Inphi Corp
Q4 2019 Earnings Call Transcript

Published: 4 Feb 2020

[Operator:]

Good day, ladies and gentlemen, and welcome to Inphi’s Fourth Quarter 2019 Conference Call. As a reminder, this conference may be recorded. I would now like to hand the conference over to your speaker today, Vernon Essi, Senior Director of Investor Relations and Corporate Development. You may begin.

[Vernon Essi:]

Good afternoon, everyone, and thank you for joining us today to discuss the financial results for the fourth quarter of 2019. A copy of today’s press release can be found in the Investor Relations portion of Inphi's website and inphi.com/investors.

[Ford Tamer:]

Thanks, Vern, and thank you for joining us for Inphi's fourth quarter and year-end 2019 earnings update. Inphi closed out 2019 with a strong Q4. We finished the year with record quarterly revenue and operating and net income. Revenue and non-GAAP EPS were above the consensus estimates. Strength in several market areas and across several of our product lines contributed to a robust quarter. We are confident, this demand will continue in 2020.

[but our Tier 1 customer,:]

With that, I again thank every member of the Inphi team for their efforts that led to the results we just discussed. I have every confidence in the projections for what lies ahead for the company. And with that, over to you John.

[John Edmunds:]

Thanks, Ford. Now let me recap the financial results. In the fourth quarter of 2019, Inphi reported revenue of $102.9 million, which was up 9% sequentially and up 19% year over year. This result was $3.1 million or 3% better than the midpoint of our guidance. To address the growth in more detail, our cloud products, including PAM DSP, gearboxes, CDRS, COLORZ, Cloud, TiAs and drivers grew 26% sequentially and 43% year over year . Together, they comprise 65% of total revenues in Q4 '19. Our cloud growth was primarily led by 400-gig and 200-gig PAN business, aided by some additional growth in COLORZ. For the year, Telco represented approximately 40% of revenues, which was remarkable given the trade war challenges we faced throughout the year.

[Q - Paul Silverstein:]

Thanks. Appreciate you all taking the questions. Ford and John, just on the risk side of the equation, in terms of China both with respect to coronavirus and with respect to any prospective build up of inventory by Huawei. You’ve all been kind enough in the past to give the best for your confidence. I recognize that you're projecting much lower volumes for Huawei in the second half of the year. What can you all share with us the insight you have? You have some of your peers this morning and last night talk about expected demand impact on whether -- whether due to employees not being able to show up to work for their customers, et cetera. Any insight you can share with us on those two issues beyond the comments in the prepared remarks?

[Ford Tamer:]

Yes. Paul, hi. This is Ford. So first on the Huawei, we've shipped a significant amount to Huawei in January. So we feel the risk to Q1 is minimal. And as we discussed from Q2 to Q4, some of that revenue for us coming from Huawei is again very minimal compared to our total revenue. So we feel the impact of Huawei has been mitigated and it's not consequential for us for the remainder of the year. Regarding coronavirus, our primary vendors that are in the manufacturing modules have already resumed work as of yesterday, as of February 3. And so they're back to work. Everything is normal. We have one component that is being manufactured at a place where they'll be back to work on February 10. And there is a contingency plan to move the manufacturing to outside of China. So in case there is any issue. So we feel that impact at this point should be mitigated.

[Paul Silverstein:]

And then, Ford, if I could ask on the upside. What are you most -- you've got obviously several a lot of win in your sales. What are you most excited about? Is it possible for ZR to be upside? Assume it could be upside in pull forward, or are you pretty confident that that's a little late 2020 with the bulk of revenue coming in 2021 event?

[Ford Tamer:]

What we're more excited about, Paul, would be upside on the PAM. So the 400-gig ZR is going to take about a year to go through qualification and field testing. There is no way to shorten the period of time. So we expect it to go production on track in Q4 of this year as we’ve been saying now on multiple calls. So we've been very consistent on the timing. We still stick to the timing. We are very excited about the quality of the product that we have sampled to customer. Superb. No need for spend, very high quality product. Came up out of the box in couple of weeks. Just phenomenal execution by the team on the 400-gig ZR. And we not only by -- we sample module. We also have -- our own partners with sample module. And we are actually very excited to already be in field testing at both cloud and major mobile operator telecom type of customers. So we see a big upside in both ZR and ZR plus for that in 2021. As far as 2020, there could be significant upside on PAM. As of right now, we do expect a higher share than what we had initially predicted.

[Paul Silverstein:]

Great. I appreciate that. Thanks, guys.

[Operator:]

Thank you. Our next question comes from the line of Blayne Curtis with Barclays. Your line is now open.

