiers, SerDes and CDR solutions remains very robust. Our industry-leading solutions for the long-haul metro and data centers markets have won hundreds of margin and line card designs. The production ramps and continued strong demand from our top-tier customers around the world will continue to drive our future revenue growth in Q4 and beyond. On the memory side. Our memory team has worked tirelessly on DDR4, and I'm excited to report that our DDR4 buffer received green qualifying component status from Intel in Q3. DDR4 is now ramping to production. Although late, the DDR4 will participate in this growing memory market. Combined with our continued focus and investment on this and future generations of DDR4, we look forward to regain our market position over time. We have several significant product and customer announcements in the third quarter. In September, at the European Conference on Optical Communication, or ECOC, in France, we announced sampling of the industry's first 45GBaud Linear Coherent drivers and amplifiers. These products will provide more flexible 400-gigabit solutions for metro and long-haul customers seeking to keep up with the demands of their products. As compared to today's best solution, the Inphi solution doubles the bandwidth capacity of these customers' fiber installations. Separately, for the metro market, we announced availability of the industry's first single chip quad linear driver to enable higher density, 100G system at a reduced size, and importantly, reduced cost. At ECOC, we also successfully demonstrated and announced immediate availability of a new 100G CDR Retimer for CFB4 and QSFP28 modules and 2.4 terabit high-density line cards. We also demonstrated an end-to-end band link, which is our future PHY technology for 4.8 terabit high density line cards, and PAM or Pulse Amplitude Modulation, which is a new advanced method for packing more data into a given transmission. Earlier in September, we attended the Intel Developer Forum, or IDF, where we showcased our 100G CMOS PHY/SerDes Gearbox, inter-operating with Silicon Photonics modules from Intel and high-performance packet processing from Tabula. Also in IDF, we showcased our latest production-ready DDR4 offerings for next-generation data centers. We are pleased to report continued increase of production orders for our DDR4 products in Q3. As you can see, innovation at Inphi continues at a rapid pace, as the need to rapidly move immense quantities of data accelerates. I am confident that all these new products will contribute new incremental revenue growth in the future. On the customer front, I am very pleased to report 2 significant customer recognition awards. For the second year in a row, we're deeply honored to receive the Supplier of the Year award by Sumitomo Electric Device Innovations, leading provider of optical solutions to the telecom, broadband and data communication markets. Additionally, at the 23rd Annual Supplier Appreciation Event held last month, we received fiscal 2014 New Emerging Technology Supplier Award. This prestigious designation is bestowed annually on a company bringing forth emerging technology critical to the success of Cisco products. This demonstrates Inphi's continued focus on our customers' success. I could not be more proud of these important recognitions, and my sincere thanks to each and every employee at Inphi for contributing to the success. Now let me turn to our biggest announcement in Q3, the now complete acquisition of Cortina Systems. We have now added a very profitable $20 million-per quarter business and could not be more pleased with this acquisition or with the integration progress to date. As we blend these 2 companies, we look forward to the benefits of scale and greater diversification in both our engineering strengths and our product lines. Currently, all aspects of the integration are running smoothly and are on track. Now let me get right into the financial implications, as we know that this is a subject of great importance to our newly expanded shareholder base. In the fourth quarter of this year, we expect to report revenue exceeding $58 million. Even more importantly, we look to report earnings of approximately $0.20 per share and financial metrics roughly in line with our long-term operating model. Thanks to the Cortina Systems acquisition, we now expect to achieve this model approximately 1 full year ahead of time. But this acquisition wasn't just about near-term returns, it was motivated by our desire to add depth and engineering talent without impacting the P&L so that we can expand our delivery of great products and continue to focus on customer success. In fact, we've already achieved early success on this mission. I met with quite a few customer at the ECOC show, and I'm currently joining you from Asia where I'm meeting with key Cortina and Inphi customers. Just prior to this trip, I spent a few productive days visiting with our newest employees in Folsom, Ottawa, Raleigh and Vancouver. I am invigorated by the enthusiasm of our talented employees, both original and new, around the globe. Before turning the call over to John for a more in-depth discussion of our financial results, let me close by saying it has been a gratifying year for us here at Inphi. With the addition of the Cortina products, customers and employees, we have achieved a major milestone in the growth of this company. When I arrived in February of 2012, with the help of our 140 talented employees, we achieved 70% of our revenue for memory products and 30% from communication. We had just announced a $17 million revenue quarter in Q4 of 2011. Our goal has been to increase our communication exposure because of the longer design cycle, the increased opportunity for differentiation, and the accelerating growth. By the end of this calendar year, and thanks to the effort of our combined 440 employees, we expect that through solid top line growth from new products and the addition of Cortina, 75% of our revenue will accrue from communication products and solutions and 25% will come from our still vital memory product line. And that's on our guidance of over $58 million for Q4 of 2014, which represents a 340% revenue increase from Q4 of 2011 3 years ago. I'm confident that we are very well-positioned as we enthusiastically look forward to the opportunities ahead. With that, and again, with my appreciation to our customers, our team and to you, our shareholders, let me hand the call over to John. John?

