Maxlinear Inc (MXL) 2017 Q4 法說會逐字稿

Greetings, and welcome to MaxLinear 2017 Q4 Conference Call. (Operator Instructions) As a reminder, this conference is being recorded.

I would now like to turn the conference over to your host, Gideon Massey, Investor Relations. Thank you. You may begin.

Thank you, operator. Good afternoon, everyone, and thank you for joining us on today's conference call to discuss MaxLinear's fourth quarter 2017 financial results.

Today's call is being hosted by Dr. Kishore Seendripu, CEO; and Adam Spice, CFO. After our prepared comments, we will take questions.

Our comments today include forward-looking statements within the meaning of applicable security laws, including statements relating to our first quarter 2018 revenue, gross margin, operating expense, tax expense, tax rate and interest and other expense guidelines as well as statements relating to trends, opportunities and uncertainties and various product and geographic markets.

These forward-looking statements involve substantial risk and uncertainty, including risks arising from competition, average selling price trends and numerous other risks outlined in our SEC filing.

Actual results may differ materially from currently forecasted results. For a detailed discussion of the risks and uncertainties potentially affecting these forward-looking statements, we encourage investors to review the section of our SEC filing captioned risk factors and our previously filed Form 10-Q for the quarter ended September 30, 2017, and our upcoming Form 10-K for the year ended December 31, 2017, which we expect to file later this month.

Any forward-looking statements are made as of today and MaxLinear has no obligation to update or revise any forward-looking statements.

The fourth quarter 2017 earnings release is available in the Investor Relations section of our website at maxlinear.com.

In addition, we report certain historical financial metrics, including net revenues, gross margins, operating expenses, income or loss from operations, pretax margin, effective tax rate, net income or loss and net income or loss per share on both GAAP and non-GAAP basis.

We encourage investors to review the detailed reconciliation of our GAAP and non-GAAP presentation in the press release available on our website. We do not provide a reconciliation of non-GAAP guidance for future periods because of the inherent uncertainty associated with our ability to project certain future charges, including stock-based compensation and its associated tax effects.

Non-GAAP financial measures discussed today do not replace the presentation of MaxLinear GAAP financial results. We are providing this information to enable investors to perform a more meaningful comparison of our operating results in a manner similar to management's analysis of our business.

Lastly, this call is also being webcast and a replay will be available on our website for 2 weeks.

And now let me turn the call over to Kishore Seendripu, CEO of MaxLinear.

Thank you, Gideon, and good afternoon, everyone. Thank you all for joining us today. We are pleased to report Q4 2017 revenue of $113.7 million, which is up 31% year-over-year and flat quarter-over-quarter.

During Q4, we witnessed strength in both connected home and industrial and multimarket revenues, which were offset by weakness in infrastructure application.

In 2017, we posted a record annual revenue of $420.3 million and delivered non-GAAP growth in operating margins consistent with our long-term target model, which included the effects of 2 acquired businesses, with historically lower margin than MaxLinear's target levels.

I would like to note that Q4 2017 brings to close one of the most transformational years with MaxLinear. During 2017, we started to realize the benefits of our strategic initiatives related to market diversification and target market expansion, which we embarked on a few years ago.

In 2017, our infrastructure business grew by more than 90% year-over-year, which offset a double-digit decline in connected home, owing to a $65 million revenue headwind from legacy Entropic revenues.

We also added significant revenue diversity with the addition of Exar's high-performance analog products consisting of power management and interface product serving industrial and multimarket applications.

Based on Q4 2017 results, connected home, infrastructure and industrial and multimarket contributed approximately 58%, 19% and 23% of overall revenues, respectively. We expect our revenue diversification to continue, aided by anticipated strong infrastructure growth, driven by our industry-leading [fine] products for cloud data centers and 4G, 5G wireless networks.

Undoubtedly, 2017 was a significant year in the company's history for both engineering execution and market expansion perspectives. Our organic strategic developments achieved significant milestones.

We started mass production of the industry's first 28-nanometer CMOS RF transceiver, spanning the 5 to 45 gigahertz frequency range addressing the microwave wireless backhaul market.

We taped out a recent example, the industry's first 400-gigabit per second PAM-4 SerDes optical transceiver IT device, or Telluride platform, for the hyper-scale cloud system data market. With the sampling of our Telluride platform, consisting of the 400-gigabit PAM-4 DSP plus integrated driver and TIA solution, we believe we are in a leadership position to address the next major inside-the-data-center upgrade cycle for data speed and capacity.

We also taped out the industry's first Full Duplex, or FDX, cable fiber node infrastructure remote PHY system-on-a-chip device, which will enable 10 gigabits per second data rate services over coaxial cable or cable MSO subscribers.

In addition, we have secured Tier 1 OEM engagements for our Blackcomb 40-nanometer CMOS massive MIMO wireless access radio solution targeting 5G macro base station and 5G active antenna systems.

