Infinera Corp (INFN) 2017 Q3 法說會逐字稿

Welcome to the third quarter year 2017 investment community conference call of Infinera Corporation.

(Operator Instructions).

Today's call is being recorded.

If anyone has any objections you may disconnect at this time.

I would now like to turn the call over to Mr. Jeff Hustis of Infinera Investor Relations.

Jeff, you may begin

Thank you, operator.

Welcome to Infinera's third quarter of fiscal year 2017 conference call.

A copy of today's earnings is available on the Investor Relations section of our website.

Additionally, this call is being recorded and will be available for replay from the website.

Today's call will include projections and estimates that constitute forward-looking statements.

These may include statements regarding our overall business strategy and results of operations, market conditions, market and growth opportunities, views on our customers and products, expectations regarding the timing of new products, expectations regarding our restructuring plan and Infinera's financial outlook for the fourth quarter of fiscal 2017.

These statements are subject to risks and uncertainties that could cause Infinera's results to differ materially from management's current expectations.

Please refer to Infinera's current press releases and SEC filings, including our most recently filed quarterly report on Form 10-Q and subsequent filings, for more information on these risks and uncertainties.

Please be reminded that all statements are made as of today and Infinera undertakes no obligation to update or revise any forward-looking statements to reflect events or circumstances that may arise after the date of this call.

Today's earnings release and conference all include certain non-GAAP financial measures.

Pursuant to Regulation G, Infinera has provided a reconciliation of these non-GAAP financial measures to the most directly comparable GAAP financial measures in its third quarter earnings release, which has been furnished to the SEC on Form-8K and is available on our website in the Investor Relations section.

I will now turn the call over to our Chief Executive Officer, Tom Fallon.

Good afternoon and thank you for joining us on our third quarter 2017 conference call.

Joining me today are Brad Feller and Dave Welch.

Today, I will review our third quarter performance, progress on new products and technologies and also go over the restructuring plan we announced earlier today.

Brad will then provide a more detailed financial review of the third quarter, our outlook for the fourth quarter and additional details around the restructuring.

In Q3, I was pleased we delivered revenue at the high end of our guidance range.

Our 9% sequential revenue growth was primarily driven by ICE 4 products with ICE4 growth exceeding our internal forecast.

On the bottom line, while we're not where we need to be, I was satisfied with our execution and stronger revenue and tighter controls around our operating expenses resulted in EPS that exceeded our guidance range.

On the new product front, I am pleased that our CX2, XT and XTS-3300 and XTM2 are all now GA and are all impacting revenue.

As of the end of Q3, we had recognized revenue from 10 customers for our ICE4 platform, solidly up from 3 in Q2.

As expected, our ICE4 growth to date is being driven largely by ICPs, both existing customers and newer ones, such as recently announced Netflix, who are seeking best-of-breed solutions for their data center interconnect applications.

Also encouraging is early revenue and growing interest from wholesalers and cable providers, who are seeing advantages to supplementing their networks with the high-capacity cost-effective capabilities of the XT series.

Moreover, we see opportunities to grow market share in subsea with our XTS-3300 demonstrating the best spectrum efficiency of any commercially-available solution on the market.

Finally, we anticipate our metro installed base will take advantage of the density and power enhancements in the new XTM2 platform.

We expect our larger metro customers will begin to qualify XTM2 in Q1 and for revenue to ramp over the course of 2018.

Whether for ICPs seeking disaggregated solutions for point-to-point links or traditional service providers that value integrated multi-service capabilities to ensure the best experience for their end customers, we are delivering the most reliable, high-capacity, power-optimized solutions in form factors customers want to deploy.

Our solutions and approach are gaining traction as evidenced by Dell'Oro naming Infinera the market leader in disaggregated transponder units in its initial reporting of this category.

With nearly 1/3 of all DCI wavelengths being implemented in disaggregated systems today, we believe our portfolio and product roadmap align with the needs of both ICPs and traditional service providers.

Beginning to contribute to this dynamic, we are seeing an uptick in customers conducting field demonstrations around open networks.

As open becomes more pervasive, we see opportunities to sell our disaggregated transponders over competitors line systems, a dynamic that would open up new markets for us and we believe favors Infinera's vertical integration cost advantage on high-capacity transponders.

On the technology front, I am very pleased with the performance of ICE4.

In particular, our recent deployment of the Seabras-1 submarine cable, demonstrating the industry's highest spectrum efficiency in a commercially-shipping product.

This capability enables our customer, Seaborn, to utilize significantly more fiber spectrum than any current competitive offering, a material cost advantage for Seaborn and an important demonstration to the market that gives me confidence in our opportunity to gain share in subsea.

Additionally, we continue on the path of executing a 2-year optical engine cadence for both our ICE5 and ICE6 optical engines that includes advancements in both our DSP and PIC capabilities.

We recently demonstrated 2 industry firsts: 100 gigabaud and 1024QAM utilizing constellation-shaping algorithms.

These technologies give us a path to greater than one terabit per wave solutions and positions us well to achieve our goal of being optical performance and cost leaders.

Combining our unique ability to perform massive integration with the PIC and leading on the DSP with baud rate and modulation, we intend to provide customers the highest capacities possible across varied distances delivering differentiated optical performance and lower transmission cost.

Overall, I am pleased with the progress we are making on both ICE5 and ICE6 in the labs and look forward to providing updates on our new optical engines during 2018.

While we are seeing promising signs of our technology adoption and are now demonstrating a 2-year optical engine cadence, we also continue to experience near-term challenges.

First, our revenue has been under extreme pressure over the past year largely attributable to customer consolidation.

While you're now seeing business gradually improve from most of our consolidated customers, our expectation is spending from recently combined CenturyLink and Level 3 will remain suppressed for an undetermined period of time.

Second, while pricing declines have moderated relative to the historically steep declines of last year, ongoing price aggression by certain competitors and technology transitions continue to drive pricing down at a rapid pace.

Third, there are signals customer CapEx could be weaker than we had originally anticipated in 2018.