[Blayne Curtis:]

Thanks. My congrats on the results. Just a couple. Just curious on Huawei, can you just remind us what it was in the beginning of the year? I thought it was high single digits. You said 11% for the year. Just trying to figure out when that started to pick up. And for Q1, is it above that 11%? I was just unclear.

[John Edmunds:]

Blayne, we don't really break out our major customers, 10% except once a year. I can share with you, I think for 2019. It was a little more of a barbell. So we had a big first quarter with Huawei, I think, before the ban came in. And then once the ban came in, second and third quarters tend to be lower. And then we did see more shipments in --- again in the fourth quarter. So that's kind of we just ended up for the year in total, then slightly above 11%.

[Blayne Curtis:]

That's helpful. Just John on eSilicon, I was curious, I thought you said legacy would decline 15%. I'm assuming that's without the portion that you're going to allocate to it. And I guess just curious how you plan on weaving this revenue in as you report by segment, if you can give any color there as to when that 40% would end up in legacy?

[John Edmunds:]

Well, you'll see it beginning as we report in Q1, Blayne, and we'll just blend all of the numbers together. Really the existing legacy revenue is only about $4 million a quarter, so it's not that significant and then we'll be adding the eSilicon piece in the first quarter and we'll be taking the other portions of revenue that pertain -- do pertain to data center cloud based products on the one hand and 5G on the other hand, those will be added into the cloud and 5G businesses that we report as well. So we don't intend to break it out separately.

[Blayne Curtis:]

Thanks. And just maybe finally for Ford on the PAN market. It looks like a very strong end of the year and continues into Q1. Just any update on the size of that market, given the strength you're seeing early on?

[Ford Tamer:]

Thanks, Blaine. We expect -- at this point, we haven't changed our sizing for that market as we progress through the year, we may very well do so. We're a bit, as I said, more bullish about our share of the market at this time.

[Blayne Curtis:]

Thanks.

[Operator:]

Thank you. Our next question comes from the line of Harlan Sur with J.P. Morgan. Your line is now open.

[Harlan Sur:]

Good afternoon, guys, and congratulations on the solid execution and strong start here to 2020. Your comments on the press release, Ford, suggest organic revenue core Inphi momentum in Q1 into the balance of 2020. Does that mean that you expect to grow sequentially through 2020 based on your design win pipeline ramps and new customer programs, even with the step down of Huawei in Q2?

[Ford Tamer:]

Yes, Harlan. We do expect to grow sequentially every quarter this year?

[Harlan Sur:]

Okay, great. And then on the strong March quarter guidance, if I back out eSilicon, looks like core inside, growing 9% sequentially, up 36% year-over-year. Could you just help us understand what's driving the quarter-on-quarter growth. Is it both your cloud and Telco businesses? I assume it's PAM for with the ramp of 100-gig Porrima for 400-gig into your second cloud customer, but whether the other product categories that are driving the sequential growth.

[John Edmunds:]

Harlan, it's a combination of different things, but obviously the cloud area is expected to grow through the course of the year. Yes, we do expect continued strong COLORZ business in the year. And long haul and metro, I think will also, still be lumpy, but it will have probably two good quarters and two flattish or slightly negative quarters in terms of sequential growth, just like it had been this past year. But overall, we'd expect growth in long haul and Telco for the year as well as growth coming in the cloud and data center. And obviously we'll layer eSilicon in on top of that.

[Harlan Sur:]

Great. And then just my last question on the cloud and hyperscale side. You had one large hyperscaler ramping. Your 50-gig PAM for last year, that's a 200-gig Polaris. You’ve another large cloud customer ramping late last year, early this year with your 100-gig PAM for Porrima for 400-gig. How many other cloud and hyperscalers do you see layering on in terms of ramping with either Porrima or Polaris Solutions this year.

[Ford Tamer:]

So thanks, Harlan. This is Ford. We expect two more customers this year. One in the U.S. and one in China. And we expect all seven major account customers to go to PAM next year, including the remainder one in the U.S. and two more in China. So overall in four years, three big China cloud customer is going to PAM.

[Harlan Sur:]

Great. Thank you very much.

[Operator:]

Thank you. Our next question comes from the line of Quinn Bolton with Needham and Company. Your line is now open.

