Aquantia Corp. (AQ) CEO Faraj Aalaei on Q3 2018 Results - Earnings Call Transcript

About: Aquantia Corp. (AQ)
by: SA Transcripts

Aquantia Corp. (NYSE:AQ) Q3 2018 Earnings Conference Call October 25, 2018 4:15 PM ET


Deborah Stapleton - IR

Faraj Aalaei - CEO

Mark Voll - CFO


Ross Seymore - Deutsche Bank

Chris Caso - Raymond James

Hamed Khorsand - BWS Financial

Joe Moore - Morgan Stanley

Quinn Bolton - Needham & Company

Blayne Curtis - Barclays


Good day, and welcome to the Aquantia's quarter three 2018 earnings conference call. [Operator Instructions] Please note, this event is being recorded.

I would now like to introduce your host for today's call, Ms. Deborah Stapleton, Investor Relations. Please go ahead.

Deborah Stapleton

Thank you. Good afternoon, everyone, and welcome to Aquantia's third quarter 2018 earnings conference call. With us today are Aquantia's Chairman and CEO, Faraj Aalaei, and Chief Financial Officer, Mark Voll.

The purpose of today's call is to provide information regarding Aquantia's third quarter 2018 financial results. During the course of this conference call, the company may provide financial guidance, projections, comments and other forward-looking statements regarding future market developments, market share gains, the future financial performance of the company, new products or other matters.

We want to caution you that actual events or results may differ materially. We refer you to the documents Aquantia files from time to time with the SEC, including our most recent Form10-K filed on March 7, 2018, and Form 10-Q filed on August 8, 2018. These documents contain and identify important factors that could cause our actual future results to differ materially from those contained in our financial guidance, projections, comments, or other forward-looking statements.

In addition, any projections as to the company's future performance represent management's estimate as of today, October 25, 2018. Aquantia assumes no obligation to update these projections in the future as market conditions may or may not change.

Also, the company's press release and management statements during this conference call will include discussions of certain non-GAAP measures and financial information. These financial measures and a reconciliation of GAAP to non-GAAP results are provided in the company's press release and related current report on Form 8-K, which can be found at the Investor Relations section of Aquantia's website at

For those of you who are unable to listen to the entire call at this time, a recording will be available via webcast for 30 days in the Investor Relations section of Aquantia's website.

And now I'll turn the call over to Aquantia's CEO, Faraj Aalaei. Faraj, please go ahead.

Faraj Aalaei

Thank you, Deborah, and good afternoon, everyone. In the third quarter, we recorded revenue of $32.9 million, representing 8% sequential growth and 23% growth year-over-year. The quarter represented our tenth consecutive quarter of sequential revenue growth.

Let me begin with a discussion on each of our main markets, first beginning with data center. Our data center business was up for the quarter when compared to the prior quarter and the third quarter of 2017. Through the first three quarters of 2018, our data center revenue is up 2.5% when compared to the same period last year.

In the third quarter, our shipments to Intel grew sequentially in dollars and ports. In the quarter, we saw a significant increase in demand for our 28-nanometer products and a decline in demand for our 40-nanometer products. As we have discussed before, the 28-nanometer products carry a lower ASP. Many investors have asked us in the past to explain the dynamics of our business at Intel. I wanted to take this opportunity to provide additional details between the various products that we sell to them.

Our revenue is made up of two kinds of products, ASICs and standard PHYs. Twinville was our first-generation ASIC product, which was built in 40-nanometer process technology. Its successor, Sageville also an ASIC, was built in 28-nanometer process technology. Both are built at Aquantia by combining Intel's controller technology with our PHY technology.

The second kind of product we sell to Intel is our standard PHY product, which does not require any integration of Intel's controller technology. This product is called Coppervale and is built in 28-nanometer process technology. Currently, the majority of our revenues come from ASICs, Twinville and Sageville and a smaller portion comes from the Coppervale. In 2018, the overwhelming majority of our revenue at Intel has come from our ASIC products.

We believe that although our data center customer sell-through was very strong in Q2, we have observed a lessened sell-through rate in Q3, thereby indicating softness in the end market, which has led to inventory build-up with our customer. Therefore, we are reducing our estimation of data center revenue contribution in the near term.