This expectation is supported by multiple industry analysts lowering their 2018 outlooks for the DWDM market outside of China, now expecting overall growth in the low- to mid-single digits with long-haul expected to be only flat to slightly up.

I believe the growing trend towards disaggregation is a key contributing factor to this outlook as customers increasingly transition to disaggregated architectures.

While this transition is a positive development for Infinera and our new portfolio, it comes with the challenge of having to sell multiple units to achieve the same level of revenue an integrated chassis-based sale would have historically achieved.

This dynamic and the possibility that we maybe experience some short-term pauses in adoption and deployments as customers transition to disaggregated networks, support our cautious outlook for 2018m but do not dampen our long-term optimism that we are well-positioned for this next generation of network architectures.

Lastly, on a company-specific front, the timing of completing the XT-3600 and 1.2 terabit line card for the DTNX, the final ICE4 products, has slipped one quarter to Q1 of 2018.

Given our expectations over the next few quarters, today we announced a restructuring plan that will cut cost to align our operating structure with current opportunities and improve future operating leverage.

Over the past 3 years, we have made significant financial investments and transformed into a multi-market company delivering a fully refreshed product portfolio and establishing a faster R-and-D cadence.

Perhaps inevitably, as we have reflected on this substantial evolution and internal expansion, we have identified areas across the company where we can be more efficient going forward.

This restructuring will include an R-and-D remote site closure, a rationalization of products and programs and a reduction in headcount across all functions of the company.

To effectively drive this streamlined operating structure and ensure we lead in optical technology, I have made some changes in our organizational structure.

First, Dave Welch will assume the role of Chief Strategy and Technology Officer.

In this capacity, Dave will renew his focus on creating differentiated capabilities that will drive our strategy.

He will also increase his external presence actively engaging with customers to align our product roadmap and technology strategies with the challenges they see on the horizon.

Additionally, I am pleased to report David Heard, who joined Infinera in June, has been named an officer of the company and will be responsible for product realization.

David has a strong track record around building businesses and driving operational efficiencies.

He has been tasked with enhancing internal alignment between our engineering, product management and marketing teams and ensuring efficiency in our overall development process.

While difficult, I strongly believe taking these actions now is the right thing to do for our company, our customers and our shareholders.

In the near term, despite challenges in the market, we believe this plan enables us to return to profitability and cash generation during FY18, which I consider an imperative.

Over time, we remain very optimistic about growth in the industry and Infinera's opportunity to outgrow the market.

Adjusting our operating structure at this juncture will allow us to refocus our R-and-D programs to ensure we lead in optical technologies, scale more efficiently for future growth and generate strong returns on the investments we have made in recent years.

Looking ahead, I believe our ramped-up cadence and optical leadership position us to capitalize on impending architectural inflections and grow faster than the market.

WE see 2018 as a year of important decisions, in which I believe we are well-positioned to address.

In metro, we continue to anticipate major cable and mobile operators will begin to award opportunities around fiber deep.

Additionally, we expect disaggregation and open will continue to increase in prevalence.

This is driven by, but not limited to, ICPs and massive bandwidth demand from cloud services.

Finally, in subsea, our belief is that providing leading spectral efficiency tied with the ability to instantly deploy bandwidth will become a prevailing expectations as customers award new routes.

With this backdrop of inflections and design decision in 2018, our refreshed product portfolio and faster technology cadence position us to win.

Customers make decisions based on optical performance, cost of ownership and ease of use.

Initial growth with our refreshed portfolio suggests ICE4 is successfully addressing customer requirements and gives me optimism that Infinera is on track to return to outgrowing the market.

Over time, my expectation is our faster technology cadence with ICE5 and 6 will enable us to lead in bringing high-capacity optical technologies to market.

In closing, we are taking decisive actions to ensure we return to profitability in 2018 and are more profitable as we grow in the future.

Our long-term opportunity is unchanged.

With the significant investments we have made in R-and-D to establish an end-to-end portfolio, I am optimistic we will be successful winning opportunities around impending architectural inflections.

Ultimately, by leading in optical technology and delivering solutions that enable our customers to win in their markets, we will be in a strong position to grow market share and return to achieving our stated business model.

With that, thank you to our customers, partners and shareholders for their ongoing commitment to Infinera and a sincere thanks to our employees.

Now I'll turn the call over to Brad.

Thanks, Tom, and good afternoon, everyone.

I will start with a review of the third quarter and then provide details around our restructuring plan, our outlook for Q4 and some high-level modeling for 2018.

As Tom mentioned, we executed well in Q3 delivering favorable results relative to our guidance ranges.

Revenue was $192.6 million, up 9% sequentially and 4% year-over-year.

Growth was primarily driven by ICE4 products as we benefitted from the growing traction with CX2 and began to recognize revenue for the XT and XTS-3300.

Q3 revenue for the 3300 was in line with our expectations in the low- to mid-single digit million range.

From a customer vertical perspective, we had 2 greater than 10% customers in the quarter, a wholesale and enterprise carrier, and a cable operator.

Rounding out our top 5 customers were an internet content provider, and international Tier 1 and a domestic Tier 1. Cable was again strong I the quarter benefiting from capacity additions and footprint expansions from large existing customers in North America and Europe including initial deployments of the XT-3300.

ICPs were slightly down sequentially as revenue growth from CX2 and XT-3300 deployments were offset by a fall-off in demand from one of our largest customers.

Results from wholesale and telco in the quarter continued to be heavily impacted by customer consolidation.

Of note, the combined spend from CenturyLink and Level 3 in Q3 was down nearly 10% sequentially and represented less than 20% of our revenue for the quarter.

Excluding these 2 customers, wholesale was up more than 10% sequentially and telco was essentially flat with spending largely subdued, aside from a couple of new builds in APAC.

Geographically in Q3, our international business drove most of our sequential growth led by EMEA accounting for 30% of total revenue and growing 22%.

This result in EMEA was driven by data center expansion from ICPs and improvement from wholesale stemming from subsea and initial XT deployments.

North America was steady in Q3 coming in at 59% of total revenue and growing 1% sequentially.

