邁威爾科技 (MRVL) 2020 Q2 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Q2 2020 Marvell Technology Group Earnings Conference Call.

(Operator Instructions) As a reminder, today's conference call is being recorded for replay purposes.

It is now my pleasure to turn the conference over to Mr. Ashish Saran, Head of Investor Relations at Marvell.

You may begin.

Thank you, and good afternoon, everyone.

Welcome to Marvell's Second Quarter Fiscal Year 2020 Earnings Call.

Joining me today are Matt Murphy, Marvell's President and CEO; and Jean Hu, our CFO.

I would like to remind everyone that certain comments today may include forward-looking statements, which are subject to significant risks and uncertainties and which could cause our actual results to differ materially from management's current expectations.

Please review the cautionary statements and risk factors contained in our earnings press release, which we filed with the SEC today and posted on our website as well as our most recent 10-K and 10-Q filings.

We do not intend to update our forward-looking statements.

During our call today, we will refer to certain non-GAAP financial measures.

A reconciliation between our GAAP and non-GAAP financial measures is available on our website in the Investor Relations section.

With that, I'll turn the call over to Matt for his comments on our performance.

Thanks, Ashish, and good afternoon, everyone.

During the second quarter of fiscal 2020, Marvell delivered solid results with revenue above the midpoint of guidance despite the challenging macroeconomic environment creating weakness across several end markets and the impact from the current export restrictions on Huawei.

We expect both of these factors to continue to impact us in the third quarter, but as we'll discuss, we remain well positioned to capitalize on infrastructure opportunities spanning 5G, data center, enterprise and automotive applications as we look forward to our next fiscal year.

Specifically, revenue for the quarter was $657 million, with a GAAP loss per share of $0.09.

Non-GAAP earnings per share was $0.16, slightly above the midpoint of our guidance driven by the higher revenue level and lower operating expenses.

GAAP operating expenses were $397 million, $22 million above the midpoint of guidance.

Non-GAAP operating expenses were $280 million, $8 million below the midpoint of guidance.

We are continuing to prudently manage expenses as we navigate the industry downturn.

Keep in mind, this expense level reflects the complete realization of synergies we sought to capture from the Cavium acquisition, plus the incremental cost savings we outlined in December of last year.

I'm pleased to report that we delivered these expense reductions 2 quarters ahead of schedule.

This achievement is a testament to the operational platform we have established within Marvell and the disciplined approach we have implemented as part of our transformation, enabling us to aggressively invest in high-growth areas even within a smaller expense envelope.

Our integration track record will serve us well for the upcoming acquisitions of Aquantia and Avera as well as the divestiture of our Wi-Fi business to NXP.

As we look forward, we are preparing for significant new product ramps within our infrastructure markets, particularly 5G, providing strong growth catalysts and secular positive offsets to the weakness across the broader semiconductor industry.

Before I discuss our core businesses, I would like to provide an update on Marvell's executive team.

Recall that back in May, Tom Lagatta, our current Head of Global Sales, announced his decision to retire later this year.

I'm very pleased to announce that we are promoting Dean Jarnac to Senior Vice President and Head of Global Sales for Marvell.

Dean joined us in 2017 as our VP of North America Sales and Global Distribution and previously held positions of increasing sales responsibilities at Samsung, Broadcom, Freescale, Altera and AMD.

Dean is an ideal candidate to assume the strategic leadership role given his demonstrated success in building strong customer partnerships and managing global organizations.

He will be an invaluable addition to our executive team.

I would also like to recognize Tom for the enormous contributions he made in the transformation of Marvell.

Tom joined us at a critical juncture and built a strong global sales team, including the recruitment and mentoring of Dean.

We wish Tom all the best as he returns to retirement.

Now moving on to the performance of our 2 core businesses.

First, in our networking business, revenue during the quarter was $330 million, down 3% sequentially, with seasonal growth in Wi-Fi products more than offset by the U.S. government's export restrictions on Huawei and the pause in demand from our base station OEMs as they transition from 4G to 5G.

More importantly, I'm very pleased to report that strong execution by our engineering and operations teams, coupled with a very close relationship with our lead customer, is enabling us to launch our first 5G products, which include our fusion baseband and OCTEON-embedded processors and our Ethernet switches and PHYs into production in the third quarter ahead of plan.

This early production start gives us and our customer further confidence in ramping up in the fourth quarter.

In fact, our 5G silicon is enabling our lead customer to deploy trial base stations this quarter in Tokyo to light up their Olympic Village in preparation for next summer's Games.