Having said that, this quarter, I'm excited to report that we have achieved a growing base of design wins with major data center switching customers, which can contribute meaningfully to our data center revenues in 2019 and beyond. This is an important accomplishment, as it can substantially diversify our data center revenues and help drive growth in that market for us.

Turning to enterprise, our infrastructure business was quite strong in the quarter with sequential revenue growth of 9% and year-over-year, sorry, year-to-date growth of 28% compared to the first three quarters of 2017. Business was up across all our enterprise customers, as we see multi-gig gaining wider acceptance in the enterprise.

In one case, one of our enterprise switching customers is designing a large family of new enterprise switches that take advantage of our latest products to build 32 different SKUs utilizing multi-gig technology. As these types of products make their way into the marketplace and fill in various deployment scenarios from a wide variety of vendors, end customer awareness is increased, which will in turn substantially grow multi-gig deployments.

Let me put into perspective the whole enterprise infrastructure opportunity, which is composed of enterprise and SMB switches and wireless LAN access points. In both cases, we are dealing with a transition cycle from gigabits to multi-gig Ethernet. 650 Group estimates of the annual multi-gig shipments, which combine the enterprise switch and wireless LAN access point markets will reach 100 million ports by 2022, a significant increase from their forecast of 7 million in 2018, representing a CAGR of 94%.

The advent of the latest Wi-Fi standards, 802.11 ax, which the Wi-Fi Alliance recently branded as Wi-Fi 6, is expected to become a significant driver for Multi-Gig adoption in wireless LAN access point in the enterprise infrastructure. According to the 650 Group, by 2022, shipments of Multi-Gig Ethernet and enterprise access points are expected to reach 24 million ports, representing a CAGR of 55% from 2018 levels.

As we discussed in our last quarterly call, 802.11 ax represents a tipping point for Multi-Gig, as this new Wi-Fi technology demands as many as two ports of five gigabit per wireless LAN access point, and an equal number of ports on the switch side. For 802.11 ax devices, we have multiple design wins with the leading OEMs for both retail and enterprise markets. The first products are expected to hit the market later this quarter, followed by a second wave of products early next year. It is logical that as enterprise infrastructure adopts those Multi-Gig switches and access points, other network devices will follow.

In another example of the growing ecosystem of Multi-Gig, we have a customer building a large family of enterprise storage devices, where the 10GBASE-T port utilizes the Aquantia action Ethernet controller to enable lightning-fast performance for applications such as database transfer, virtual machine storage, and larger data backups. Here is some perspective on how quickly this sector is growing. In the enterprise market, in 2016, our first customer went into production with the world's first Multi-Gig switch and access point. Two short years later, we are delivering substantial revenues and growing an extensive ecosystem of vendors supplying switches, access point, and storage devices.

Now, let's turn to our Access business, which is made up of two markets, client connectivity and service provider, which we call carrier access. With these markets together, this quarter, we experienced year-over-year growth of more than 190%, but were down 27% compared to the prior quarter. This decline in quarterly Access revenue was a result of a decline in the carrier access portion, while we saw a simultaneous increase in the client connectivity business. One of our carrier access customers experienced a fluctuation in orders during the quarter from the respective end customer. We believe this is temporary, based on the orders that we have received for Q4, and we expect both Access segments to grow significantly in the fourth quarter.

Our client connectivity business continued to ramp in Q3. The source of this revenue growth is from our many different customers and their respecting applications, which are bringing innovation to the marketplace. We anticipate that this business will continue to ramp nicely in the fourth quarter. The sheer breadth of customers generating revenue for us in this market is very encouraging and bodes well for us and our customers, as the end user adoption cycle makes progress. Several customers in the gaming space, including Asus and Gigabyte, recently announced new products featuring Multi-Gig LAN on motherboard, and we believe the adoption of LOM will continue to pick up speed.

On the professional computing side, we have design wins with Dell, HP, and Lenovo, the major workstation and business productivity vendors. These OEMs offer our NICs, our network interface cards, as a configure to other options across several SKUs, and in some cases, are selling our NICs as aftermarket purchase options. The Multi-Gig client connectivity ecosystem includes the various network interface cards, motherboards, and network-attached storage devices we previously announced.