In wireless infrastructure, we are making great progress towards becoming the end-to-end network infrastructure solution provider for 5G wireless rollout.

At the 2018 Mobile World Congress this month, we have several joint customer product announcements and demonstrations relating to wireless access, fronthaul and backhaul solutions for 5G wireless network infrastructure deployment. These include our 5-gigabit per second next-generation modem SoC for 5G wireless backhaul and fronthaul transport using the microwave frequency band; our 20-gigabit per second virtual fiber Ethernet modem SoC, which allows operators to use cheaper and more readily available coaxial cable in place of deploying expensive new fiber for emerging 5G wireless fronthaul applications; and the [MxL210x] wireless RF transceiver plus modem single-chip, providing point-to-point, 2.5 gigabits per second connectivity link using MaxLinear's AirPHY technology, which solves both the link budget problem as well as the self-installation challenge at home to enable fixed broadband wireless access to the home.

As you can see, we are really excited by our expanding technology footprint. We solidified our position as one of the industry's leading CMOS analog and mixed single communication IC platforms.

Moving on to some of the fourth quarter's notable business highlights. Our connected home revenues increased approximately 1% sequentially, generally consistent with seasonal patterns. Within our connected home, cable data stepped up after a seasonally weak Q3, accompanied by strong growth in satellite digital outdoor shipments to international operators, along with continued growth in G.hn home connectivity and seasonal strength in hybrid TV tuners. These strengths offset revenue weakness elsewhere in our connected home business.

Specific to our connectivity portfolio, we have a strong growing design win pipeline for G.hn opportunities validated by a strong operator adoption of our Wave-2 technology.

G.hn is emerging as a critical high-bandwidth wired backbone connectivity solution, particularly outside North America across a wide range of telco, automotive and industrial end markets.

Our infrastructure revenues were down 11% sequentially but grew by nearly 60% year-over-year on a quarterly basis. As stated earlier, for the full year 2017, our infrastructure revenue increased by more than 90%. Our wireless infrastructure revenues were down in Q4 2017, owing to seasonality and weakness in telecom operator deployment. However, we are encouraged to see the first meaningful RF backhaul transceiver revenue from a Tier 1 OEM in Q4 2017.

We are in the early stages of increased industry activity leaning into the 2019 5G wireless investment cycle, with the last mile wireline access infrastructure, namely c.Link and G.Now, retreated from record highs in Q3 2017 but experienced a strong 2017 overall.

In 2018, we expect strong growth in our last mile wireline access business. In our high-speed optical business, despite a strong Q4 relative to its slow Q3 base, the stagnant Chinese metro market and delays in the ramps of renewed TIA and driver design wins, owing to excess inventory in the channel, are proving to be revenue headwinds.

In the long-haul and metro optical market, we continued to broaden our portfolio to include 45 gigabaud and 64 gigabaud TIA end driver products and believe we are gaining customer design win traction relative to our competition.

While we are very excited about our growth prospects in infrastructure, many of our new initiatives are yet to ramp to critical scale and we are not immune to the timing-related uncertainties associated with large infrastructure build-outs.

Lastly, our industrial and multimarket revenues increased 8% sequentially, driven primarily by strong demand across a broad range of interface and power devices.

We continue to believe that Exar provides a large and exciting opportunity to expand our analog content across a range of communications in industrial and multimarket platforms. We expect it to prove to be an ideal vehicle for continued end market diversification as well as serve as a conduit to increase silicon content on our existing platforms.

With that, let me turn the call over to Mr. Adam Spice, our Chief Financial Officer, for a review of the financial and our forward guidance.

Thank you, Kishore. I will first review our Q4 2017 results and then briefly discuss our outlook for Q1 2018.

On revenue of $113.7 million, GAAP and non-GAAP gross margins for the fourth quarter were approximately 45.8% and 62% of revenue, respectively. This compares to GAAP gross margin guidance of 47% and a non-GAAP gross margins range of 61% to 62%.

The underperformance relative to the GAAP guidance was due to higher-than-estimated acquisition purchase price accounting impacts, and the overachievement to the midpoint of our non-GAAP gross margin guidance was due to increased supply chain efficiencies.

The delta between GAAP and non-GAAP gross margins in the fourth quarter was primarily acquisition-related, reflecting the amortization of $9.7 million of inventory step-up, which was completed in Q4; $8.5 million of purchased intangible assets; $100,000 in depreciation of stepped-up of acquired assets; and $100,000 of stock-based compensation and stock-based bonus accruals.

Q4 GAAP operating expenses were approximately $57.8 million, which was $800,000 above the GAAP guidance with the overage primarily related to the Exar acquisition-related restructuring charges and higher prototyping costs and activities related to the sampling of our 400 gigabit per second PAM-4 SerDes device.

GAAP operating expenses included amortization of purchased intangible assets of $8.9 million, stock-based compensation and accruals related to our stock-based bonus plan of $7.5 million and $2 million, respectively, $800,000 in restructuring charges, $300,000 in depreciation related to the step-up in acquired fixed assets and $100,000 for acquisition and integration costs.