[John S. Edmunds:]

Thanks, Ford. Now let me recap the financial results for Q3. As Ford mentioned, Inphi reported record revenues of $36.3 million in the third quarter, which represented 7% sequential growth and 36% year-over-year growth. Perhaps more impressive is that we have sustained year-over-year growth of 38% for the first 3 quarters of fiscal 2014. This has been driven primarily by sales in the 100G coherent communications applications, bringing the communications business year-to-date to 90% year-over-year growth, and in Q3, now comprises almost 60% of total company revenues. For the second consecutive quarter, our communications business grew sequentially by approximately 18% in Q3. We expect healthy sequential growth in this organic business in Q4 as well. Sales of product in the high-speed memory applications are also up about 4% for the first 9 months of the year, though they are down slightly for the third quarter. Our DDR4 buffer was qualified by Intel in September, so we're now expecting a mix of growth in DDR4 product in Q4. That will most likely offset the beginning of any expected decline in Q3 in DDR3, resulting in essentially flat revenues for that 25% slice of our total business in Q4. On a GAAP basis, we incurred a net loss in the quarter of $6.9 million, or a loss of $0.22 per share. This included stock compensation expense of $6 million and the associated offsetting tax benefit of approximately $1.7 million. We also had acquisition transaction expenses, mostly in legal and accounting fees, of $1.1 million in Q3. Other GAAP income tax reporting differences were additional tax reductions of approximately $5.3 million in the quarter. These were primarily driven by nondeductible acquisition expenses and timing differences in inter-period tax allocation, coupled with changes in forecasted mix of taxable income between the U.S. and international jurisdictions. GAAP tax expense is complicated and difficult to follow. However, one way to think about the difference is to realize that we've paid only $40,000 in cash-based income taxes so far this year. Q3 '14 results compare with the net loss of $2.8 million, or a loss of $0.09 per diluted share in Q3 of 2013. To give you more detail in comparing numbers, let's now turn to some non-GAAP measures. On a non-GAAP basis then, net income for the second quarter was approximately $3.9 million, or approximately $0.12 per diluted share. This was at the high end of our Q3 earnings guidance of $0.10 to $0.12 per diluted share. This also compares with non-GAAP net income of $1.1 million, or $0.04 per diluted share for the same quarter 1 year ago. Non-GAAP gross margin for the third quarter of 2014 was 65%, which was up 40 basis points from the 64.6% reported in the second quarter. In Q3, gross margins were up due to improvements in yield and the higher mix of communications products. We currently expect Inphi's gross margins to improve to between 66% and 67% in Q4 due to a higher mix of communications products brought in through the Cortina acquisition. Non-GAAP operating expense for the quarter totaled $18.5 million. This was up about $0.3 million from the previous quarter, but less than $0.7 million increase that we had expected. Fundamentally, we didn't hire as quite as many people as we had originally planned, in part, due to the acquisition. We also had more employees taking vacation in the quarter than we expected. As we have forecasted previously, we also benefited in the quarter from additional nonrecurring engineering fees from customers, which offset R&D spending. And we also accrued incremental annual bonus expense, which is part -- which is in part what is driving the additional sales and marketing and G&A expense in the quarter. In Q4, we will continue to do some targeted hiring, and we will add Cortina operating spending for the full quarter. So for non-GAAP, on a run-rate basis, we would expect operating spending to be in the range of $28.1 million to $28.7 million, with approximately 68% to 70% of that being invested in R&D. With regard to the non-GAAP tax position in Q3, the projected effective tax rate for the year is now running at 27.2%, up from the 26.5% estimated last quarter based on a higher mix of revenue being booked in the U.S. This adjustment caused the catch-up rate to be used in Q3 of 28.4%, to bring the year-to-date rate to 27.2%. This also excludes the potential benefit from the expected eventual restoration of the federal R&D tax credit. We believe the 2015 effective tax rate can be further reduced from the 2014 levels to something in the 16% range when the federal R&D tax credit is reinstated. Other income in Q3 was approximately $0.2 million coming solely from interest income, consistent with Q2. Now turning to the balance