The newest addition to this ecosystem includes our action products, which are Multi-Gig USB-to-Ethernet controller chips. These allow any USB interface to connect to the Ethernet network at 2.5 or 5 gig speed, something that has never been done before. These new action controllers enable our customers to build innovative Multi-Gig Thunderbolt and USB Type C docking station and dongles.

These Thunderbolt and USB implementations open up the opportunity for us to reach the laptop market, which according to IDC, showed global notebook shipments of 162 million units in 2017. This new class of high-performance, small-footprint device is perfect for laptop users who want mobility, combined with the reliability and performance of wired Multi-Gig Ethernet connectivity when at their desk. The end devices for these action controllers like Akitio docking station we announced last week, are the perfect solution for a market which values the slim and light form factors. Laptop users can have one cable that charges their computer, delivers Multi-Gig networking, and supports displays, multiple USB devices, as well as various other types of connections.

The other portion of our Access business which supplies the service provider market began shipping in volume in Japan earlier this year. Our previously announced design win with KDDI in Japan is the first deployment of Aquantia's Multi-Gig solutions in the service provider space. We have also started shipping to two other service providers in Japan, and we believe that large app need service providers plan to deploy Multi-Gig service in late 2019, and in time for the 2020 Olympics.

In other parts of Asia, service providers are developing Multi-Gig services using our unique high-speed technology. We have been collaborating with these vendors and their customers to develop the breakthrough innovation that will enable them to deliver Multi-Gig services that were not possible up to this point. We are conducting field trials at this time and have received our first production order with the potential for wide-scale deployments in the 2019 time frame. In these markets, each home represents a minimum of three ports of Multi-Gig Ethernet. That translates to a total opportunity of 150 million ports in the next 5 to 6 years.

In the automotive market, this quarter, besides the great traction among European and US car manufacturers, we have also made significant headway into the Japanese and Chinese markets, and we have furthered our engagement with major car manufacturers and Tier-1 suppliers. The success of our engagements worldwide supports our belief that we will start to see significant return from our automotive efforts in 2020, 2021.

A couple of months ago, we announced the formation of the NAV Alliance, an industry-wide organization with the goal to build an ecosystem for next-generation automotive Multi-Gig in-vehicle networks. NAV Alliance founders include the world's largest automaker, Volkswagen Group, which includes Audi and Porsche, the two largest Tier-1 suppliers, Bosch and Continental, and the leader in AI computing, NVIDIA, as well as Aquantia, the first company to introduce Multi-Gig automotive Ethernet connectivity solutions.

According to Strategy Analytics, the market for Ethernet connectivity in the car will reach 590 million ports by 2023, a tenfold increase from 2018 levels. We anticipate the start of production for autonomous vehicles in 2020 timeframe. Evidence to this is the flurry of activities we are witnessing with highly successful trials in different platforms, including some of our partners in the NVIDIA and NAV Alliance.

This month, an NVIDIA test vehicle based on the DRIVE Pegasus platform, which integrates our 10-gig Accelerate controllers and PHYs completed a fully-autonomous, 80-kilometer long, driverless highway loop here in the Bay Area. As we continue to engage with more companies in the automotive ecosystem, we are securing our position as the go-to supplier that can meet the stringent connectivity requirements and speeds necessary to achieve the highest levels of autonomous driving.

Before I turn the call over to Mark, let's summarize where we are. Despite the near-term elevated inventory situation in our data center market, all of our other elements of growth are performing as expected. We are introducing new products that are being well received and adopted by major customers. Our strategy of diversification is continuing to bear fruit in our new markets. The enterprise infrastructure and access markets are growing rapidly for us. Our investments in automotive are resulting in important design wins in multiple geographies with market-leading car manufacturers. Our product portfolio is proliferating, and margins have been expanding nicely. In spite of the reduction in our growth rate for fourth quarter, we expect our overall revenue to continue to significantly grow in 2019.

With that, I would like to turn the call over to Mark for more details on the financials for the third quarter and provide our outlook for the fourth quarter of 2018.

Mark Voll

Thank you, Faraj. Let me begin with a review of our financial results for the third quarter of 2018. Revenue for the third quarter was $32.9 million, above the midpoint of our guidance of $31.75 million to $33.75 million. Third quarter revenue represented an increase of 8% compared to $30.4 million in the second quarter and an increase of 23% compared to the $26.7 million in the third quarter of 2017.