Payouts under our 2017 performance bonus plan are expected to be settled primarily in shares of MaxLinear stock, which are expected to be issued in Q1 2018.

Non-GAAP OpEx was $38.3 million, slightly above our prior guidance of $38 million and up approximately $400,000 sequentially, with a slight overage to our guidance related to higher prototyping expenses related to the previously referenced 400 gig PAM-4 SerDes device.

Rounding out our commentary on operating expenses. At the end of the fourth quarter 2017, our headcount was 753 compared to 772 at the end of the third quarter of 2017. We continue to evaluate our staffing levels globally, particularly following our recent acquisition activity, to strike a balance between driving near-term operating leverage and staffing key long-term growth initiatives.

Moving to the balance sheet and cash flow statement. Our cash, cash equivalents and restricted cash balance increased $300,000 to approximately $74.4 million. Our ending cash position reflects the effect of $20 million in debt prepayments during the quarter towards our term loan. When including an additional $10 million prepayment that was made earlier in Q1 2018, it brings the total prepayments to $80 million so far and brings our loan balance down to $345 million.

Our cash flow generated from operating activities in the fourth quarter of 2017 was approximately $21.7 million versus $37.7 million generated in the third quarter of 2017. The sequential decline in cash flow generated from operating activities was largely attributable to less revenue linearity in the quarter and decreases in deferred revenue, resulting from the increased distributor channel sell-through and a catch-up in a large customer's rebate payments that were accrued for but unpaid at the end of Q3 due to that customer's internal processing delays.

We continue to focus on deleveraging aggressively, targeting an operating cash balance of approximately $75 million with any excess cash generation above this level deployed towards debt payment.

Our days sales outstanding for the fourth quarter was approximately 53 days or 8 days less than the prior quarter.

Our inventory turns increased to 4.2 in the fourth quarter compared to 3.5 turns in the third quarter and are a focus of our ongoing Exar integration efforts to better align with MaxLinear's target model of approximately 6 inventory turns.

That leads me to our guidance. We expect revenue in the first quarter of 2018 to be in the range of $110 million to $114 million. Built into this range, we expect connected home revenues to increase approximately 1% sequentially and account for roughly 61% of overall revenue; contribution from infrastructure to decline approximately 1% to represent 19% of overall revenue; and industrial and multimarket to be down approximately 13%, contributing approximately 20% of overall revenues.

More specifically, within connected home, we expect cable data, cable video and G.hn connectivity to be up sequentially, offsetting declines across satellite and terrestrial TV [tier] solutions.

Within infrastructure, we expect sequential growth to resume in wireless backhaul, which will be more than offset by sequential declines in optical, given persistent China macro issues as well as softness across a range of power and interface applications.

Within industrial and multimarket, we are witnessing general softness across a range of power management and interface solutions.

We expect first quarter GAAP gross profit margins to be approximately 54% (Sic-see press release "53%") of revenue and non-GAAP gross profit margins to be approximately 63% of revenue. As a reminder, our gross profit margin percentage forecast could vary plus or minus 2% depending on product mix and other factors.

We continue to fund strategic development programs, targeted at delivering attractive top line growth as we look forward in the first half of 2018 and beyond, with a particular focus on infrastructure initiatives and our goal of increasing the operating leverage in the business.

As such, we expect Q1 2018 GAAP operating expenses to decrease approximately $1.3 million quarter-on-quarter to approximately $57.5 million, with the largest decreases coming from lower restructuring, amortization of intangibles and accruals related to performance-based equity. We expect Q1 2018 non-GAAP operating expenses to be up $1.2 million sequentially to $39.5 million, with increases driven by seasonal payroll tax and vacation step-ups, partially offset by lower projected prototyping expenses.

We expect GAAP tax expense to be approximately $500,000 and a non-GAAP tax rate of 7%.

We expect interest and other expenses in the quarter to be $4.2 million.

In closing, we're pleased to announce Q4 2017 results, a quarter in which we delivered year-on-year revenue growth and produced $21.7 million in operating cash flow, which enabled a further $20 million of deleveraging.

As we look back at 2017, we are encouraged by our ability to weather the legacy Entropic revenue declines and China macro headwinds in the optical market while adding strategic technology platforms in high-performance analog and G.hn through our corporate development efforts.

As noted earlier in our remarks, we're excited about our growth prospects across a range of markets, but we are in the early stages in many of these growth initiatives and predicting infrastructure ramp timings can be challenging at best. As a management team, we are committed to practicing continued restraint on near-term expenditures to ensure maintaining strong cash flow generation and preserving operating leverage in 2018 as we progress through our ongoing transformative investment in growth process.