sheet, cash flow from operations was up approximately $0.8 million in the quarter, driven primarily by $3.9 million in non-GAAP net income, plus another $3 million in depreciation expense, all offset by a net $6.1 million being invested into working capital, primarily an AR of $8 million and inventory of $1.8 million, offset by other working capital funding. AR is up primarily due to a $3.5 million increase in earned NRE payments, which offset R&D spending in the P&L that was in receivables at the quarter-end. AR has increased otherwise due to relative linearity of shipments being back end-loaded at the end of the quarter. We had approximately $4.6 million in capital expenditures in Q3, primarily from mask sets, lab equipment and facilities expansion. Cash and investments in marketable securities decreased by $3 million for the quarter, from $122 million to $119 million. This represented a slight decrease per share from $3.71 to $3.57 per diluted share at September 30. Post the Cortina acquisition, cash and short-term investments are expected to be in the range of $60 million to $65 million. Accounts receivable increased $8 million from $14.4 million, to $22.5 million, again due to NRE milestones earned and accepted at the end of the quarter, as well as the linearity of shipments being more skewed toward the latter half of the quarter. This also represented an increase in day sales outstanding, from 39 days in June, to 56 days at the end of the September. Inventory increased $1.8 million from $7.4 million in June, to $9.2 million at the end of September. Inventory days also increased to 64 days in September, from 56 days at the end of June. This implies an inventory turns measure declining slightly from 6.5x in June to 5.6x at the end of September. To some degree, our mix of inventory shifting over to communications products that have longer lead times and also require greater lead time investment to have inventory ready for changes in demand. To a lesser degree, we've also recently built up inventory of DDR4 product in Q3 in anticipation of the launch of Haswell and DDR4 product demand in Q4. Payables increased from $7.5 million to $10.8 million, and days payable outstanding went up from 56 days at the end of June, to 77 days at the end of September, due to the timing of inventory received during the quarter. Now let me recap our Q4 business outlook. I remind everyone again that the following statements are based on current expectations, as well as -- and include forward-looking statements. Actual results may differ materially. We do not plan to update nor do we take any obligation to update this outlook in the future. As a range of guidance for Q4, including a full quarter of Cortina, we forecast non-GAAP revenues to be approximately $58.6 million, plus or minus $1 million. We expect non-GAAP gross margins to be in the range of 66% to 67%. We expect non-GAAP operating expense to be in the range of $28.1 million to $28.7 million, of which approximately 68% to 70% will be invested in R&D. We are currently estimating the non-GAAP effective tax rate to be 27.2% for the year. We are confident these components would then align, resulting in non-GAAP operating margin of 17% to 19%, and net income of between $7.3 million and $8.3 million, which on approximately 39.1 million estimated diluted shares, would result in estimated non-GAAP earnings of between $0.19 and $0.21 per diluted share. We also estimate noncash stock compensation expense to be between $6.5 million and $6.9 million. In addition, we expect approximately $1.5 million in acquisition/transition expenses, about 1/3 of which are legal and accounting expenses, about 1/3 from outside consulting to combine the 2 general ledgers, and about 1/3 from the expenses, primarily from additional employees to administer 2 general ledgers through the end of the calendar year. We also expect that approximately $3.1 million of deferred margin on Cortina's books at the acquisition will be consumed in the purchase accounting, but will be added back to non-GAAP income to reflect the company's run rate business. This would imply a GAAP net loss in the range of $0.1 million to $0.9 million and would imply a net -- a GAAP net loss per share in the range of breakeven to $0.02 per share. We have not estimated any purchase accounting adjustments, for instance, increases in inventory turning value, as well as increases in the tangible asset amortization, which may increase the GAAP loss further. We will not update this outlook during the quarter and up until the next time -- time of the next quarterly earnings, unless Inphi publishes a notice stating otherwise. So please ask any questions you may have today during the general Q&A period. And now, we'd be happy to take your questions.