Offsetting strong growth from North American cable in Q3 were decreases in combined CenturyLink-Level 3 spending and weakness from one of our large ICP customers.

LatAm and APAC both grew in Q3 accounting for a combined 11% of overall revenue driven primarily by solid results in subsea.

Now on to margins.

Non-GAAP gross margin in Q3 came in slightly above the midpoint of our guidance at 39.1% as we incurred high early costs from our ICE4 units and were impacted by an increased volume of new customer footprint builds.

We also incurred costs to commercially transition customers to our new products though some of the costs we originally expected in Q3 were delayed to Q4.

While yields in overall cost performance of ICE4 are on track, we will continue to be margin challenged until our overall volumes recover more substantially.

Regarding OpEx, non-GAAP operating expenses were $90 million in Q3, well below our guidance range as we took steps to tighten spend in advance of our restructuring.

Even at a lower OpEx level, we remain committed to bring new products to market and execute on our faster technology cadence.

Enhancing our operational efficiency and refocusing our R-and-D initiatives should enable us to achieve this cadence even at lower levels of R-and-D going forward.

Putting everything together, in Q3 we had a non-GAAP operating loss of 7.8% and a bottom line net loss of $0.11 per share, solid improvements over recent quarters but still well below acceptable levels.

On a GAAP basis, we incurred a net loss of $37 million or $0.25 per share.

The difference between our GAAP and non-GAAP results was attributable to approximately $12 million in stock-based compensation, $5 million of acquisition-related costs and $3 million in amortization of debt discount.

Turning to the balance sheet, as of the end of Q3, cash and investments were $309 million, down $24 million from the end of Q2.

Our cash decrease primarily stemmed from burning $21 million in cash from operations in the quarter.

Of note on working capital, I was pleased overall inventory declined as ICE4 sales began to offset our ICE4 inventory ramp.

As for other cash drivers in Q3, CapEx came in at $11 million and we received $7 million in proceeds from equity issuances.

Now for some thoughts regarding our financial outlook, beginning with our restructuring plan.

We have started to see early success with our ICE4-based platforms and are confident in our ultimate opportunity with ICE4 and future products based on our ramped-up technology cadence.

That said, given the challenges Tom covered that could impact our growth in 2018, we believe now is the time to undertake a restructure initiative to sight-size our cost structure.

As we are very conscious of both our top line growth and our profitability levels, we believe this restructuring initiative will enable us to return to profitability and cash generation during FY18 and also be more profitable as we grow in the future.

Under this plan, we will lower expenses across all areas of the company by reducing our workforce by approximately 10% and also reducing our contractor spend.

As part of the plan, we are making several changes we believe will help our R-and-D efficiency.

Specifically, we plan to consolidate our development sites including closing our Beijing design center, more broadly leverage our engineering resources across regions and product line development, prioritize R-and-D initiatives including decision about what development will be performed in-house versus externally, and more effectively optimize our use of prototypes.

Outside of engineering, we have also made changes to allow us to operate more efficiently as we scale the business including reducing our facilities footprint and writing off certain equipment and licenses which will not be utilized in the future.

Relative to a pre-restructuring trajectory of around $400 million in FY18 OpEx, we expect this plan will save approximately $40 million annually as we plan to continually drive down expense levels as we finalize the ICE4 products such that we will exit FY18 on a much lower run rate.

From a GAAP standpoint, we expect to incur restructuring charges of approximately $25 million attributable to severance benefits, facility closures and other restructuring-related costs.

We expect to recognize most of the restructuring charges in Q4.

There could be additional charges taken in future quarters as we finalize our product rationalization efforts.

Now turning to our outlook for the fourth quarter of 2017.

We currently project revenue of $190 million plus or minus $5 million.

The midpoint of this range is slightly down on a sequential basis and up 5% on a year-over-year basis.

This expectation reflects our anticipation of a continued ramp in the CX2 and XT series offset by a further decline at the new CenturyLink due to the delay in the transaction closing and softness at certain ICP customers.

In Q4, we also anticipate the initial revenue contributions from our recently launched XTM2 16QAM metro edge platform.

Although we expect more substantial contributions over the course of FY18 as key existing customers adopt the platform.

As it is always difficult to predict whether and to what extent we could benefit from year-end budget flush, we have not baked much upside into our guidance.

Regarding margins, in Q4 we currently anticipate non-GAAP gross margin will be 38% plus or minus 200 basis points.

In Q4, we expect footprint deals will again grow and anticipate capacity adds will be lower than normal as several customers have indicated they're delaying capacity adds until next year.

While footprint deals are often margin-challenged upfront, they represent great opportunities for profitable growth as we move forward.

Additionally, we expect to continue to incur costs related to fulfilling commercial bridge deals and working through our early inventory of ICE4 units, which carry much higher costs than those products built recently.

Although we are not pleased with our current gross margin level, we are confident many of the issues impacting us are temporary and that we will improve our margin profile in 2018.

As we release the final ICE 4 products and shift our mix to the ICE4 platforms, we expect to more substantially benefit from the lower inherent cost structure of these products and demonstrate the leverage of our vertically integrated manufacturing model.

As for OpEx, we currently anticipate non-GAAP operating expenses in Q4 will be $91 million plus or minus $2 million.

This expectation accounts for our restructuring plan taking effect in the middle of the quarter.

Even as we reduce costs, we will continue to invest in bringing the remaining ICE4 products to market as well as lab trials to support the adoption of the new products that have already been released.

Below the line the combination of interest and other expense in Q4 is expected to net out to approximately $500,000 and tax expense should be approximately $1 million.

Putting this all together for Q4, the midpoint of our projected guidance translates to a non-GAAP operating loss of 10% and non-GAAP EPS is expected to be a loss of $0.14 per share plus or minus a couple of pennies.

As for GAAP EPS, we project it to be lower than non-GAAP EPS by about $0.29 per share primarily related to stock-based compensation expense and restructuring-related charges.

On the balance sheet, given our expectation of an operating loss, included in the restructuring-related costs we expect our cash balance will decline in Q4.

We intend to continue to tightly manage working capital and CapEx for the remainder of the year.