As we move forward in 2018, we are encouraged by the diversity and depth of our product portfolio as well as the continued efficient acquisition that our company has demonstrated. We remain confident that our recent acquisitions, combined with organic the initiatives highlighted earlier, uniquely position MaxLinear shareholders to benefit from the growing demand for bandwidth across consumer, connected home, wired and wireless infrastructure networks and a diverse growing demand for high-performance analog and mixed-signal solutions across industrial, automotive and the multimarket applications.

And with that, I'd like to open the call to questions. Operator?

(Operator Instructions) Our first question comes from Ross Seymore from Deutsche Bank.

I guess the first one is, what's going on in the industrial and multimarket business? We haven't really had to deal with the seasonality of that before, but the down 13% guidance is a bigger step down than I would have expected. So any color there would be helpful.

Hey, Ross. This is a -- it's also a new category for us, the industrial and multimarket. So we're told there's some seasonality in the business, but it's -- given the nature of the business, so much of it goes through distributors. It's very hard for us to see this as anything but a seasonal impact rather than any end market demand issues. So we don't worry that it is down, as we have talked about it. However, it's also new for us, but we don't see any end market issues other than the fact that it's a new business for us.

And I guess as my follow-up, you mentioned a couple of different times about the difficulty in projecting the infrastructure side with these big projects, and I think all of us appreciated it. But is there -- I guess what's the underlying tone that you're trying to convey there? Is there something that has changed in that segment of your business as you look for 2018? Or is it just something where you're being more prudent after what happened in the high-speed interconnect market in aggregate, nothing specific to you guys, in 2017?

We are being cautious, of course. And there are 2 elements to our infrastructure business, right? One is the optical side of the business and the other part of the business is the wireless infrastructure backhaul business. And then we have our wireline last mile access infrastructure business. If you look at those 3 elements, they're very different. However, the wireless infrastructure business and the wireline access business have a number of customers leading to ramp to new product designs in 2018 and through 2019. So given the number of ramps that are there in various customers, we are being cautious because in the past, they have moved. On the optical interconnect side, obviously, what's happened to the China telecom market on the optical side is still quite stagnant, the marketing, and we are not seeing yet any relief there. However, the big movement for us is an opportunity to be the leader in the 400 gigabit per second data center market by being the first provider and sampler of a really working silicon. And that, if everything goes well, starts ramping really at the end of this year. However, there's a slim possibility that moves into the beginning of the next year. So those are dynamics. We are just being cautionary about the market because there are so many product ramps involved.

Yes, Ross, and I would add some to that, too. I'd say that we also commented on the fact that we're early in the phase for a lot of these initiatives for ourselves. And so we have a little bit more lumpiness to our business because we don't have enough customer diversity yet to kind of weather when one customers starting to ramp and then they pause for a little bit before the other one begins their ramp. So I think there's just a function of your markets take longer and then it takes a while to build up some customer diversity to give you some predictability to your revenue forecasts. And I think also, I would also -- we also deal with some competitors in our markets that tend to be a little bit more aggressive on when these markets are going to be develop. And I think you can attest over the years that we tend to be pretty conservative when it comes to predicting when the real need occur -- happens in these markets. And we don't want people to necessarily expect quicker ramps than we think are realistic. So again, I think some of our competitors may be a little bit over their skis as far as the timing of some of these new major technology ramp timings, and we're just trying to be a little bit more conservative.

And I guess one follow-up on that is just that competitive landscape, Kishore that you see on the single lambda PAM-4. I know you guys were taping that out either late last year or earlier this year. So just talk about your progress in that. I know you can't control exactly when the customers are going to ramp it, but how did the tape-out go? And how are you feeling about the competitive landscape in that market?

We are really excited. I think we have made good on our promise that we will be the first one with a real product that meets the power and performance requirement of a very stringent standard for QSFP-DD and OSFP sockets. We are in all our Tier 1 and optical module partners and their customers. They are all engaged in the ecosystem. We have sampled the product. We have looked at the performance that indicates the right bandwidth. We feel very, very good. And we are very optimistic when I say this right now, that by the OFC, we should be able to show you some good product performances. We're actually very, very excited not only because it's a form of deep market. There was an enormous amount of risk we have taken when we entered this market. We were the ones who are being the most aggressive that the leapfrog technology is really 400 gig, not some intermediate points between 100 gig and 400 gig. And I think we'll be able to validate what we set out on. And right now we're pleased we're in the frontrunner status with respect to this hyper scale data center market for 400 gig -- gigabit speeds between servers, racks and routers.

Our next question comes from Gary Mobley from The Benchmark Company.

Speaking of your 3 pillars of long-term infrastructure growth, I was hoping that you could share with us your expectation for revenue contribution for the cable fiber node product. And then related to ASC 606, can you remind us what percentage of revenue goes to distribution channel and whether or not there's any revenue benefit in your Q1 guide from revenue recognition change based on the sell into the channel?