[Operator:]

[Operator Instructions] And our first question comes from Tore Svanberg from Stifel.

[Erik Rasmussen:]

This is Erik, calling in for Tore. Very nice results and great operating margin. Can you just talk about the opportunities in DDR4 and how you see this market transitioning? I know you talked about production orders in Q3. But how big is the market? What is your -- how -- what is your split between DDR3 and DDR4? And then how do you see that business in Q4?

[Ford G. Tamer:]

Thanks, Erik. This is Ford. A recent iSuppli data for DDR3 versus DDR4 volume split next year shows that the overall market will grow about 50% next year, growing to about 225 million from about 150 million this year. The market PAM expansion is primarily due to higher ASPs from DDR4. And according to the same iSuppli report, the crossover from DIMM volumes to -- from DDR3 to DDR4 will occur in the third quarter of 2015. For the next quarter, in Q4, we expect our memory business to be flat, made up by a small increase in DDR4 and continue to transition from DDR3 to DDR4 later in 2015.

[Erik Rasmussen:]

Okay. That's helpful. So how -- are we talking -- you're saying production orders in Q3 and Q4. We kind of still sub 1 million in revenue there? Or is it a little bit more than that?

[Ford G. Tamer:]

We haven't broken up the revenue for DDR4. What I would say is a good measure of market share would be on the Intel website. If you go look at the Intel latest DDR4 RDIMM system-level validation, and their latest report is dated October 15, 2014, there are 9 qualified Inphi modules at a total of 46 mentioned in the report.

[Erik Rasmussen:]

Okay. Switching to the optical comms business, it seems to me that the optical space there with some signs of choppiness in the spending environment, yet you guys continue to grow and there are signs for that momentum to continue. Can maybe you just walk through the puts and takes of what's driving that strength, and then the confidence that -- as you head into '15?

[Ford G. Tamer:]

Sure. So we break that into 2 segments

[Erik Rasmussen:]

That's helpful. And then maybe one final thing. Can you give the breakout next quarter of what the core Inphi versus Cortina is in terms of the midpoint of your guidance?

[Ford G. Tamer:]

We have not broken that up. But we do -- we did mention that Cortina will add about $20 million of revenue per quarter. So the traditional Inphi business is still growing at the same 7%. John, any more detail there? Okay.

[Operator:]

Our next question comes from Quinn Bolton from Needham.

[N. Quinn Bolton:]

John, Ford, I just wanted to follow up on Erik's question on the memory interface. Like, you talked about the lower market share here, initially, in DDR4, and that's going to be a headwind. But as you look out to 2015, can you comment what you expect in memory interface on a year-over-year basis, '15 versus '14?

[Ford G. Tamer:]

Yes. So as we mentioned, we're exiting 2014 at a -- roughly, at a 25% of our revenue coming from the memory segment and 75% now coming from the communications segment. As we previously mentioned in our prepared remark, both our DDR4 register and buffer have now been qualified at Intel and have achieved green status. And we look forward to regaining market share over the next year. We are continuing to invest in the memory space and are focused on renting our generation 1 product right now in the marketplace. In parallel, we're working on our generation 2 and our generation 3 products, which we believe will be market leaders in terms of performance and enable us to regain our more traditional leadership position.

[N. Quinn Bolton:]

Okay. Maybe just -- you talked to -- referenced the Intel validation results on the Intel site. I think from looking at those, it looks like at least on the LRDIMM side, you have yet to be validated, or at least on any modules listed on the Intel website, validated for LRDIMMs. You talked about receiving that data buffer qualification from Intel. When do you think we'll start to see the first LRDIMMs using the Inphi DDR4 data buffers? Is that -- would you expect that here in the fourth quarter? Or do you think it may take longer for Intel to refresh those validation results?

[Ford G. Tamer:]

Our validation green status has been achieved at the component level. We're currently working with Intel and the module makers to achieve the green qualification at the LRDIMM module level and expect that in the next few weeks.