Finally, regarding our outlook, while we do not typically provide guidance beyond one quarter, given the uncertainty around market dynamics and pace at which our business will recover from customer consolidation, I wanted to provide some perspective on how we're anticipating our revenue in 2018.

With our addressable markets expected to grow only low- to mid-single digits, I feel good about outgrowing the market, but I am less confident in our ability to grow 10% or greater for the full year 2018.

That being said, with the actions we are taking to lower our cost structure, I believe even at a lower revenue level, Wall Street's current expectation that we get back to profitability during the second half of 2018 remains reasonable.

In conclusion, I will echo Tom's sentiment that the opportunity in front of us remains strong starting with ICE4 and continuing to strengthen as we bring our ICE5 and 6 based products out on our accelerated development cadence.

We enter this new phase as a company well-positioned to grow profitably armed with a product portfolio that can address the wide-ranging requirements of our customer base and a cost structure that as volumes ramp should be advantageous for us and for our shareholders.

We are confident that the optical market will grow nicely over time and that we will outperform the overall market and return to delivering best-in-class financial results.

With that, I'd like to turn the call over the operator to begin the Q-and-A portion of the call.

(Operator Instructions).

Our first question comes from George Notter with Jefferies.

A lot of information here.

I guess I wanted to start by just asking you how you feel about the new product portfolio and how it stacks up competitively vis a vis the other solutions that are in the marketplace.

I guess you're taking some actions on cost.

You're talking more about the ICE5 and 6 product pipeline.

And I guess I just want to make sure that you still feel good about where you're at competitively with ICE4.

Or conversely, is there some hesitation in terms of you guys are looking competitively with ICE4 and therefore we're kind of dialing back the costs and looking forward to the new stuff a year or 2 from now?

Yes.

I'll start, George, and then Dave can -- I'll ask him to add to anything.

So I think ICE4 is actually an exceptionally well-performing product portfolio and set of technologies.

If you look at how -- you can't see the numbers, but we look internally how ICE4 is ramping in comparison to ICE3 ramped, we're actually on track to be at the same level or faster in the first couple quarters of both bookings and shipments.

And ICE3, Gen 3, as you remember, was material to creating a series of successful years as we significantly outgrew the market.

CX2 is doing very well.

XT-3300 is doing very well both in current customers and in new customers.

And the XTS-3300, if you've seen our reports on spectral efficiency and the wins we're having, I do believe it is the best performing optical capability in subsea shipping today and it's created, quite frankly, a huge amount of interest that I think we'll be able to capitalize on in the next year.

The challenge with ICE4 is not the products that are shipping.

It's the products that we still have to make ship, the 3600 and the AOFX-1200.

Slipping them a quarter is -- it's an error on our part.

I take responsibility.

Those products address about 2/3 of our customer base.

So the ICE4 that's shipping and addresses about 1/3 of our market is doing, in my mind, exceptionally well, exceeding our expectations.

It's being extremely well-received in subsea.

It's opening up some new accounts in the market.

We announced Netflix today.

And I don't want to talk about what materiality or how big that could be.

The point to me is Netflix will have recently done a complete and exhaustive review of all products in that form factor that are available in the market and they picked us.

And they picked us not because we're anything less than the best solution, the best value for what they're trying to accomplish.

This restructuring has nothing to do with the receptivity of ICE4.

It has to do with, quite frankly, as we laid out.

We are still uncertain of when the top line from a CenturyLink-Level 3 will fix itself.

It has exceeded all of my expectations for how long that could go on and how challenging it has been.

I also think that when you look across the industry and look at the industry forecast, you look at the component suppliers in the industry, you look at across the competitive landscape, there's a fairly consistent thematic.

Carriers slow down on spending and ICPs have been erratic.

You see it in the component guys.

You see it in the module guys.

And you see it in the systems guys.

So I think that this restructuring, quite frankly, it's disappointing to have to do, but I think it's imperative considering the macro landscape in the environment.

Dave, on the technology?

Yes I'll add a couple comments to Tom.

We are seeing a great acceptance of our Gen 4 optical engine.

And the most stringent applications, which are the subsea, I think the subsea community has, one, both validated and demonstrated their strong interest for that platform.

That will continue to propagate through to the other applications as well.

And again, to reiterate Tom's comments, the Gen 3 product when it came out, we saw similar types of response functions and we still have a couple products left to go on the Gen 4 and getting that out there.

Brad indicated on the conference call that we are comfortable with the growing faster than market and that's driven by what we think is a superior technology.

Got it.

And then just as a follow-up, you mentioned earlier that ICPs are sort of opening up shelves and allowing alien wavelengths over those shelves in open line system configurations.

I guess that's always -- from what I understand, it's always been a requirement for the ICPs and I understand that a lot of those customers historically haven't taken advantage of that capability in the products they buy from vendors.

And it sounds like maybe that's changing now.

I guess I'm wondering why is that changing now and then how does that affect you in terms of margins and price with those kinds of customers?

So you're bundling it, George, that we said ICPs.

We didn't say ICPs are expecting open.

We said it as a more general statement.

And the ICPs have requested open now for quite a while.

And there has been a limited level of deployment on that.

The interesting part is we are now doing a significant number of open trials outside of the ICP space.

It's being, I would say, explored fairly seriously in multiple geographies, in multiple verticals.

And there is real intent, I think.

I think it's a longer sales cycle.

There is a caution that goes along with it.

But I think that it's an opportunity for some of these guys to lower their overall spend and put in place leading optical engines.

We have a significant initiative doing trials right now with a number of customers that extends into early to mid-next year.

We call it Open ICE.

So use your current infrastructure and put the latest ICE technology across it.

We have won a couple of deals.

They are not -- I would not -- well, they're very interesting and they're important.

I think that the momentum is building, that this is a very significant opportunity for us moving forward.

And I think it's going to be, quite frankly, as much outside of the ICP space as I think it will be in the ICP space.

There has been, even the guys driving it hard, a reluctance to operationalize some of this because of the complexities associated with running other people's transponders across different optical line systems.

But I think that the savings can be significant enough, the capability significant enough that this will be a transaction that starts occurring in the industry.