Data center products revenue consisted of $17.5 million or 53% of revenue, representing an increase of 16% from the second quarter and an increase of 17% from the third quarter of 2017. Enterprise product revenue was $12.5 million or 38% of revenue, an increase of 9% sequentially and an increase of 16% from the same period a year ago. Access product revenue was $2.6 million or 8% of revenue, representing a decrease of 27% from the second quarter and a more than 190% increase from the third quarter of 2017.

Automotive product revenue was $258,000 or 1% of revenue, representing an increase of 7% from the second quarter and an increase of more than 150% from the third quarter of 2017. Gross profit in the third quarter was $19 million or 57.9% of revenue. This compares to $17.5 million or 57.6% of revenue in the prior quarter and $15.1 million or 56.5% in the prior year third quarter.

In terms of our operating expenses for the third quarter, total operating expenses were $21.3 million or 65% of revenue. Operating expenses of $21.3 million compares to $18.7 million or 62% of revenue in the prior quarter and $16 million or 60% of revenue in the third quarter of 2017. Total non-GAAP operating expenses were $19.2 million or 58% of revenue compared to $17.6 million or 58% of revenue in the prior quarter and $15.6 million or 58% of revenue in the third quarter of 2017.

Loss from operations was $2.2 million or 6.8% of revenue. This compares to a loss of $1.2 million or 3.9% of revenue in the prior quarter and a loss of $0.9 million or 3.4% of revenue in the prior year quarter. Non-GAAP operating results was a loss of $102,000, or 0.3% of revenue. This compares to a loss of $21,000 or 0.1% of revenue in the prior quarter and a loss of $0.5 million or 1.8% of revenue in the prior year quarter.

Non-operating income in the quarter was $0.3 million, consisting mostly of interest income. Turning to income tax, our tax provision in the third quarter was $143,000 compared to our prior guidance of $100,000 for the third quarter tax provision and compares to a benefit of $68,000 in the prior quarter. As we had previously stated, we expect fluctuations in our effective tax rate in 2018 as we transition our business to our international structure.

Our third quarter GAAP net loss was $2.1 million or a loss of $0.06 per diluted share compared to a prior quarter net loss of $0.8 million or a loss of $0.02 per diluted share and a loss of $1 million or $0.21 per diluted share in the third quarter of 2017.

Excluding $2.1 million of stock-based compensation, our non-GAAP net income in the quarter was $66,000. Our third quarter net income compared to the prior quarter non-GAAP net income of $338,000 or earnings of $0.01 per diluted share and non-GAAP net loss of $0.9 million in the third quarter of 2017.

Our third quarter 2018 loss per diluted share was calculated using 34.4 million shares. During the quarter, we had two customers that represented more than 10% of our revenue. Intel was 50% of revenue and Cisco was 30% of our revenue in the quarter.

Let me now present our guidance for the fourth quarter of 2018. Revenue is expected to be in the range of $33 million to $34 million. Our revenue guidance assumes revenue below our prior forecast for the data center market with all of our markets performing as we have previously anticipated. The midpoint of our guidance range would represent a 20% growth over the fourth quarter of 2017.

Gross margin is expected to be in the range of 57.5% to 59.5%, the midpoint representing a 50 basis point increase from our third quarter results. Our GAAP operating expenses are expected to be in the range of $20.5 million to $22.5 million, which includes stock-based compensation expenses in the range of $2.1 million to $2.3 million. We estimate our tax provision for the fourth quarter to be $400,000. We estimate our fourth quarter 2018 earnings per diluted share will be calculated using 35 million shares.

Looking ahead, we continue to see our business evolving as we had previously guided. For 2019, we project annual revenue growth similar to what we experienced in 2018, around 20%, driven by the growth of our enterprise and Access businesses, which we anticipate to experience higher growth rates. We expect to see incremental improvement in gross margin, as we have demonstrated throughout this past year. As we grow revenue, we see leverage in our business that would allow us to generate increasing operating margins.