[N. Quinn Bolton:]

Okay, great. Turning to the communications business, you talked about seeing stronger demand for 100G in the second half of 2014 versus 2000 -- the first half of '14. Can you talk about whether you're seeing any extension in lead times? Or is there any signs of double ordering, any supply constraints? Or do you feel that you're able to meet this demand and that this is -- that there's no evidence of double ordering going on at the 100G component level.

[Ford G. Tamer:]

Sure. Let me just say 1 sentence, and then I'll turn it over to John to go into more details. We do -- we did a survey of our OEM customer base for 2015, and are projecting continuing strong growth for the 100G coherent in the order of 50% year-over-year. And a recent check with Infonetics at the September ECOC conference also confirmed the same forecast. Now let me turn it over to John for any evidence in that double ordering. John?

[John S. Edmunds:]

[Audio Gap]

[Ford G. Tamer:]

John, we're having a problem with your line.

[John S. Edmunds:]

Can you hear me now?

[Ford G. Tamer:]

Now, yes.

[John S. Edmunds:]

Sorry. I don't think we've seen any of that sort of frenzied order pattern of double orders or that sort of thing. But the order volume continues to remain robust. We have good visibility moving out because of the longer lead times in communications. So we just continue to knock on wood. And if you follow the Infonetics forecast, you'd expect a very healthy 50%-plus growth in the next year as well.

[N. Quinn Bolton:]

Can you say what your lead times are? What are typical lead times in that business? Is it 12 weeks? Is it longer?

[John S. Edmunds:]

No. With some of the components, it can be as many as 12 to 16 weeks or longer, depending on the packaging and what has to be assembled.

[N. Quinn Bolton:]

Great. And then just a last question. You mentioned the PAM4 modulation technology a couple of times now. When do you think that gets deployed in networks? And is that going to be more of a longer-reach technology? Or is that really more for some of the backplane and electrical applications rather than fiber-based applications?

[Ford G. Tamer:]

So we will be sampling our PAM solution in the first half of 2015, and then we will be expecting revenue in system-type of solutions and module-type of solutions in end of '15, early '16. That technology gets deployed first to sort of increase the bandwidth and reduce the cost and power of some of these applications in both first modules and then in line cards, and eventually backplanes. So same progression that we've seen on a regular SerDes, starting in the modules, then go to line cards, then go to backplane.

[Operator:]

Our next question comes from Doug Freedman from RBC.

[Doug Freedman:]

If you could maybe give us a little bit of an update on Cortina and your ultimate goals with the design teams that you're acquiring there from Cortina. I know the revenue growth you've talked about, not expecting the legacy revenues to really be much of a growth driver. But if you could talk about your plans for the design teams that you're acquiring, that might be helpful.

[Ford G. Tamer:]

Sure, sure. So I -- this past week, I did visit Raleigh, Ottawa and Folsom, also visited a new team, we -- a small team we got in Vancouver from a different company. And also now, calling from Asia, where we had a very productive few days with Cortina customers. So we're very excited about -- the team joining us is the entire infrastructure business unit which joined us. I'm here actually in Asia with Hojjat Salemi, one of the cofounders of Cortina, who's now on my staff; and some other Cortina team members. It's a phenomenal complementary expertise and they will be focusing on optical analog and data center products. And we've already finalized the roadmap. They're already at work, hard at work on these products. And those include suppose their own networking products, as well as some are forward-looking optical and data center products. So we're very excited about what this team will do. And as I said, they're already super busy and at work on a variety of products for future growth.

[Doug Freedman:]

If you could, I'm not sure if this is for John or Ford. A few details on the memory business if you wouldn't mind. What percentage of your memory sales are LRDIMM at this point in time? And how do you see that progressing going forward? And I have a follow-up to that one.

[Ford G. Tamer:]

John, go ahead.

[John S. Edmunds:]

In the DDR3 space, the LRDIMM revenues have been relatively flat and are expected to be flat into the fourth quarter for us. We would expect that area to pick up over time, and it's really kind of a function of whether it happens in DDR3 or in DDR4 or both. But overall of our revenue base, it's probably about 25% or so of the revenues in that side of the business.