Additionally, we are on track for our customer to start deploying fusion processors for massive MIMO offload in remote radio heads in the fourth fiscal quarter.

Even more exciting, we have secured another strategic design win at our lead customer with our follow-on baseband solution in the next technology process node to provide additional processing capacity within a smaller power envelope for their next-generation 5G base stations.

This win builds upon our very successful multiyear partnership, where we provide advanced baseband processors, enabling our customer to drive higher performance, lower power and faster time-to-market.

In addition, fusion baseband development for our second Tier 1 base station OEM is also proceeding on schedule, and we remain on track to sample early next fiscal year and ramp production at the end of next fiscal year.

Based on the design wins we have secured so far in using industry analyst forecasts for base station units while holding our customers' market share at their current positions, we estimate that in a few years, our 5G revenue potential can exceed $600 million per year.

Of course, our revenue potential can flex above this if our lead customer is able to gain share as they drive towards their long-term goals and from additional design wins, which we are currently pursuing within our comprehensive 5G platform to address baseband, transport, switching, fronthaul and massive MIMO opportunities at multiple base station OEMs.

Further, next calendar year is expected to be the inflection point for 5G adoption, with industry analysts such as Dell'Oro projecting 5G macro base station penetration to grow from about 10% this year, to rapidly increase to 38% next year and then on to 55% in calendar year 2021.

Moving on from base stations to our enterprise and data center markets, our revenue grew sequentially driven by stronger-than-expected demand from our Chinese customers who have not been impacted by export restrictions.

However, these markets have remained generally soft.

Therefore, we believe that some of our relative strength could be due to these customers building inventory to guard against any future supply chain disruptions.

Nevertheless, on the new product front, our refreshed Ethernet products continue to win new designs, which will drive multiyear growth.

Our PHY team had a very strong quarter, with multiple design wins in 3 separate platforms at a Tier 1 U.S. networking OEM.

These include gigabit and 10-gigabit copper PHYs for a very high-volume enterprise access switch, 25-gigabit optical PHYs for a high-capacity enterprise access switch and 56-gigabit PAM4 optical PHYs for an enterprise aggregation switch.

Equally exciting, I'm happy to announce the first strategic design wins for our 12.8-terabit switch platform which we had introduced earlier this year.

The switch will be powering a next-generation firewall appliance from a Tier 1 networking OEM, and we have also secured an enterprise aggregation switch at a large Asian networking OEM.

These PHY and switch wins will go into production late next fiscal year, with the bulk of the ramp starting in fiscal year 2022.

In our automotive product line, as you may recall, earlier this year, we'd announced design wins with 16 automotive OEMs spanning Europe, North America and Asia.

We have now started the initial ramp of our gigabit Ethernet secure switches and PHYs at some of these OEMs for their upcoming model year 2020 rollouts and expect to grow more substantially next fiscal year when we ramp the remaining majority of these design wins in support of model year 2021 launches.

These multiple design wins were secured across a variety of applications, including infotainment, ADAS, telematics, central gateway and body domain controllers.

The investments we made to establish in-house automotive-grade capabilities are now starting to pay off, and we recently achieved a key qualification, ASPICE level 2, an important software process development certification tailored specifically to the auto industry for developing high-quality embedded systems.

This represents a key differentiator versus our primary automotive Ethernet competitors.

We believe that our technology investments, which will be further enhanced by the multi-gig capabilities from the upcoming acquisition of Aquantia to have positioned us to become the leader in the automotive Ethernet connectivity market.

This market is projected to grow rapidly from a low base today to well over $0.5 billion over the next several years.

The steep trajectory is not driven by automotive unit growth but rather by the growing proliferation of high-speed in-vehicle networks, connecting the increasing number of sensors in cameras for driver assist and higher levels of autonomy as well as richer infotainment and more advanced telematic offerings.

The coming increase in bandwidth and the sheer number of endpoints, which need to be connected and shared, will require secure standards-based Ethernet fabric designed for speed and scalability.

In a few years, we believe that the overall automotive market can become another substantial growth engine for Marvell.

In addition to Ethernet connectivity, we see opportunity to leverage additional technologies such as our processing, security and custom design capabilities in the automotive market.

Moving on to our outlook for the third quarter.

We expect a low single-digit sequential decline in networking revenue.

This outlook reflects softness in demand from the enterprise networking end market due to current macroeconomic conditions and in particular, a recent significant forecast production from a key enterprise networking customer as well as the seasonal decline in Wi-Fi, partially offset by growth from our base station products driven by the start of our 5G production shipments.