With regard to our balance sheet, at September 30, 2018, our cash, cash equivalents and short-term investments were $67.3 million compared to $61.7 million at the close of the second quarter. Cash flow from operations for the third quarter was $7.3 million compared to cash flow used for operation of $3.4 million in the prior quarter. Accounts receivable of $15.8 million represents 44 days sales outstanding, which compares to $16 million, or 40 days sales outstanding at the end of the prior quarter. Inventory of $15 million represents 3.5 turns compared to $16.4 million or 3.2 turns in the prior quarter.

That concludes our prepared remarks. Operator, you may now open the call for questions.

Question-and-Answer Session


[Operator Instructions] The first question comes from Ross Seymore of Deutsche Bank.

Ross Seymore

Really just wanted to dive mostly into the data center dynamic. Faraj, if you could go a little bit deeper into what's going on with how the inventory built, et cetera. I'm sure you're aware, Intel itself reported tonight, and their business looked to be very, very strong. So, given their strength, can you talk a little bit about how the inventory had built up so much? And how long do you think that digestion period may last?

Faraj Aalaei

So, in the sector that obviously we're playing, the business actually did really - the sell-through was really well in Q2. In fact, it kind of reached record numbers in Q2, and you know, the expectation was that that trend is going to continue. However, what we've experienced on that sell-through in Q3 was, you know, short of that performance in Q2. And with Q3 being weaker, then the question is, okay, what is Q4 going to look like and how do you handicap that? And frankly, you know, because we're a step removed from that, we'll have to wait until Q4 is done.

But, you know, the purchase in Q3 and purchasing in Q4 would have been in anticipation of that Q2 kind of level sustaining or even growing, and since it's gone the other way, then that creates basically inventory situation if you kept shipping at those rates. So in terms of how long it's going to take, it's hard to say, honestly, but it's probably one or two quarters. Fundamental business is still there; it's just fluctuations from quarter to quarter.

Ross Seymore

Got it. And I guess it's an inventory management dynamic. Before you guys were public, I know you had a couple of these sorts of inventory adjustments, but you went to great lengths to describe to us how you guys have gotten tighter with Intel to try to avoid these sorts of situations, and I know this one is kind of an incremental one. It's not as big as the prior one, years ago. But just talk about what you guys do to try to manage that inventory, albeit not really knowing what the sellout is going to be at any given time.

Faraj Aalaei

Yes, so it's really - it's a process of just staying in close contact with your customer, asking the type of question we do at the end of every quarter. How'd the last quarter go? What sold out? We also look at the analyst reports of how many ports actually ship out into the market. We know what we ship out to our customers. Then we try to triangulate it. It's not a perfect science. But I think we've done really well over the last, let's say, 2.5, 3 years, the last 10, 11 quarters, being able to kind of keep it within a reasonable range of understanding what the sell-through is and managing it. And like you said, this one isn't huge, but it's one that we felt we need to react to.

Ross Seymore

And I guess, as my last question, both you and Mark talked about next year's growth remaining strong, roughly 20% year-over-year. It seems like the data center side itself might be a little bit weaker, at least the traditional part. Just talk about what gives you confidence in what would otherwise imply an accelerated growth in your other businesses or in the new portions of data center to still get you to that 20% bogey.

Faraj Aalaei

Sure. For one thing, what we just discussed was the fact in data center, we actually have had some significant wins on the switching side with some suppliers who can produce a fair amount of volume. As those products go into production, that in and of itself diversifies our business on the data center side. But on the enterprise and access side, just the level of activity and the growth that we see in that market, as evidenced by sort of like what we've been doing the last two or three quarters in that area, gives us huge confidence that that business is going to continue to grow and we're going to continue to do well in it.

And, you know, you look at the Access business, you know, we went essentially from standing still beginning of this year with our first deployments in Japan to what's on the docket in that country with the other service providers ramping. If you look in that particular market, KDDI, which was our first shipment, it's about 17% of total broadband subscribers. As other service providers get on board, that continues to grow. Other parts of Asia, as we said, we've gone through now successful field trials. We're starting to see the beginning of production orders. That bodes very well for us, as those markets are large. As we said, between those two markets, about 150 million ports need to kind of ship and get deployed in the next five years, by edicts of their government, so we are in a great position in those markets.