[Doug Freedman:]

And then as my follow-up, because of your late entry into qualifying your DDR4 solutions, how should we think about not only the ASP premium that you can capture with DDR4, but is there any margin impact from the transition when you start to see more DDR4 shipments later in 2015?

[John S. Edmunds:]

Well, I think there will be, obviously, a pickup in terms of the amount of revenue, in particular, that we get for the register in DDR4 versus DDR3. But generally, in -- overall, in a LRDIMM solution, the revenues will be slightly higher, and we really don't expect to see a lot of impact to gross margin. It is a multichip solution as opposed to the monolith that we shipped into the DDR3, so there's some additional cost with respect to that. But overall, the gross margins will be atleast comparable in terms of what we're seeing today.

[Operator:]

Our next question comes from Ross Seymore from Deutsche Bank.

[Ross Seymore:]

First question, really, the computing side comes into play again. Trying to think of what sort of growth rate you think is attainable there? Because there seems to be so many moving parts between the transitions between DDR3 and DDR4, the market share being a little bit late to the market. How should we think about that growth potential in 2015? Or if you want to talk about it from a relative basis, what percentage of sales do you think you exit 2015 after the 25% that you talked about exiting 2014?

[Ford G. Tamer:]

So I would say that we expect the memory business to remain relatively flat in 2015. Depending on this transition from DDR3 to DDR4, we could see a slide down and expect to exit the end of 2015, and that's in our projections, in the low-single digits, 20%. So it would be, call it, the mid -- anywhere from 22% to 25%. So -- I'm sorry, the revenue is growing, so probably closer to 20 -- somewhere between 20% and 25%, Ross.

[Ross Seymore:]

Okay. So I guess, just to clarify, Ford, the revenue, do you expect it to grow or -- in 2015 as a whole?

[Ford G. Tamer:]

No, we -- the computing revenue is not expecting to grow in 2015 as a whole. The computing revenue is expecting to stay flat, with a combination of a decline in DDR3 and growth in DDR4. And again, given our larger share in DDR3, we expect the overall revenue, even though there'd be a decline in DDR3 and the gross in DDR4, to remain on the flattish side, Doug. If then the communication side of the business grows, the computing, as a percent of total revenue, will decline. So we expect to exit 2015 in the range of 20% to 25% of our revenue coming from the computing type of business. And the one thing I would point out in the memory side of the business is we did prioritize our investment in both the register and buffer on DDR3, where we're the only vendors to have both register and buffer, as well as our communication products ahead, and that has paid off. So I think the good news, at least from an investment return point of view, is our investments have paid off, the investments we made. Obviously, we're late on the DDR4 and expect to recover from here.

[Ross Seymore:]

Great. I guess as my next question, just I know you haven't really baked in much growth in Cortina longer term. But generally speaking, that $20 million a quarter rate, how should we think about that changing over time? Is that something that is very sticky and will stay that level for the next couple of years? Is it going to kind of fade away? Is it reliant upon some of the new investment, making up the difference as it fades away? Just generally, how should we think about that $20 million [indiscernible].

[Ford G. Tamer:]

Sure. So that is composed of 2 segments, if you wish. One is the networking business, where Cortina probably has hundreds of different design wins for their 10-gigabit, 15-gigabit and 40-gigabit product. And this is really what's ramping right now in long haul, metro and data centers worldwide. So we do expect to see some significant growth and sustained growth from the Cortina networking piece of their $20 million of product. On the transport, this is typically sticky revenue. This transport market is roughly divided 25% amongst Inphi-Cortina, PMC, APM and high silicone PGA vendors, and we expect that to decline slowly over time. And so as a combination of the growth in the networking business, as well as the slower decline in the transport-type of business, we do expect that revenue to stay flat on the $20 million a quarter. John, any more color you want to add there?

[John S. Edmunds:]

No, I think you covered it, Ford.

[Ross Seymore:]

Great. I guess, one last one for John, quickly. How should we think about OpEx from this fourth quarter exiting level, where you guided to the OpEx being in the, what do you say, 28 -- low $28 million range. How should we think about that going forward?

[John S. Edmunds:]

We would expect that to grow modestly through the course of next year. We think we've added a lot of headcount organically this past year, and then that, combined with the acquisition, gives us the resources to really execute on the plan -- development plan we have for last year. So we don't expect to add a lot of additional expense. But if you just multiply the $28 million by 4, you'd get $112 million. And then you're probably looking after something in the range of $116 million to $119 million or something like that in the OpEx for next year.