[Quinn Bolton:]

Hey, guys, congratulations on the nice results and outlook. I wanted to follow up on Harlan's question there. Obviously, the PAM business seems like it could drive a lot of the growth and upside in 2020. You've talked about the 50-gig Polaris, 100-gig Porrima. You've got Vega, you've got an IP licensing business. Is there any way you can give us some sense of how each of those contribute to your overall outlook? Are they all similar sized as the 200-gig Polaris, the biggest piece of the bucket? And then a follow up on the U.S. cloud and the Chinese cloud customer ramping this year. Will that be 200-gig PAM, or is that for the 400-gig PAM? Thank you.

[Ford Tamer:]

Thanks, Quinn. For the breakout, we don't break it up. But the way I'd characterize it is the two biggest contributors are the 50-gig PAM Polaris inside the margin and a 100-gig PAM Porrima inside the module. Both of them are each quite larger than the retimer business. Between the two, the 50-gig PAM is ahead this year, so we will see the 50-gig PAM Polaris for 200-gig be ahead of the 100-gig. Although the 100-gig is going to grow very nicely. As far as the new customers, one of them will go towards Polaris. The second one is still undecided and could go either Polaris or Porrima TBD, yet.

[Quinn Bolton:]

And then just a final question, just you sort of mentioned Huawei an 11% customer in 2019. Wondering if you had any other cloud or particularly module vendors that broke above the 10% threshold in calendar '19? Thank you.

[John Edmunds:]

Quinn, the other one will be Microsoft this year and they were in the mid to low teens range. I don't remember the specific numbers we sit here, but they will also be reported as more than 10% customer.

[Quinn Bolton:]

Thank you.

[John Edmunds:]

Obviously, because of the COLORZ business, right?

[Quinn Bolton:]

Yes. Yes. Okay. Thank you.

[Operator:]

Thank you. Our next question comes from the line of Ross Seymore with Deutsche Bank. Your line is now open.

[Ross Seymore:]

Hi, guys. Congrats on the strong end of the year, and the beginning of this year. I just had a question on your telecom portion of your business. I believe you talked about Huawei being strong in that barbell approach you said, John, with the end of the year finishing strong. But that segment overall, you said went down 8% sequentially. So correct me if I'm wrong, but if Huawei is a big portion of that and it was strong, I guess what was weaker in the quarter? Is it just lumpiness or is there anything going on in the fourth quarter there?

[John Edmunds:]

I don't know that I'd point to anything specific, Ross. I think for us it was just lumpiness in certain product areas that we think will return and show growth again in the first quarter. That's why I say that that business seems to seesaw back and forth. It will have a strong quarter and then a little bit weaker quarter and then come back with a very strong quarter again. So, overall we'd expect I think we've had growth of about 18% for the year and we'd expect something a little better than that for next year. And there's just a mix of things moving back and forth. But our platform growth are going to continue to grow. The coherent DSP in particular, and some of the other components, the 5G business, also grew nicely this year, it's expected to grow again next year. So there's several good growing pieces there. And then the more sort of germane components that we've sold in the past that we've done well on, we're looking to collect together on platforms and sell in chipsets and platforms more effectively, just as we do in the data center and also in the coherent platforms.

[Ross Seymore:]

Thanks for those details. I guess my second question, I don't know if this is for you or for Ford, but as we think about eSilicon, you've talked about 40% of that business, at least initially being in your legacy segment. So I guess two parts to this. One, could you parse out even a little bit how the remaining 60% of the business in the first quarter is going to parse between your two other segments? And then much more importantly, as we go throughout the year, as we work from this roughly $20 million number towards that total year revenue level at about 100. Can you just talk a little bit about the lumpiness of that and kind of the linearity of it back half versus front half? Because I know that business has a bunch of company specific design wins that will kick in at different times and probably doesn't look perfectly linear.

[Ford Tamer:]

Yes. Thanks for the question. We do expect eSilicon to sequentially grow again every quarter for the remainder of 2020. And to step up really in the second half as the production ASIC at both the data center and the 5G go to production of midyear. So -- and then furthermore, we expect this to step up another step up into 2021, if we would have a full year of production for those ASICs. As we -- when we close the acquisition of eSilicon, we did say that the revenue growth would be accretive to our growth -- from 2020 to 2021. At the time, we said we'd expected the revenue from eSilicon to grow about 50% from 2020 to 2021. So as you could see, we see sequential growth this year and we see sequential growth year-over-year.

[Ross Seymore:]

Perfect. Thanks, guys.

[John Edmunds:]

Ross, just to try to answer your question. Initially, I think we'll see more 5G business come on and then back half a year it'll be split relatively evenly between datacenter and 5G.

[Ross Seymore:]

Perfect. Thanks, John.

[John Edmunds:]

Yes.