Turning to our storage business.

Storage revenue for the second quarter came in above our expectations at $275 million, declining 1% sequentially, better than our guidance for mid-single-digit sequential decline.

As expected, our storage business was impacted by the export restriction on Huawei, but we benefited from stronger-than-expected demand from a broad set of storage controller customers in the HDD, SSD and fiber channel end markets.

It appears that previously elevated inventory levels have slowly started to subside at some of our storage controller customers.

We also believe that PC builds picked up in the second quarter with better CPU availability and that demand for high-capacity nearline drives also stabilized.

Growing our revenue from the enterprise and data center market is a key strategic objective for Marvell, and to that end, we are now shipping in multiple nearline platforms, with capacities up to 16 terabytes.

In addition, I'm also pleased to announce that we now have a design win in the next-generation platform, targeting even higher capacity points well into the 20-plus terabyte range.

Our storage controller team had a very strong showing at the recently concluded Flash Memory Summit, or FMS, where we introduced 2 breakthrough products, an NVMe over Fabric Ethernet SSD controller and a family of PCIe Gen 4 NVMe SSD controllers.

We demonstrated both of these products as well as additional technology solutions, including artificial intelligence, SSDs, Fiber Channel over NVMe and centralized automotive storage at this premier industry event.

Our Ethernet SSD controller enables an SSD to directly connect to an Ethernet network without the need to go through a host, such as a server.

Disruptive new data center architecture significantly reduces cost by eliminating power-hungry CPUs, smartNICs, DRAM, and PCIe switches, while also reducing latency, increasing throughput and lowering downtime due to SSD failures.

These controllers fully integrate with Marvell's data center Ethernet switches, and their introduction marks a key milestone in advancing our end-to-end Ethernet storage strategy.

Also, at FMS, Toshiba Memory showcased the world's first direct-to-Ethernet SSD, their dual-ported 25-gigabit Ethernet solution, leveraging Marvell's Ethernet SSD controller technology.

Our new NVMe SSD controllers represent the industry's first PCIe Gen 4 SSD controllers to be fabricated on a 12-nanometer process, which consumes lower power while delivering better performance.

Multiple ecosystem partners, including AMD, Lenovo, Micron and Toshiba Memory, have expressed very strong support for these new products.

More importantly, these controllers provide the core architecture for our upcoming embedded and do-it-yourself custom SoC flash controllers, which we expect will start ramping late next fiscal year.

Moving on to our outlook for our storage business in the third quarter of fiscal 2020.

We expect an increase in demand for our storage controllers from the data center and enterprise markets, especially from high-capacity nearline drives, and some additional catching up in the SSD market from the undershipment in prior quarters.

Demand for fiber channel adapters should also trend up in the third quarter.

In contrast, demand for HDDs from PCs and gaming is expected to remain soft with sub-seasonal growth for this part of the fiscal year.

As a result, we expect an approximate high single-digit sequential growth in our third quarter storage revenue.

In closing, while we remain very -- while we remain in a very challenging macroeconomic environment, which has certainly worsened recently and has impacted our guidance for the third quarter, we continue to focus on things we can control.

We're winning new designs, optimizing operating expenses, introducing new products on or ahead of schedule and expanding our product pipeline.

It is particularly exciting to see the production of our first 5G products accelerated into the third quarter, faster than prior expectations.

Of course, we expect our 5G business to ramp more substantially in the fourth quarter and well into our next fiscal year from new 5G base station deployments in multiple geographies.

In addition, we also expect to benefit from our customers starting to replace FPGAs with our purpose-built 5G solutions in the [pre-5G] base stations we have already shipped.

At the same time, our storage business is starting to recover, and we are increasingly pivoting this business towards enterprise and data center applications.

Further, with the Aquantia and Avera acquisitions, Marvell is well positioned to capitalize on a broader set of opportunities, leveraging our unique standard, semi-custom and full ASIC design capabilities towards realizing our vision of becoming one of the world's leading suppliers of infrastructure semiconductor solutions.

With that, I'll turn the call over to Jean for more detail on our results and outlook.

Thanks, Matt, and good afternoon, everyone.

I'll start with a review of our financial results for the second quarter and then provide our current outlook for the third quarter of fiscal 2020.

Revenue in the second quarter was $657 million versus our guidance of $650 million at midpoint.