And just generally, in Access, in computing, in enterprise, we see just bigger engagement, customers putting out more SKUs of their products, and Aquantia is leading that market, so we feel that we're going to get more than our fair share.


The next question comes from Chris Caso of Raymond James.

Chris Caso

If I could just go back to data center a sec, and perhaps it would be helpful for you to give us an idea of where you think port growth on a unit basis has been a data center business, because that gets skewed by the ASP changes that happened. Just as a reference point, I know we're all digesting the Intel numbers because they come in the same time, it looks like their enterprise revenue was up about 1% year-on-year in the most recent quarter. That's a little below what I think most were expecting, which tend to support your comments. But I guess perhaps looking at the unit numbers would give us some sense of how much inventory may have been built in that segment.

Faraj Aalaei

Yes, so the unit numbers for this year we're expecting to end up someplace about 25% or so higher than the prior year. And our expectation, frankly, up to now, was that we were going to get the same kind of unit growth next year. So we'll see. You know, this could be just a short, one-, two-quarter thing, the softness, and it comes back roaring strong again, but that's sort of like what the history suggests.

Chris Caso

Okay. Just to follow up in general with regard to your guidance, if you're able to give a little more granularity within the different segments. I mean, it sounded like, you know, I suppose to get to those numbers, you're expecting data center to be down, still showing a similar magnitude of enterprise growth, and I guess Access would probably grow the strongest, based on your commentary. Is that the correct way of reading it?

Faraj Aalaei

Yes, I think enterprise will continue very nice growth, but Access is going to grow even faster, from what we see. Just adding these several service providers, and we named a few of our customers in the computing side like Dell, HP, Lenovo, which have now launched products of various sorts based on our chips. And there's new designs in the books that are coming up that will be introduced the first half of next year, which will continue that growth path for us. So, we are very bullish on the growth of our enterprise as well as our Access business next year.

And it's just, when you look across these markets, these Multi-Gig markets, whether it's Access or enterprise, just the number of vendors that are now involved and the filling in of their product portfolios gives us a lot of confidence that this market is going to just continue to grow, and hopefully, grow faster.


The next question comes from Hamed Khorsand of BWS Financial.

Hamed Khorsand

My first question was on the Intel side with the increase in revenue this past quarter. Is that new programs that you're getting included in on Intel, or is it just the changeup in their ordering metrics?

Faraj Aalaei

So at Intel, as I described on the call, we've had this historical product, the first generation of ASIC which, just a couple of years ago, they started to roll out our second-generation ASIC. And of course, as you know, these products have a long lifeline, so as a new one comes up, the other one kind of gradually goes down over a two, three-year period. So those 28-nanometer platform, whether it's an ASIC or a PHY, have been kind of growing steadily over time as becoming a bigger mix.

So we supply these products to Intel, and then Intel, whether it puts those on the NIC cards or supplies them as part of their motherboard solution, bag of chips, basically, to server manufacturers. And so, for us, they're not new design wins from us to Intel, but Intel constantly, obviously, is trying to promote its business by getting new designs and new configurations going with their customers.

Hamed Khorsand

Okay. And then, on the Access side, are you seeing new design wins, or are you just seeing activity from previous design wins that's drawing for the growth in Q4?

Faraj Aalaei

Actually, we're seeing both. So in the Access business, there are several new vendors that we have got design wins with which we just took through field trial, and those field trials have now been successful and gotten a thumbs up from the service providers. And so the next phase for these is for them to get into production with the service provider and for the service provider to roll out service, which we expect to happen in 2019.

Hamed Khorsand

Okay, and then could you comment on where you stand with the North America field trials with service providers?

Faraj Aalaei

So in North America, what we've done is we have design wins with the equipment manufacturers that supply the North American service providers. We have not gone into the field trials with those, although there is a lab evaluation of these systems taking place over the last several months. So our expectation is, is that the North American service providers obviously will probably kind of lag the Asian service providers, but they are doing lab evaluation of these design wins that we have.


The next question comes from Joe Moore of Morgan Stanley.

Joe Moore

You talked about some delays on the Access side in Q3, service provider delays. Was that a single platform? Was there more than one, and you know, do you see that fully coming into the fourth quarter, or will it take longer?