[Operator:]

Thank you. Our next question comes from the line of Tore Svanberg with Stifel. Your line is now open.

[Tore Svanberg:]

Yes, thank you and congratulations on the record results. Ford, could you talk a little bit about COLORZ II versus 400ZR. Just so I understand this correct, COLOR -- COLORZ II will still be only for your largest customer, right? Well, 400ZR is basically sampling to everybody.

[Ford Tamer:]

Yes. Thanks, Tore. COLORZ I is really the name we gave our 100-gig product and we're calling this COLORZ II to stress the fact that there's a good amount of leverage that we can have on the firmware and some of the labs and interop testing that as we move forward to 100-gig ZR to the 200-gig ZR. COLORZ II is the name for the module. Canopus really is the name for the DSP. And so what you're saying is, will Canopus -- you're suggesting Canopus is having a larger adoption just because it's going to go more customers, it go to more module customers outside of Inphi or go to more OEM customers as well. And so from that point of view, yes, Canopus may have a larger adoption compared to COLORZ II. COLORZ II, though, is going to have a higher ASP. So did I answer your question?

[Tore Svanberg:]

Yes, it did. That's very helpful. Second question is, I know you're sampling your 7-nanometer coherent DP right now. Any further thoughts on that on the PAM4 7-nanometer product?

[Ford Tamer:]

We have plenty of thoughts on that, but we haven’t announced them yet, Tore. So stay tuned. I think we'll have some news very soon.

[Tore Svanberg:]

Okay. I'll try another one then. So obviously the eSilicon acquisition is not just important for revenue, but having custom silicon and 2.5D packaging technology is important as well. Could you just talk a little bit more about how that positions Inphi not just here for 2020 and 2021, but perhaps, from a more scalable offering and where the landscape is today, with all of the different acquisitions going on, that would be really helpful. Thank you.

[Ford Tamer:]

Yes. Thanks, Tore. So the eSilicon acquisition gives us a tremendous amount of flexibility because it brings a lot of resources to projects where the customers are requesting for us to customize and we didn't have in the past or the resources required. So. So right now it gives a lot of flexibility in potentially taking on incremental projects for these cloud vendors or telecom vendors that are very strategic in nature.

[Tore Svanberg:]

Great. Thank you.

[Operator:]

Thank you. Our next question comes from the line of Richard Shannon with Craig-Hallum. Your line is now open.

[Richard Shannon:]

Hi. Ford and John, thanks for taking my questions as well. Ford, you mentioned you’re expecting better share in PAM4. Not sure if that was related to the entire year or just the start of the year, but maybe you can give us a sense of why that is. I know your first to market. Are there reasons of interoperability or competitors not executing or just understanding what's going on there.

[Ford Tamer:]

Yes. Thanks, Richard. So the first reason is you heard about the different variance of PAM. So obviously there is a 50-gig PAM DSP, a 100-gig PAM DSP, you got a 50-gig PAM retimer. We've got IP and we've got a portfolio of 28-gigabaud TiA driver and a portfolio of 50-gigabaud TiA driver. For quite a few different optics form factor, so some customers are taking to market, some other are taking directly modulated laser DML to market. Others are using externally modulated laser or EML and finally, silicon photonics. So that gives you quite a few variants of TiA and drivers for these four different optics variants. We are the broadest portfolio of anyone in the market. So in some markets we've got competition, in some markets we’ve no competition. So we enjoy in some market a higher share that so far has not been matched by competitor. And in the markets where we are matched by competitors, we have enjoyed, as you suggested and we suggested a first mover advantage, which has allowed us to capture this higher share. So I think it's a combination of a broader portfolio, a high quality product, phenomenal performance in power and then first mover advantage.

[Richard Shannon:]

Okay. Fair enough. Thank you for that. Second question is for John. John, you made -- in your prepared remarks, you talked about something related to gross margins that frankly went by very quickly, something about 50 and 100 basis points of help, I think, throughout 2020. Can you repeat that? And then I'll probably have a follow-up based on what I heard there, but I completely missed that.

[John Edmunds:]

Yes. Richard, basically what I suggested is that our gross margins, as we integrate eSilicon are going to step down in the near-term, but they will improve -- the total gross margins would improve 50 basis points by the end of 2020 and by something in the neighborhood of 100 basis points in 2021. That's based on a combination of new products, shipping coming out of eSilicon that have higher gross margins. And it's also based on some cost and price improvements that we expect to achieve through the course of the year on their existing base business.

[Richard Shannon:]

Okay. And to be clear on that, John, is this a measuring point of that should 100 basis points. Is that based on your first quarter guidance or some other measuring point?