We're going to push it very hard.

The view that we have is the winner of that will be the guy who has the lowest-cost, best-performing transponder.

And if you look across the landscape, most of the world is not covered with our amplifiers so this is an opportunity for us to grow market and grow margin.

Yes, George.

To touch on your question on the margin piece.

Obviously, we make the majority of our margin on the transponder portion of the business.

So when that happens and we're able to put our transponders across other people's line system that will be a good thing from a margin perspective.

On the flip side, we obviously have to be cautious as we build new footprint with customers that they follow through and fulfill that with us because that is obviously a more costly portion of our infrastructure and less margin that's there.

Our next question comes from Doug Clark with Goldman Sachs.

This is Balaji Krishnan on for Doug.

I wanted to touch on your commentary on CenturyLink and Level 3. Perhaps if you could give us a sense for overall orders from CenturyLink and Level 3 for the industry.

Are they tracking well below where they should be?

Meaning are you losing share or do you think that your share is stable?

And then based on prior M-and-A at large customers, how long does it typically take from closure for the order rate to be at a normal run rate?

We are -- never say 100% sure.

But we are highly confident we are not losing market share within the CenturyLink or Level 3 footprint, either combined or separated.

I believe that they have just gone on a complete avoidance of capital expenditure.

I think they're waiting to get the, quite frankly, the contract signed.

They're working now to get a contract that reflects the new opportunity of the joint business.

So it's an opportunity for them to achieve part of their stated goal of CapEx reduction and price reduction.

I was wrong in assessing now low it could possibly go.

They have really cut spending.

How long it could take?

This one is unusual in the size of it so I have a hard time poking it saying so something like this can take X number of quarters.

In other ones that are closer in size, it took probably 2 quarters for it to get back to what I would consider a normal run rate.

So I'm anticipating that happening certainly next year.

I don't anticipate it happening this quarter.

And it could maybe not happen in Q1.

We just don't know.

But we feel confident we're not losing market share.

The one piece of information that Brad had talked at is the right now they are buying -- our customers in general are buying a larger fraction of footprint than they have in the past.

And we expect that that certainly will translate into line cards.

That's actually interesting in that in general we're selling a lot more footprint.

We're actually even selling footprint to those guys.

So we're anticipating having, when they start buying again, it will hopefully and I expect it to be in regard to fill.

So they have an outlook.

I have an outlook that the opportunity for next year is more on a more margin-rich fill.

And just to clarify, you mean you're selling footprint into CenturyLink and Level 3 currently?

Yes.

Okay, got it.

And then another follow-up on the one ICP that was weak in the quarter.

Any light you can throw on what's going on there?

Do you think it was temporary or anything more longer-term lasting from here?

As you guys, if you watched the Inphi announcement, the Color Z is starting to make some significant traction.

I believe that that is impeding some of our business with that customer, that one customer.

We're not seeing that as a pervasive trend, though.

It's one customer.

Got it.

And final question on the delayed products.

What's -- I know it's hard to get these product launches exactly right.

But what's causing the delay here?

Is it qualification issues or just something else going on there?

I think it's a couple of things.

One, we prioritized -- we have multiple engineering groups, but we have, I would say, some experts in certain areas and we prioritized this expertise on getting the CX2 and the XT and XTS-3300 out first.

That caused a little bit of a ripple in putting resources on these next platforms.

We've also had some challenges atypically with some technologies we've selected from external.

Nothing dramatic, but it has certainly impeded our ability to bring these to market as crisply as I had anticipated.

And third, I would say I was overly optimistic just in general on our ability to refresh our entire product line in that short a period of time.

It's a cascading type of problem.

When you slip a couple of things it just pushes the later things out.

I do anticipate -- we have a significant demand for certainly the AOFX upgrade.

And I anticipate it being in kind of pre-GA release into customers' hands in the earlier part of the quarter, Q1.

Our next question comes from Alex Henderson with Needham.

Thanks.

I wanted to talk a little bit about the OpEx guide that you just suggested for 2018 based on the restructuring.

I think your comment was, and correct me if I'm wrong, but you're at $400 million run rate and if you take $40 million out you should be at $360 million.

But when I looked at the numbers, taking the fourth quarter and multiplying by 4, gets me $365 million and that suggests only a $5 million improvement over the course of the year.

You're at $365 million as is.

At the midpoint of your guide for OpEx again suggesting only $5 million improvement.

How do I reconcile this $400 million number and particularly in light of the fact that you had suggested that OpEx would be fairly flat in 2018 in recent conversations?

Yes.

So, Alex, the guidance for the fourth quarter already includes the impact of the restructuring.

So the numbers that I quoted of the $400 million is the pre-restructuring numbers, right, what it would have been had we not taken these actions.

And then as you get into '18 you've got the reset of the bonus, you've got the impacts of focal, you've got the reset of commissions.

So if we had done nothing else, based on the trajectory we were on, that's the $400 million number.

So that is what we started with and the $40 million is off of that number.

Just to be clear, so you're suggesting that in the fourth quarter you would have been closer to $100 million in quarterly OpEx costs?

Because that's certainly a pretty big increase from $90.3 million.

Yes.

So if you look back at our Q1 and Q2 spend, Alex, and you look at the trajectory we were on, Q4 would have been around $100 million of spend.

So you're seeing in Q3 and definitely in the guide in Q4 already the impacts of tightening and holding back spend and starting to get in advance of the restructuring.

Okay.

Second question.

In your guidance on the top line it's implying kind of mid-single digits, maybe a hair better than that for the year.

One would think that by the back half of the year with Level 3 and CenturyLink completed that you would be able to get at least a 10% improvement in revenues just form then rebounding part of the way back to where they were.

If I recall correctly, there was a lot of comments made by various executives over the last quarter or so suggesting that CenturyLink was running, quote-unquote, "red hot" and really was turning away business because of lack of CapEx.

So how do I reconcile those three comments?

Yes.

So Alex, we do believe that when CenturyLink and Level 3 start to spend again, that will be good upside for us.

They have been running quite a bit lower than we had ever expected.