Faraj Aalaei

Yes, so Joe, what we were talking about was actually just the order pattern of one of the service providers, and it has come around in Q4. We've filled in the backlog, and our expectation is that that actually grows very - the Access market will grow very nicely from Q3 to Q4.

Joe Moore

Okay. I mean, I guess it depends on where Q4 comes out between the segments, but it looks like that your Data Center business will actually sort of be where I thought it would be and Access will be a little bit lower. Is that the right way to interpret it? And then, if I think about next year, 20%, can you just give us some qualitative assessment for which of the segments, could Data Center be sort of flattish next year and the other two segments drive the growth? Is that the right way to think of it? Or just, anything you can do to help us on that trajectory.

Faraj Aalaei

Yes, I think the best way to describe the way 2019 is set up right now is that a lot of growth will come from Access and enterprise market. Those will be the engines of growth for us. The Data Center, it somewhat depends on these design wins that we have, when they actually ramp up in production. It's most likely, frankly, second half of next year, so it will be sort of like a measured impact on our 2019. But it very much depends on the timing; when they go to production, they start shipping. We have won those design wins; customers have those boards in their labs and are bringing up. So it looks very good for having a meaningful impact on second half of next year. We'll see the size of that once they start to go to production. But no doubt the Access and enterprise are going to be huge driving engines for us.


The next question comes from Quinn Bolton of Needham & Company.

Quinn Bolton

Just wanted to follow up on Joe's question. With the ramp of the new switch design wins for data center, do you still think 2019 is a flattish year for that segment?

Faraj Aalaei

Yes, again, it could be. We're trying to weigh a couple of things. One is, this softness that we're seeing, is this short lived or very quickly recovering back up? In the past, when you look at historical trends, frankly, Q3 was - typically, Q3 and Q4 were good, strong quarters. This is like a little bit of a change from the trends we've seen. So that's on one hand.

On the other hand, the new design, the timing of when the new design wins go to production and the speed by which they ramp. So you've got these three different elements that says that hey, it's hard to kind of estimate exactly whether it's going to be a flat year or it's going to be an up year in data center, and we'll work through that as we get our switching customers into production and as we return to growth of revenue with our lead customer in that space.

Quinn Bolton

Got it. And then the second question I had just on the data center business, how do you tell whether this is just a temporary slowdown for Intel on the 10GBASE-T NICs relative to a shift in share between 10-gig NICs and 25-gig NICs. Mellanox put out a press release today touting their 25-gig NICs and saying that they've shipped 2.1 million, I think was the number, nine months to date here in '18. But the important part was they were claiming - they're now starting to target and see design wins in corporate data centers. I mean, do you have any market share or evidence that says this is kind of more just an inventory build, or do you think there could be some share shift between 10 and 25, or 25-plus NICs taking place here?

Faraj Aalaei

Well, I mean, we for sure know that, let's say, the inventory at least, if not the whole thing, is a major contributor to what we're seeing. As far as 25-gig NICs in the corporate data center, that is, quite frankly a small possibility, because one of the challenges in this industry in the corporate side has been to upgrade people from 1 gig to 10 gig, and 25 gig is far into the future for any such deployments. Now, it depends what you call of course corporate data center. There could be some very gigantic guys at the edge who are buying a few 25-gig NIC cards. But the majority of everything that even ships today in these markets is gigabit and 10-gig. So, yeah, I don't think that's a big contributor, quite honestly.

Quinn Bolton

Okay, I just figured I'd ask. The last question, you sort of talked about the Intel business, the PHY-only design is a smaller percentage of your revenue. I know there's a second-generation 14-nanometer port being bid. Any updates? Do you still expect to sort of get an answer on that by the end of 2018, or could some of these inventory issues perhaps just, make that second-generation 14-nanometer port perhaps not as important to the platform going forward?

Faraj Aalaei

Yes, so let me - first of all, there are obviously some conversations back and forth with customers that, under NDAs, I can't disclose here. But let me just kind of explain some facts around this whole next-generation chip that people have been talking about. First of all, this device, if and when it's put into production, will have minimal impact on our overall topline revenue. The reason for that is, based on our understanding, the time of its introduction will be no earlier than the next cycle of processor, which is expected to be released by mid-2020. And the extent of its impact is tempered by the end customer adoption, which, if you look back in the case of Coppervale one, it took two years or so after introduction before it started to even be cut in, right?