[Operator:]

Our next question comes from Sundeep Bajikar from Jefferies.

[Sundeep Bajikar:]

First of all, can you give us some color on relative revenue contributions from each of the 100G components? Perhaps just tell us how each component is tracking relative to your expectations?

[Ford G. Tamer:]

Sure. So on the long haul and metro first, amplifier and driver products, as I mentioned, these are growing extremely strongly and still not being impacted by some of the report of strong CapEx. So we're still excited in the growth of these products in both Q4 and 2015. On the networking side, as I mentioned, we have now a much larger installed win of different gross modules and line card designs at different stages of design and qualification, all at Tier 1 customers. And we're talking about hundreds of line card and module design. We expect the 10G and 40G to keep ramping for the next 5-plus years and a strong ramp in 2015. We expect 100G to continue to grow at a good rate as a primary first CFP customers continue to roll out their solutions. And then next, as our enterprise line card customers continue to ramp, we expect probably a doubling of the volume growth in 2015 compared to 2014 in that line card business.

[Sundeep Bajikar:]

Okay, great. And then as a follow-up, can you give us an idea of how Inphi's 100G design wins are distributed across module and line card? And perhaps as part of that, just give us some color on whether you see any dependency or risk related to carry your CapEx spending either in the U.S. or other parts of the world, specifically in the new 100G on the module side and on the line card side.

[Ford G. Tamer:]

Right. Sundeep, as we discussed, we have not seen an impact on the demand of our 100G coherent product. And we do believe, as we discussed in the past, that this is a result of the coherent technology of really providing benefits from both a CapEx and an OpEx point of view to these service providers. So we do believe that the investment in the coherent technology will be prioritized over any other investments they will be making because of this reduction in CapEx and OpEx and increase in revenue flexibility that the technology can provide. So that 100G, if you wish, coherent technology has been maybe bucking the trend as far as being deployed and being accelerated in its deployment. I think on the enterprise and data centers, we have been primarily module-focused. So if you look at our revenue to date, it's probably been a majority coming from module. But as we go into 2015, we're going to see the line card take that over and probably get to a 50-50 module versus line card going into 2015. So we'll see significant growth from new deployments of line card on 100G that we didn't have before.

[Operator:]

Our next question comes from Jorge Rivas from Craig-Hallum Capital Group.

[Jorge Rivas:]

It's -- this is the Jorge on behalf of Richard today. First question, on the 100G products for the metro that you expect to ship in the first quarter, can you comment on the customer breadth for those particular products? And also, in which part of the network are those products going to be deployed?

[Ford G. Tamer:]

Sure. So we have a mix of traditional customers, as well as new types of customers. And the primary -- one of the new primary driver of that growth is really coming from data center-to-data center-type connectivity. So we're seeing some nice growth in -- where we used to be called metro, it really is today data center-to-data center connectivity, and we expect that to drive a good growth into both 2015 and 2016. Actually, if you go back to that Infonetics report that I mentioned, their "line" is that 2015 will be the start of that metro deployment and their quote says that they expect 100G coherent will grow in the metro in a big way 2016. So you'll see some very good growth prospect in 2015, followed by even stronger growth, in our mind, in 2016 as that data center-to-data center connectivity takes hold.

[Jorge Rivas:]

Okay. And one last question, maybe for John. The announcement of the acquisition from Cortina, as we go through the process, I'm wondering if you guys have been able to identify any additional synergies, maybe a manufacturing that can be a tailwind probably to your gross margins.

[John S. Edmunds:]

Well, as you know, we have guided with the acquisition of Cortina to -- for gross margins to move up in the range of 66% to 67%. We do think there's some opportunity to improve in there. There are some savings that we bring from the Inphi side, and there's some test capability in the Far East that will help Inphi lower some of its cost that Cortina had. So there's some opportunity there, yes, but it's difficult to quantify or really project when we'll see the benefits of those. But I would expect through the course of next year to see some level of influence being applied by those improvements.

[Operator:]

Our next question comes from Guy [ph] Richard from Northland.