Obviously, in Q3 and again in Q4 they continued to decline.

So when they start to recover, that will be a good thing.

The challenge we have is; when will that be, how much will that be when it comes back?

You are correct to assume, though, by the second half of the year that their spend should be growing at a nice clip.

The overriding aspects of the market is the other thing that gives us caution for '18.

But just to be clear, wouldn't CenturyLink alone give you over a 10% growth rate in the back half if the company kicks back into gear and starts spending anywhere near where they had been prior to this merger?

I mean it's pretty hard -- Are you suggesting other things are going to be down to offset that if that occurs?

So Alex, you're assuming that they're going to spend the levels they did prior to the mergers independently.

We don't know that.

Yes, if that's possible, you're right.

That would drive more than 10%.

But it would be, I think, imprudent of us to assume that.

I think a couple things are going to have to happen.

One, they're going to have to have a network strategy.

I don't know what that is.

They have 2 long-haul networks.

How many are they going end up with?

Less than 2. So there's some overlap in their networks that we have to go understand what are they going to do from a business perspective.

Two, we're going to go through a contract reconciliation.

A dollar before might not be a dollar afterward of business.

So we have a lot of those things that are completely uncertain right now because it's been closed for about a week.

I understand your premise.

I think the premise is based upon some logical thinking.

But it's assuming that we know what's going to happen within that new customer.

We don't.

One last question if I could.

You commented about pricing conditions and the language you used was pretty much the same as the language you'd used in prior quarters.

Has there been a change in the trajectory of pricing over the course of the last couple of months?

Because, as far as I can tell, it's pretty much as it has been.

Yes.

It's not that it's gotten worse, Alex.

It just continues to be challenging.

So obviously we said it wasn't as bad as last year, but we continue to see situations where competitors bid very aggressively on deals.

It's not a new phenomenon.

It just continues to be painful.

It's a hyper-competitive environment.

One last question if I could.

You had talked about getting back to a gross margins in the back half of '18 in the past I think suggesting at least in the mid-40%s.

Can you give us some sense of what your assumptions are relative to the gross margins over the course of '18 as the mix shifts to your new products?

And I would assume you get some improvement in card sales.

Yes.

So the biggest thing is the transition cost for some of the commercial deals that we've done and the high cost of the early ICE4 units will largely be done by the end of this year.

And then as you start to grow with more and more of the Gen 4 units, both the XT and CX, and you get the rest of the portfolio out, you'll get to experience that much better cost structure.

That also has benefits in terms of what is in going through the fab, which will drive volumes there as well.

So mid-40%s is doable in the back half?

Yes.

(Operator Instructions).

Our next question is from Michael Genovese with MKM Partners.

You've already answered the question about the one large ICP where there seems to be a shift to another technology.

But just in the ICP market in general, it seemed like you also said there was some delays in spending through the end of this year and into next year.

So can you talk about the ICP market more generally what you're seeing in terms of spending patterns and seasonality over the next couple of quarters?

Sure.

So, Mike, the comment you're referring to is one and the same.

It's the same customer.

The rest of the ICP space has actually been fairly good.

We have seen then start to adopt the, both the CX and XT platforms.

We continue to win more and more customers in the ICP space.

And we have multiple more that are trialing the new products.

So the comment you're referring to is one and the same.

Otherwise, spend still seems relatively healthy.

Okay.

I don't want this to be my follow-up event though it is a follow-up.

But just in terms of seasonality around the Christmas season, a slowdown in that market.

Is that a phenomenon that we should expect or not in that market in general?

No.

We don't see anything specific in that space for seasonality.

The only piece I'll point out is, as I mentioned in my prepared remarks, we have seen a couple of big customers that have basically pushed all of their capacity adds out into next year, which is a challenge for us.

And then the year-end budge flush side of things is really hard to call.

So we haven't really baked any of that into our assumptions either.

Okay.

And then the very first question you were asked about your technology versus other technology.

I just wanted to ask it in a more focused way.

Because ICE4 is a 1.2 terabit DSP.

I think Sienna has a 2.4 terabit.

I think Acacia doesn't have their 1.2 out yet.

But besides looking at that headline number, what other facets and technology capabilities should we be looking at?

I mean are we just comparing the 1.2 to the 2.4 or is there other features that your product has that makes it more attractive, say, in a DCI application versus somebody else who has a 2.4?

Yes.

So what people have when they refer to a 2.4, they talk about 2.4 terabits in a box.

What we have is 1.2 terabits in an optical subassembly.

It's coming out of a common PIC.

It's a unique piece of technology.

The comparative piece of technology that deployed in the market today is a 200 gigabit equivalent.

Our 3600 is a 2.4 terabit box.

We believe that there's no need to, at this point in time, to drive the capacity per box higher.

The cost and economics leverage is not gigabits per box.

It's gigabits per optical subassembly.

That does that completely different set of technologies that drive that.

Photonic integration is what enables a multi-wavelength, higher-capacity per optical package capability.

So the optical network is key for that, but it's not the only piece of technology critical for the application.

We see a significant adoption by our customers for the enablement that that gives you for instant bandwidth and the ability to have bandwidth-on-demand networks.

That's a very strong offering within the service provider space and in the subsea application space for that.

We also see the software interfaces, the CLI software interface, the ease of turn-on that we've facilitated for our box in a ICP application.

Again, Tom referenced this, Netflix, a win that we've got.

That's as much about software, ease of use and reliability and rapid turn-up as it is about what the optical engine is capable of doing.

So it's a complete offering and I'd say from our customer growth, we continue to win market share because of that suite of offerings.

In regard to DSP, I believe we're the only people today offering SD-FEC gainsharing, which matters in the subsea space.

I think we're the only people offering today commercially Nyquist subcarriers, which matters in performance in subsea.

And I believe something that's terrestrial that people don't talk about a lot is the ability for lightning strike on our receptivity.

People who run terrestrial networks, they often don't realize that a lot of it is aerial and they're surprised when they find out they have aerial networks and they're surprised because lightning strikes cause disruption to their transmission.

It's a fairly specific capability you have to design into the DSP.