So the first part of the question, Gary, was for you what are the various related contributions for our infrastructure revenues from the 3 pillars of growth we have talked about. One is wireless infrastructure, the other one is the optical data center market and the third one is the wireline last mile access. Here it's cable fiber node Full Duplex modified chip. So if you really look at these 3 products, while I cannot talk to you about the immediate relative contribution, but I can speak to you in the -- when the product cycles are fully set in, right? We expect our wireless infrastructure revenues over the next 3 to 5 years to grow somewhere between $100 million and $200 million. Our cable fiber node product really is somewhere between $50 million to $100 million. And our optical should also be in the range of about $100 million to $150 million revenues for data center. So those are the relative size of these markets. We could think of our wireless in the really long term being the #1, the data center market being a close #2 and the cable fiber node being a solid $50 million to $100 million contributor, depending upon market share and our success in being the leader in the marketplace. Having said that, we also taped out our cable fiber node in Q4 last year, and it's actually right now, in the lab, and we're going to commence testing. And this product being successful in being able to demonstrate the Full Duplex cable fiber node capability, we'll be by far the leader in making the cable fiber node for 10 gigabit per second data services for cable MSOs a reality in the very near future.

Okay. Gary, if that answers your question on the fiber node, I'll take your 606 question.

Sure.

So 606. A percentage of our revenue that goes through our distributor, it's about -- call it the mid-20s, mid-- okay, let's say 25% to 27% revenue goes through the distributor. And if you think about the impact for Q1 of 606, yes, there certainly is a possibility for more -- for some benefit from a channel refill. I would think, obviously, given the way that 606 treats what would have been deferred revenue on the balance sheet going into 2018, there certainly was motivation for us to lean down the channel going into the end of the year because any of that deferred revenue that remained on the balance sheet at the end of the year will then be recognizable as revenue. Well, it ultimately did ship through, because as you would recall MaxLinear has been on a sell-through model. With 606, we moved to a sell-in model. So yes, there probably will be some channel refill opportunity in Q1, but we think it's a relatively short-lived and probably relatively minor impact. It's not that big. So I don't think it's a real needle-moving event for us.

Just to be clear on that, that benefit, if any at all, is not included in the guidance?

I would say we took a very neutral view on whether or not the 606 kind of channel refill impact would really kind of move the needle for us in Q1. So I'd say we took a balance view that there might be a little bit but really not a lot.

Our next question comes from Tore Svanberg from Stifel.

I just wanted to follow up on the infrastructure, and I perfectly understand the conservatism. But just wondering with some recent conversations you've had with customers, especially given the new tax policy and so on and so forth, would you say that there's a little bit more discussions now for infrastructure deployments?

So Tore, you asked a very really, really macroeconomic question. We do not have those kinds of conversations with our customers. We sell really what I call very nerdy, analog, mix, single communication DSP products, and we are typically not in conversations with CFOs or our customers. So I would like to say that that's true, they're going to invest more because of the tax policies. But at our levels of conversation, we're really focused more on performance and platform opportunities than timing of ramps.

One thing I can say is, we've talked about this before. We've definitely seen a lot more activity around our infrastructure initiatives, particularly around this 400 gig PAM-4 opportunity. I would say, as you know, historically MaxLinear has not been a field-of-dreams company, where we build it, and they will come. But there was a little bit of that with this PAM-4 because it was a strategic new focus for us as a company and we were a new entrant to the market focused on the metro and the long-haul through acquisition of Physpeed. And so for us, we really needed to lean forward and take a few gambles on this marketplace. And I would say that over the course of the last 6 to 9 months, customer engagements or prospective customer engagements have really ramped up quite a bit. And particularly, after we now have sampled the product, I think that's taken another level up in activity and interest. And I would say that it's really confirming and validating the risk that we took when we decide to lean forward and invest in this market as fast as we did and end up being what appears to be a first-to-market with the productizable solutions Kishore mentioned earlier. So I think that's really what we're more focused on than kind of -- we don't see, as Kishore said, really these macro level impacts of increased domestic investment. It's really more product and project-focused. And that, we have certainly seen a very significant step-up in activity particularly related to the PAM-4 and also for the -- our wireless optic, particularly the 5G massive MIMO antenna arrays and so forth. So we feel very comfortable about those signs of investment growth for us.

That's fair. And a follow-up on Telluride. So what are going to be some of the tangible milestones that we should track? Obviously, you're sampling the product now. You talked about OFC. But as we move throughout the year, what are going to be some of the milestones? And will you at all be able to actually announce design wins for this product?

A very good question, Tore. And let me walk you through what are the tangible milestones. We hope and expect that you guys will come to our OFC booth, and we can actually show light entering the device, coming out of the device at the right bandwidth and data rate. That would be the first big milestone. And we will have press releases regarding this subject. Not only about our PAM-4 DSP and integrated driver but also our companion TIA device. So you'll see the full menu of demonstration, and you'll be hopefully able to go to 1 or 2 main module maker -- optical module makers and see a demo in a form fit function of our new SFP and QSFP-DD functions. So they're working very vigorously on that right now. The next step beyond that is obviously some level of customer-related product finish or entry or ramp announcements. That really will happen probably towards the end of the year, not any time now. However, there's a good chance that we may announce a partnership across an optical module maker and to the -- all the way to the end network systems provider. So that's what we're working on right now. So I would say those are the major milestones on the 400 gig PAM-4 product deployment from MaxLinear.