[John Edmunds:]

It's based on the -- the improvement will be based on total gross margin improvement. So in other words, just those improvements should drive a total gross margin improvement at that level. And obviously they’re only going to be mid-teens kind of percentage of the overall business. So they would have to have much greater improvement in order to drive the total gross margin improvement. And I would say the course of the year, not just the first quarter.

[Richard Shannon:]

Got it. Okay. And that is all the questions for me. Thank you.

[John Edmunds:]

Thanks, Richard.

[Operator:]

Thank you. Our next question comes from the line of Fahad Najam with Cowen and Company. Your line is now open.

[Fahad Najam:]

Thank you for taking my question. If I can start with a very quick first question on, what was the revenue in the year from your Metro DSP business, the M200? And how should we be thinking about that business in calendar '20?

[Ford Tamer:]

So I will let John give you the number, if he wants to, but I think we increased the business very significantly from '18 to '19, and then we see still a very significant improvement from '19 to '20. John, do you want to give more specifics?

[John Edmunds:]

Yes, the business overall, Fahad, I believe increased by around 84%. But we don't break it out separately, so we're not going to give you the details.

[Fahad Najam:]

Got it. And you see the same traction going forward in calendar '20 as you saw in '19, directionally speaking.

[John Edmunds:]

Yes, we continue to see strong business. And you know would it be again, something I’m probably in a order of 50% improvement?

[Ford Tamer:]

Got it. Ford, if I may ask you a big picture question, specifically regarding both PAM4 on one hand and also on the ZR side, we see a divergence in both these end markets, but ZR, Acacia and group of pushing open ZR. and on the PAM4 side, we have the Open Eye expect being pushed by your competitors. To what extent do you think that this impact the total PAM opportunity for you guys? And if you could comment on both the dynamics that you're seeing from here, especially in the PAM4 side first and then on the ZR, how do you see that play out?

[Ford Tamer:]

Yes, I think these are --- Thank you. Very good question, Fahad. These are in our mind, two very different trends, right. On the coherent side of the world, we're absolutely committed to Interop and come to a joint forward error correction that would standardize across the industry. So we're working frankly very closely whereas both Acacia and to come to a joint proposal. In the short-term our CFEC has been adopted in quite a few standards. And I believe Acacia and NEL will be putting them in their products. Their oFEC has been adopted as well, and we committed to put in our products. So in the short-term, we're going to be cross-licensing each other and we'd both put the FEC in the products. However, longer term, we do believe that it's in the industry interest to standardize on one fact. So we don't end up having to do twice the work and the customers are not having to do twice the firmware. And so, longer term, we're committed to an interop and getting through hopefully one standard and we've got a very good relationship with all the different players in the ecosystem, including Siena and others and Nokia and others. Huawei they do these facts. So I think what we're seeing eye to eye with many of the industry players towards standardizing and the cost of these DSP right now, Canopus at 7-nanometer. The next one is going to be 5-nanometer. The cost of doing DSP is going to start approaching $80 plus million. And it's not sustainable to have all these different standards compete and increase the costs for everyone, right? So I think you're seeing a very good convergence there. I think there is a very good working relationship and a commitment to get to Interop longer term. And that's good for everybody in the market. On the PAM side of the equation, you see companies like Inphi and Broadcom and MaxLinear and others all following IEEE standard, which is what PAM is for. We are all collaborating and competing, if you wish in different places, but that's fine. Again, I mean, a standard is good for the customers. That Open Eye is a corner case where, you've got some vendors that don't have enough strengths in their solution to close the link and they define some kind of weird, nonstandard book-ended solution that you'd have to have the products on both ends. If you put somebody else's product on the other end, it's not going to work. And that's really just a artifact of the fact that their solution is not strong enough to be able to be a standard solution. So there again, we believe that this Open Eye is just a non-starter and doomed to fail. We don't believe this is going to go anywhere.

[Fahad Najam:]

Thank you very much. That's all for me.

[Operator:]

Thank you. This concludes today's question and answer session. I would now like to turn the call back to John Edmund for closing remarks.

[John Edmunds:]

Thank you, Sarah. Ford and Vern and I would like to thank you for joining us today and let you know that we plan on attending the Morgan Stanley Conference in San Francisco on March 2nd, and we'll be at OFC in San Diego on March 10 and 11. Although we caution you that OFC is primarily an industry and trade conference, not an Investor Conference, per say, but we are happy to talk with investors while we're there. And again, Ford and Vern and I would like to thank you for joining us today. We look forward to speaking with you again in the future.

[Operator:]

Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.