I believe ours is leading in the market.

And I think it will matter to people who are making terrestrial decisions.

So when you talk about DSP, it's important not to say DSP for ICP.

DSP for various segments have different requirements.

And I think we have really, quite frankly, the leading DSP out there today.

Yes, absolutely.

Our next question comes from Stanley Kovler with Citi.

I wanted to ask you guys just a, first of all, a high-level question.

One thing that we've talked about in the past is that a lot of the traffic across the networks is staying on the cloud network and so the utilization rates on the telco networks may not be increasing as much as we've seen in the past.

I guess, do you have a sense of what your customers are doing in terms of that capacity utilization at this point?

And can you confirm that when the orders come back, Tom, I think you referenced this a couple times, should we expect that the order patterns will be very incremental so in other words the recovery will be flatter than we've seen in the past?

And I have a follow-up.

I don't understand your question in regard to cell traffic staying in the cloud.

So I apologize.

In regard to orders coming back, I think a couple things are going to drive opportunity.

I think we're going to continue to see opportunity in subsea where we will, I do believe, we will gain market share with our new optical family.

I believe that the DCI market is exceptionally competitive, but I think we're well-positioned to win some new customers with our CX2.

The XT-3300 I think will continue to grow somewhat linearly.

I think a big opportunity is for customers to upgrade their DTN-

X network with the AOFX.

And I think that that could drive some significant revenue probably in the second half of next year as people certify it in the first half and make expansion plans.

And the 3600 I also think could be a significant step function in the second half of next year.

It's really hard to answer do I see it coming back linearly or more step function.

A lot of it is going to be dependent upon the continued growth of bandwidth, which I think everybody believes will continue to grow unabated, versus ASP compression.

And right now one of our competitors said in their market they see ASP compression happening at the same rate as bandwidth growth.

I quite frankly don't think it's that bad.

But it's a pretty compelling competitive environment right now.

I think that over time the industry grows significantly more than the 0% to 2% long-haul is forecasted for next year.

And I think that our new technology should allow us to continue to grow market share.

Brad being conservative and I think reasonably conservative on saying he has a hard time telling shareholders to expect us to grow 10% next year.

We have to reflect what we see today.

I do think that there is significant opportunity that if things transpire in a positive way he could end up being conservative.

But we have to tell you what we believe is a reasonable look right now.

Thanks.

My question about the traffic staying on the cloud wasn't related to mobile.

I just meant in general, traffic overall.

But my follow-up question is really just near-term outlook for Q4.

Can you guys parse through what impact relative to maybe what we've had as growth expectations or you've had as growth expectations between the XT-3600 push-out, maybe some DTN-X and other issues.

Can you bridge us to where we were as far as expectations were based on various items and what they accounted for in terms of the lower revenue?

Sure.

The biggest thing, Stan, is the CenturyLink-Level 3. I mentioned that that is down sequentially in the fourth quarter by another 10% or so.

That's a big problem for us.

Also, the big ICP customer dropping their spend down pretty significantly.

Those are the 2 items between what we were expecting and the world was expecting versus what we're guiding to.

Our next question comes from comes from Meta Marshall with Morgan Stanley.

A quick question on the Netflix announcement that was made today.

And Arista had made an announcement kind of a couple of years ago about their relationship with data center interconnect and Netflix.

And so just trying to gauge kind of what the difference is between the 2 announcements where there.

And then second, as you guys talk about moving towards trialing or selling more open line systems and kind of building upon George's question from earlier.

But just how does that change the pricing model of a traditional kind of razor blade pricing model if you're not necessarily selling the chassis or you're not necessarily selling the line cards?

Like do you think that that pricing model has been figured out and kind of accepted by the industry?

The first question regarding Netflix, I candidly don't know what Arista announced with them.

It doesn't surprise me that Netflix would be a customer of Arista and a customer for us.

We provide an optical transport interconnecting data centers or data centers into the cloud.

And that's what they're using us for.

That would be certainly different than what Arista would be using for, which is probably inside the data center for switching.

So I can't answer your question on how it's different.

We sell a different application and a different architecture.

So I don't think it's in any way competitive with Arista.

It's not unusual for us to be in customers with Arista.

Does that answer your question at all?

Yes.

I mean they had made indications that it was more for their optical systems and in the past.

So I don't know if it was just shorter reach or -- but they had made indications that it was more for optical applications rather than switching.

But we can take it offline or follow up with kind of Arista as well.

We should probably take it offline.

And I'll call Jayshree and ask her.

Meta, on your Open ICE question, yes, it is a bit of a different model.

And, as Tom pointed out, lots talk about it, lots of desires to have it.

Not seeing it deployed widely yet.

I think it will be at some point.

But obviously you have to be more cautious as you put in line systems because historically you could afford to price line systems very aggressively because you knew you were going to get the transponders.

Now, we need to be more careful in those situations.

We need to make sure that if they're putting other people's traffic on those line systems that we're getting licenses for those, that we still make our money.

Or we go to customers and say, "Well, you got to pay us a more fair return for the line system and in exchange for that you get more flexibility." So it does take more management on a deal-by-deal basis.

But longer term, it's something we'll deal with and I think will ultimately be a good thing for us.

Yes.

I think that the whole industry right now is trying to figure out the appropriate business model for this new paradigm of open.

The razor and blade model has worked in the industry for a very long time, but it is no longer going to be the model I think that -- it will be there for a while, but it's going to transition into a more open line system model.

You watch what Google and Facebook are doing with their initiatives.

They're trying to drive these open line system models.

If you look at us or our competitors, as Brad said, when you sold your amps, you were anticipating filling it with your transponder.

There's going to have to be an appropriate level of business structure around if you use my amp and you don't fill it, how do I get paid?

How do I get a reflected value from that?

Most people, most customers accept that.

But we're in that transition of from-to and I think that there's going to be a lot of dancing done.

And that's going to slow some of this transition down.

But I think the transition, quite frankly, is inevitable.

I'm going to mitigate the commentary a little bit.

There are in general two types of networks.

There are point-to-point networks and then there are optically-meshed or layer 1-meshed networks.