Very good. And my last question is on connected home. So I know, historically, you target this to be sort of a high single-digit or 10% growth segment for you. As we look at '18, are we thinking about that in the same vein? And you haven't mentioned, DOCSIS 3.1, so I was just wondering if it could add some color on your growth prospects for connected home this year.

One thing, just to clarify a little bit, Tore. I think in our recent investor presentations, we've modified that connected home growth to more like 5% to 10% as the range that we've taken. And part of that is just because of the uncertainty, again, of the ramp timing of things like DOCSIS 3.1. But also, I think it's really more of a function of the fact that we've grown a fairly significant satellite business. And as we all know, satellite -- the domestic satellite market is definitely having some challenges as far as subscriber losses and so forth. So I think that's partially what informed our kind of taking the range from 10% -- or the spot to 10% down to a range of more 5% to 10%. And right now I don't think there's any reason at this point in the year why we would change that view or what that range should be for the connected home. Because again, we do have the DOCSIS 3.1 should be a good space for us because of our competency in 3.0. Combine that with the nice growth that we talked about coming from our G.hn acquisition that Kishore mentioned earlier. I think that the more question mark in that connected home portfolio is really the satellite, which is more macro-driven. And of course, we've also -- we talked in the past also about some operator on and off again behaviors with the MoCA deployments. So I think hopefully that provides some context. And Kishore, do you want to add anything?

Absolutely. That's absolutely correct. And I just want to point out the DOCSIS 3.1, one of the reasons we didn't mention it here is because while DOCSIS 3.1 ramps should be starting in earnest, they're not as far along as they should be. However, for MaxLinear's perspective point, the ramp you're talking about is at a major operator, right, cable operator. And for us, that's a very neutral effect on our shipment because we're such a large incumbent in DOCSIS 3.0. So for us, the start of a DOCSIS 3.1 ramp at this particular operator is not a very revenue-meaningful event. However, when the revenue on DOCSIS 3.1 ramps with some other operators outside of the biggest operator in North America, we should see some ASP pickup and that should help grow our revenues related to the cable DOCSIS.

Our next question comes from the line of Christopher Rolland from Susquehanna.

It's David Haberle in behalf of Chris Rolland. Maybe just to piggyback off on that last thought on the DOCSIS 3.1 transition. Can you talk about -- you said you're very early with the largest customer. Have you experienced any inventory build from them at all? Is that a contributor in 4Q or 1Q here?

I think you touched upon a very interesting topic of inventory. So I cannot comment about inventory buildup at the customer. We did ship a reasonable amount of backlog for this particular customers, but we're not seeing throughput at the rate we would expect normally. However, it has created some level of uncertainty about the amount of inventory that we would normally see in the channel for DOCSIS 3.0, and that's also tempering our forecasting process. So yes, we are in the midst -- in the middle of a transition with 3.0, 3.1. And what I call the non-spontaneous ramp to 3.1 has created some level of inventory transition uncertainties that is affecting our guidance as well. So I think that's a fair question. But I don't think there's an excess inventory buildup. You must keep in mind that there are 2 big vendors on the OEM side for the North American DOCSIS 3.1 market. So what effects may not affect the other companies and vice versa. So I think we are good with where we are positioned today.

And then just to change gears real quick to the force-touch edge sense business that you guys have. I think last quarter, you mentioned that it was mainly one design win that was driving the revenue there but that you had other design wins that will ramp. Do you guys have any better line of sight and visibility of when those ramps might occur? And then also, can you give us a sense if those design wins are with that same OEM or different OEMs?

Yes, David, I'm glad you asked this question about the force-touch. I think it's very well known that our force-touch design wins in the phones that have been in the market have been with HTC. And you have heard recently that HTC sold off a bunch of business to Google. And so that's created some uncertainty as to what HTC's future plans are. And the problem with a smartphone business has always been that way, right? And that's the reason we have become such a strongly operator and infrastructure-focused company and industrial multimarket-focused company. And that's at the heart of our transition when long ago a consumer company to now where we have positioned ourselves. On the other hand, we've had a couple of other design and prospects in progress on the OEMs. But I'm still not confident to tell you that we have won those sockets because the nature of the force-touch feature is such that it's a feature -- in phones, generally, it's a feature that's nice to have and it could come in and stay or can be removed at the last minute. And we have seen some level of skittishness on the OEMs on how they function. And more importantly, we are seeing a level of consolidations in the phones that is not necessarily obvious to the larger world, but there is a certain level of consolidation going on in the 4 makers. And so we are feeling pretty cautious about what I call emphasizing force-touch as a growth vector for us in this year or moving forward in the consumer handset market. Our hope is it stays. And if they're going to be investing in force-touch, it's really to expand its presence in the industrial multimarket applications.