At this time, it's actually not clear to us whether the customer will even carry through with the development of this next generation, and it is our understanding somewhat that they may not even be targeted at the current markets we serve today. So, in a way, if it goes through and we win it, it could actually be an upside for us. And so, that's kind of the way we're looking at it.

To answer your question about the timing of the final decision by the customers, honestly, I don't know that today. But it's one that, at this point, we kind of look at it and say, this could be a way for us to, I guess, win it and get into other markets, because it's not very clear that will it come into the segment that we're in, even. That certainly doesn't seem to be its primary purpose at this point.


The next question comes from Blayne Curtis of Barclays.

Blayne Curtis

Yes, and I thought you said 25% port growth continues next year. I was wondering if you can give us perspective as to where the 28-nanometer part was as a percent of the mix, because I thought that transition would wrap up at some point maybe early next year. So kind of when you look at that outlook, I think you said, maybe it could be flat; depends on the ramps in switching. I'm just trying to understand the moving pieces in data center. You've had a flat business, but you've had this big ASP headwind. As you look to '19, can you just kind of give us some broad strokes on that rough mix? Sounds like PHY is not going to be that much. When does 28-nanometer kind of stop being that headwind?

Faraj Aalaei

Yes, so the way-we're still shipping 28, and we're still shipping 40. I think it stops becoming less of an issue, the 28-nanometer versus 40-nanometer headwind by probably middle of next year. We're still shipping a fair amount of the 40-nanometer. It's been kind of going down. Actually, this quarter looks like the shift is kind of accelerating, which was interesting to us, because it just kind of was steadily going down and the other one was going up and looked like it took a step - 40-nanometer took a step down. So the increase on the 28-nanometer mix actually is going up in Q4. But generally, I think, it stops becoming an issue in terms of a headwind to revenue probably by middle of next year, I would say.

Blayne Curtis

And then, maybe as a follow-up to that, I want to understand the December guidance. So it sounds like there's an increasing ASC headwind in Data Center, and then it's this unit correction. Is there a way you can kind of frame just the magnitude of this in terms of the December guide? Is there a way to think about maybe Data Center as it relates to the levels you were doing in the first half, or any color would be helpful?

Faraj Aalaei

Yes, probably, I would say that in Q4, it will probably be similar to - in terms of unit volume, it will be similar to what we did in the first half of the year, per quarter.

Blayne Curtis

Okay, and then just one for Mark, just on the gross margin that actually ticks up. I'm just kind of curious the drivers beyond that. Is it mix or costs or what? Thanks.

Mark Voll

Well, as we said before that we had a lot of test costs in our product and we're working to eliminate some of those test costs. And so, we've been able to - again, we thought we were going to be able to do that a little earlier in the year. It was delayed somewhat, but now we're starting to see that. It's happened the last couple quarters. We're projecting we'll see more of that as we go into Q4 and then later into next year.

Blayne Curtis

Actually, if I could squeeze in one more, too, on OpEx. I had OpEx down, I think it was because maybe tape-out was coming back out of system. I think you're guiding OpEx up a bit. Just kind of, what's the reasoning for that, if I had it right? And then, does it then fall off in March? If you can clarify, thanks.

Mark Voll

Yes, so we're guiding for OpEx to be flattish to - from Q3, largely because of an accelerated design activity, and most of those expenses are going to be in R&D. Again, as Faraj alluded to, we continue to do more and more activity in the automotive business, and so that's requiring more and more resources, and so we're allocating more and more resources to those opportunities.


This concludes our question-and-answer session. I would like to now turn the conference back over to Mark Voll, CFO, for any closing remarks.

Faraj Aalaei

Want to thank everyone for joining us on today's call, and we look forward to reporting our continued progress next quarter. We do want to mention that we will be attending the Needham Networking and Communications and Security Conference in New York on November 13, the Raymond James Technology Investor Conference in New York on December 4, and Barclays Global Technology Media and Telecommunications Conference in San Francisco on December 5. If you plan to attend any of these conferences, we welcome the opportunity to meet with you. This concludes the program. You may now disconnect from the call.


As stated, the conference is now concluded. Thank you for attending today's presentation. You may now disconnect.