The difficulty in deploying an open line system for an OTN network is very difficult.

Right?

The operational aspects of trying to deploy a point-to-point optical link over an optical line system, it's a little bit easier to implement the task as you don't have to keep track of the service across multi-domains within a network.

When you're trying to keep track of a service over multi-domains, you need the one neck to grab across that infrastructure.

So when Tom and Brad are talking about the Open ICE and those opportunities, it's really for point-to-point pieces within a network.

ICPs between your data centers, they have a certain amount of services within a service provider have it.

The bulk of the service provider business, however, is not a point-to-point application.

It is a meshed application.

And so you've got to look at the 2 markets differently there.

Our next question comes from Simon Leopold with Raymond James.

Just wanted to get 2 quick clarifications before a question.

One the clarification front, it does look like the gross margin for services took a step down.

So I certainly understand the explanations on gross margin pressure around the products.

But just wanted to get an understanding whether there was something in the services segment that we should appreciate.

It's generally been above 60%, I think, most of the time.

It looked like a little over 58%.

So first clarification is regarding gross margin on services, what's going on there.

Sure.

So Simon, as you know, the mix between our deployment services and our ongoing maintenances -- maintenance services are different so any quarter where we have stronger footprint adds that's going to come with some of that EF-and-I, which has a lower margin profile.

But we are still in that 60% range.

So it will fluctuate a bit, a couple points each way, depending on the mix of the services that we're doing.

Great.

That's very helpful.

And then I wanted to follow up on an earlier question around the OpEx guidance.

So to kind of put a fine point on it, do you expect operating expenses for the full year '18 that you're budgeting for roughly $360 million?

I just want to make sure I'm understanding this restructuring forecast.

Yes, you're in the right ballpark, Simon.

Great.

That's very helpful.

And then really the question I wanted to ask is we've talked a lot about kind of the web-scale trends and the overall market.

I was hoping maybe we could double-click on where you're going in terms of metro opportunities because it does seem as if that's the market where you've had little presence.

You've got the product cycle angled to it.

So how should we think about maybe a portion of your business or whatever quantification you can offer around metro?

Sure.

I'll try and answer the metro question.

Two places that we're having a significant successes in the metro today are a high-capacity metro distance links.

As far as 100 gigs in the metro, I think we're certainly within the leadership position there because of that.

We are also having a lot of success in what I would call metro access types of applications.

And we see a lot of opportunity.

Much of the cross-selling opportunity that we've talked about for the last couple years we've had good named account wins, but we haven't seen the dollar volume coming from that.

We expect that that's still in process.

The missing piece of our metro strategy is the connectivity bandwidth management element that connects those high-capacity links with the lower-capacity access ports.

And that's a product that we are definitely focusing some energy on.

We see the metro market as a good growth opportunity for Infinera going forward.

Yes, Simon.

Let me just touch on one other thing.

I think what we're alluding to longer term is the opportunity with some of the fiber deep initiatives.

Whether it's the cable guys pushing more and more fiber deeper out into the metros, we see those as very big opportunities to leverage both the strengths that the Sweden team has brought to the portfolio, but also linking those with our XT portfolio.

And if you look at the amount of spend the guys are talking about there as well as what 5G drives onto the network, those will be very big opportunities longer term.

They're not over the next few quarters, but I think those are great metro opportunities for us as we go forward.

So fair to -- just trying to quantify this -- but to think of your metro products through acquisition and organic combined, on the order of 15% to 20-plus% of sales in '18?

Is that reasonable?

It's in the ballpark.

It's in the ballpark, yes.

And then I think longer term -- yes, so, Simon, I think longer term as we get traction with some of those other initiatives, it could be more than that.

Our next question comes from Jeff Kvaal with Nomura Instinet.

A couple maybe left to go.

One is I think you mentioned at one point earlier in the remarks that spending is erratic.

And I'm wondering if you could spend a little bit more time fleshing that out a little.

And then secondly, what type of -- how much revenue, how much PIC4 revenue or ICE4 revenue needs to pass before the margins start picking up for that product line?

So we'll break it in 2 parts.

In the erratic part, it's really guided -- we've talked about it from the ICP perspective.

And what we see is that they are not consistent buyers quarter on quarter.

They'll make some significant purchases for certain metro expansions or long-haul expansions.

And then it will kind of go away for a period of time.

So it's not a steady consistent level of run rate that we see in typically other segments.

I think we see that and I think other people in the industry have commented on that recently.

And I don't know if that's new way of doing business or if this is a point in time where they are doing transitions of technologies.

But we do see it being -- still lots of opportunity.

We still continue to participate in a lot of ICP and we'll continue to pursue new ICPs.

But it is a bit of a lumpy or erratic level of spend.

Yes.

So Jeff, on your ICE4 question.

The ICE4 volumes are increasingly very nicely.

So we mentioned in Q3 that we were low- to mid-single digit millions in revenues.

In Q3 that grew several times over.

Obviously, the law of small numbers to some extent.

But in Q4, that continues to grow significantly again.

So the Gen 4 piece themselves, the only thing that's dragging on those margins today is just the high cost of the early units, which should largely subside by the end of this quarter.

The thing that hurts us more is just the overall volumes going through the factory.

So Gen 3 being down because of big customers like Level 3 and CenturyLink not buying in the volumes they would otherwise.

So once we get the higher cost of the early Gen 4 units out, we get through some of the commercial deals and you start to see some bounce back, you start to get to see some very nice looking margins again.

Okay, Brad.

So I guess the point is it's not as much about the ICE4 volumes per se as it is about overall volumes.

Your factory absorption is really the limiting factor on your gross margins.

Correct.

ICE4 has -- anytime you introduce a new generation, there is a period where you have to consume the early units.

As Brad said, that is mostly going to be done through Q4, but the overall loading of the PIC fab, whether it's Gen 2, 3 or 4, is causing more margin pressure across the company.

Thank you for joining us this afternoon and for your questions.

We look forward to updating you on our continued progress.

The conference is now concluded.

Thank you for attending today's presentation.

You may now disconnect.