Our next question comes from the line of Quinn Bolton from Needham & Company.

Guys, I just wanted to follow up, I think it was on Ross's question, regarding the sort of weakness you're seeing in the analog side of the business, power management and interface. Did you say that this is just a seasonal pattern that you sort of anticipate every year? Or is this is a broader and more macro-driven slowdown in that broad catalog business?

So Quinn, I really cannot honestly answer that question with the statement that we know what's happening at the macro level. Our understanding is typically, there is a step-down seasonality. However, we are yet to confirm that for ourselves because we have been in this position of this business for less than 1 year right now. So I wouldn't -- I would not stretch this to a macro impact. However, we see that this business, majority of it goes through distributors, so it's hard for us to forecast really direct and end customer demand to be able to answer your question.

Yes. I think, Quinn, I'll try and throw in a little more color on that. So I mean, as you know, it's a very diversified business, right? So there's a lot of different SKUs, a lot of different customers, a lot of applications. But if you really want to pick out the -- we talked about the questions raised earlier about HTC. So part of the challenge right now, as we look into Q1, is softness on that force-touch as that initial HTC design is getting a little bit kind of, I would say, aged, if you will. So with -- that's probably I'd say the leading contributor to the drag in that multimarket and industrial guide for Q1. And then just more broadly speaking on the industrial side, we're just seeing across a range of applications just a little bit of softness and it's difficult to forecast that business as well because it is so [DIS-D] forecast. You do have to rely on your distributors quite a bit more than what kind of the -- I would say the legacy method of bottoms-up forecasting by customer is. So I would say if you want to point to a couple more details again, it's more of the HTC force-touch impact in Q1 and then just broad-based kind of difficulty to forecast. And when we can't have as much granularity as we like to or we're used to as a management team, we tend to take a more conservative approach to guidance.

Got it. Do you guys have any visibility in the end customer inventory levels? Or does it kind of -- does your visibility stop with inventory held at the distributors?

It depends on the parts of our business. If it's a direct business like MaxLinear platform, these businesses are in the operator business, for example, in cable and satellite. We do make it a point to get direct information as much as possible fairly regularly about our OEM's inventory position, not necessarily the operator's inventory position. However, in the Exar portfolio of catalog-ish analog products that go mostly through distributors, we really track the distributor inventory and we don't have much visibility to the end market inventory because it's really several thousands of customers are buying those components. So I think -- so you can look at us -- 2/3 or 3/4 of our business being more platform-oriented, where we track our OEM inventories. And then in the non-platform businesses, where -- the distributor OEM, primarily Exar business, we only track distributor inventory. I would say that in 3/4 of the business that we have a direct knowledge off of the business, we do have a sense of inventory and it is what I call not away from the normal levels because on average I think it's fine. It's definitely not low inventory levels but it's not excessively high either. So I think it's about normal.

Okay. And I just I wanted to follow up on the single lambda PAM-4. It seems like there's pretty widely reported delays now for Tomahawk 3. That could push out by as much as 6, maybe possibly even 12 months. And that solution may not get into the market until sometime well into -- or systems based on Tomahawk 3 well in the market until 2019. And so I guess, as you guys look at the rollout for single lambda PAM-4, you talked about possible shipments by the end of '18, early '19, what does the 6 or 12-month delay in Tomahawk 3 do to that business? I mean, it seems like that's the big growth opportunity because TIAs, drivers, it just seems like that's kind of in a pretty challenged market and likely to remain so.

Quinn, a very good question. I'm hearing for the first time about delays to Tomahawk. But firstly, I looked at the sampling of the Tomahawk 3, as you called it, at the end of last year, beginning of this year as very, very positive. Firstly, that's confirmation that you do have a PAM-4 interface to which the 400 gig PAM-4 chip can interface to. So that's good news. Secondly, production readiness, we've always assumed, is going to take at least 12 months. And if it's delayed in terms of getting ready to production, that's one manner. The other matter is the chip itself, Tomahawk switch chip, does it have issues? From all accounts, I do not think it has issues. I'm not aware of any issues. So if there are natural lead time issues in getting ready for production, the best we can do is be prepared and ready to go to production and we will be. Anyhow, there's a lead time for optical manufacturers to get ready, too, as well in the sense that optical modules, the biggest -- the longest fold in production getting to production with these QSFP, OSFP chip module form factors is really the optics and it's not the silicon folks. So I feel that well, the delay may delay our ramps on the PAM-4 400 gigabit per second single lambda chips. However, we need to do what we need to do to be ready so that we are not the gate to going into production.