[Auguste P. Richard:]

Gus Richard. A couple of quick questions. First of all, can you talk about the regional strength of long haul -- Asia, Europe, North America -- last quarter and sort of how you see them unfolding in the end markets?

[Ford G. Tamer:]

The reports, I think, of ups and downs have been coming fast and furious by a different geographies. We haven't seen a -- any specific -- in our demand, we haven't seen any specific trends that would differentiate our ordering patterns by geography. John, I don't know if you'd have any more color on geography ordering, but...

[John S. Edmunds:]

No, we don't -- I don't have it right in front of me, Ford. We're -- we'll be -- of course, we'll publish that in the 10-Q later on in another week or so.

[Ford G. Tamer:]

Yes. So Gus, over the past months or so, I've been at the ECOC show in France in September, obviously, spent a lot of time with North American customer in early October around after the acquisition was announced on October 3. And now in Asia, both Japan and China. And we see very strong demand and very strong positive reaction from our customers around the world for both the acquisition, the organic product, and what the 2 teams can do together. So we do remain very excited about what the future potential for Inphi is going to be in both the service provider and the data center markets around the world. And we have just won some very significant new design wins, and I'm very excited about what the potential of these new design wins and new product categories could mean for us in 2015, 2016 and beyond.

[Auguste P. Richard:]

Got it. And then just 1 follow-up. On the gross margins, the transport business and Cortina's sort of on the decline in the network business, and this is going to offset that to keep the revenues stable. It feels like your comm business in classic Inphi is going to continue to outstrip memory. So on the 1 side, I've got a positive variance on mix from Inphi and a negative variance in mix in terms of gross margin. How do you guys think about model for gross margin? And then you get the cost downs coming. How do you think about where the model should be for gross margins as you get out of next year?

[John S. Edmunds:]

So Gus, I think -- it's difficult to say, of course, because there's always going to be ASP erosion that you have to manage. We do have new products coming to market, and so those help reinforce the existing margin model. But for right now, we've got out for the next quarter to 66% to 67% range. I did mention earlier that there are some synergies that we think can be had in the acquisition that should help support that margin level as we move forward. We have good relationships with our suppliers that we've always been able to leverage as well. So our goal is really to maintain something in the 65% to 70% range over the next year or so.

[Operator:]

[Operator Instructions] And now, we do have another question from Tore Svanberg from Stifel.

[Tore Svanberg:]

So I just wanted to clarify something. So for the December quarter, you're basically guiding your core Inphi business to be up in the mid-single digits, and then for the memory interface to be flat, and then $20 million from Cortina?

[John S. Edmunds:]

Yes. That's essentially a rough feel for our expectation at this point, Tore.

[Tore Svanberg:]

Okay, very good, John. And a question for Ford. Ford, you talked about 100G coherent growing another 50% next year. I assume that 50% number was specifically 100G coherent. And wouldn't the 100G coherent still be the bulk of your revenues next year? Meaning, the core Inphi revenues.

[Ford G. Tamer:]

Yes, that's correct, Tore. On both [indiscernible] you are correct.

[Tore Svanberg:]

Okay. And just so I understand the mix of Cortina today, obviously the transport business will decline, however, the networking will grow. I mean, it just seems to me that the networking business could grow faster than the transport declines. So if we look at that $80 million run rate or $80 million for next year, are you just being conservative?

[Ford G. Tamer:]

I think it's early at this point. I think you are probably right. I think the increase in the networking could probably grow faster, providing a positive delta. But at this point, we're just being early, so we're being maybe on the cautious side.

[Tore Svanberg:]

All right. Lastly, you talked about the design win for CFP4. That was primarily for 2016 revenue, right? Nothing next year?

[Ford G. Tamer:]

That's correct.

[Operator:]

Looks like there are no further questions in queue, so I'll turn the call back over to Mr. John Edmunds for closing remarks.

[John S. Edmunds:]

Okay. Ford and I would like to thank everybody for joining us today. Let you know, we do plan on attending the RBC Conference in New York on November 11 and the Barclays conference in San Francisco on December 10, and then we'll be at the Needham Growth Conference in New York on the, we believe, the 13th. So again, thanks for joining us, and we look forward to talking to you again in the future.

[Operator:]

Thank you, everyone, for your participation in today's conference. This concludes the presentation. You can disconnect, and have a great day.