Maybe a clarification. Just on when you guys say first shipments or shipments in late '18, early '19, are those kind of production-ready -- is that a production ramp or is that kind of the preproduction qualification units that are going out to the module guys and kind of establishing the ecosystem? And then 12 months after those first shipments is when you would expect the volume ramps, so you may be kind of talking volume ramp in late '19, calendar '20 -- early calendar '20.

Okay, so let me talk of production ramp that is our production. Silicon production, really, silicon products being starting to ship in some level of volume quantity, what I call initial volume quantities. Regarding product qualification, hopefully our expectation is that we'll get that done before the end of 2018. Now our customers' product qualification could be dependent upon the optics also and the silicon all assembled together and being qualified. That's a different qualification cycle than our own products being qualified and production-ready. We are on track to ensuring that we will be production-qualified and ready to go as far as silicon is concerned.

Hey, Quinn, this is Adam. Just to give you a little more color. We don't have a lot of revenue currently forecast in our model for 2018 for Telluride. So that goes back to the commentary that we provided when Ross asked his question. We're very, very conservative. So a pushout of production of Telluride because of Tomahawk delays, what have you, really doesn't move the needle for us in 2018, right? So it's -- we've been very conservative on the timing. And that's why I said earlier, when I hear other people talk about ramps in the second half of 2018, other competitors talk about that, it just makes me a little bit nervous for what other people might imply for ourselves because again we've never been saying that, right? We've been saying that we would start some shipments in the second half 2018, late second half 2018. But it really wouldn't be needle-moving. And our perspective has not change on that. We've always taken a more conservative stance.

(Operator Instructions) Our next question comes from the line of Tore Svanberg from Stifel.

Yes, I just had some follow-ups and then some housekeeping ones. So first of all, G.hn seems to be doing really well for you. And it's interesting, right, because the investors are talking about your China optical exposure, but your G.hn business is probably 3x the size of that. So can you just explain a little bit what's driving that adoption and why you're seeing such a strong momentum there?

First and foremost, G.hn addresses a large part of geographies in the world because it works over powerline, coaxial and (inaudible). So that's a nice thing about it as compared MoCA, which is purely coaxial cable and primarily North America. So on the G.hn side, there is what we call the last mile access related version called G.Now, which we refer to collectively as G.hn. And the G.Now is being used by telecom operators because G.fast, that has not delivered the promise, not that they have the capability to provide the broadband for the last mile into the multidwelling unit. And in China and South Asia and whether it's Thailand or Korea and in other Asian countries, they are looking at G.hn or G.Now as a way to do the last mile access into multidwelling units. So that's a nice thing about it because typically in these countries, fiber comes all the way near the buildings, but once into the building, you really need something that's much more robust. And G.Now provides a very cost-effective way of doing it. Secondly, on the G.hn side, there are -- it is being used as a wired backbone for Wi-Fi coverage in the various countries in Europe and other Asian countries for Wi-Fi extenders. So those designs are going on well. And finally, you may have noticed that there's a big backing of G.hn from Huawei and other players in China, China telecom players, who are really showing their strong presence on the standards bodies. And they really get G.hn as a whole networking standard wired connectivity inside the home both for distributing video and also for extending Wi-Fi coverage inside the home, which is really terrible in China, for example. So those are the things that are really driving those -- the ubiquity of reception for deploying G.hn. Not forgetting to mention that you also have these utility companies that are using G.hn for metering applications, right? So I think just the range of applications that G.hn can address is the real secret behind it because it's a readily available powerline medium throughout the world big. So I think it's quite exciting. And it's really going very strongly. And in fact, we expect the growth to continue into next year and the following year as well.

That's really helpful, Kishore. And just some housekeeping ones for Adam. So Adam, high-speed optical, China, I mean I think last quarter that was pretty low. Is it still sort of sub 1%, 2% of revenue at this point?

Yes. Well, are you talking about Q4 or Q1 guide?

Q4, like your most recently reported quarter. Was it still sort of in that 1% to 2% of revenue?

It was around -- yes, it was in the 2% to 3% of revenue actually in the fourth quarter.

Okay, that's helpful. And then also the legacy businesses, both video SoC and analog channel stack switch, are they approaching 0 now or just sort of around 1%?

Yes, the video SoC was -- yes, we said it was going to be bouncing around the bottom, and that was around a little less than $1 million in Q4. And then the analog channel stacking was almost 0.

Ladies and gentlemen, we have reached the end of the question-and-answer session. And I would now like to turn the call back over to management for closing remarks.

Thank you, operator. As a reminder, we will be participating in the Morgan Stanley Technology, Media and Telecom Conference on March 1 in San Francisco, and at the 30th Annual ROTH Capital Conference on March 12 in Newport Beach, California and the Susquehanna Seventh Annual Technology Conference in March 14 in New York. We hope to see many of you there. With that said, we thank you all for joining us today, and we look forward to reporting on our progress to you for the next quarter.

This concludes today's conference. You may disconnect your lines at this time. Thank you for your participation.