Networking represented 50% of our revenue in the second quarter, with storage contributing 42%.

Other products revenue was $52 million and accounted for 8% of total company revenue.

After a strong second quarter, we expect other products revenue to decline sequentially in the third quarter.

GAAP gross margin was 53.4%.

Non-GAAP gross profit was $415 million or 63.3% of revenue, reflecting the change in mix within the quarter.

GAAP operating expenses were $397 million, $22 million above midpoint of guidance driven by acquisition and divestiture-related expenses.

Non-GAAP operating expenses were $280 million, below the guidance range provided in May.

As Matt discussed earlier, we were very successful in achieving our planned expense reductions 2 quarters ahead of schedule.

Non-GAAP operating profit was $136 million or 21% of revenue.

GAAP loss per diluted share was $0.09.

Non-GAAP earnings per diluted share was $0.15, slightly above midpoint of our guidance driven by higher revenue and the lower operating expenses.

Now turning to our balance sheet.

In the second quarter, we returned $56 million to shareholders through $16 million in share repurchases and $40 million in dividends.

We exited the quarter with $573 million in cash and cash equivalents and a long-term debt of $1.7 billion.

We have paused share repurchases and debt reduction while we work toward closing the acquisitions of Aquantia, Avera as well as the sale of Wi-Fi business to NXP, and that we currently anticipate all of these transactions to be completed within our previously communicated time frame.

Given the ongoing uncertainty in the market, we actively reduced the China inventory in the second quarter and plan on doing the same again in the third quarter, with the aim of reducing channel DOI 2 weeks below our target level.

Let me now move on to our current outlook for the third quarter of fiscal 2020.

Please note that this guidance does not include any contributions from the pending acquisitions of Aquantia and Avera as well as the divestiture of Wi-Fi business.

Similar to last quarter's outlook, we expect our revenue in the current period to remain impacted by the U.S. government's export restriction on Huawei.

Specifically, we are forecasting revenue to be in the range of $660 million, plus or minus 3%.

We anticipate our GAAP gross margin will be approximately between 53.5% and 54.5%.

Our non-GAAP gross margin will be between 63% and 64%.

We project our GAAP operating expense to be in the range of $385 million, plus or minus $5 million.

We anticipate our non-GAAP operating expenses to be in the range of $280 million, plus or minus $2.5 million, which at the midpoint will be similar to the prior quarter.

We continue to maintain tight control for discretionary expenses given the current uncertain macroeconomic environment.

We expect interest expense to be $20 million.

As a result, we anticipate GAAP loss per diluted share in the range of $0.09 to $0.05 and non-GAAP earnings per diluted share in the range of $0.15 to $0.19.

Operator, we are now ready for Q&A.

(Operator Instructions) Our first question will come from the line of Vivek Arya with Bank of America.

And thank you for all the updates on 5G.

Matt, I had 2 questions.

First, on 5G.

I think you mentioned you're starting to ship to your lead customer a quarter ahead.

I assume that's just a partial quarter ramp, and then the real ramp should be in January.

And longer-term, I think you gave a $600 million-or-so opportunity.

Is that just with the lead customer?

Or does that include your entire 5G opportunity with multiple customers?

Great.

Thanks Vivek.

Yes, I'll answer both of those.

So I think you're right, Q3 is definitely a partial quarter and it's the initial production.

Originally, we had been signaling going back some time that we would have this initial production in the fourth quarter, so we were able to pull some in, which is I think a great achievement, where our operations team and design teams have done a fantastic job.

So we'll get some shipments in Q3, with the ramp, a full quarter's worth of contribution in Q4.

And then on your second question, the $600 million is the all-in number.

That's with -- think of it as our lead customer plus our second customer, plus factoring in other opportunities that we see in the pipeline.

And of course, we said, depending on base station units and percentages of share of our customers, it could go higher.

But again, the way to think of it is the $600 million is what we can -- what we've won so far.

And then anything that we can get that's above and beyond that on some of these new opportunities we're pursuing would be incremental.

And again, just to be clear, this is a run rate number we believe is achievable over time.

So this isn't something that's going to start like next quarter as an example, but over time, we believe once these all roll into production, it will be a substantial revenue driver for the company.

Got it.

And for my follow-up, on the storage side, sales are down about 30% from their peak quarterly level.

I realize that there is a Huawei effect in there.

But what does recovery look like for storage?

Do you think revenues get somewhere in between where you were at the peak versus where you are right now?

And what would be the catalyst for recovery?

And let's say, if you don't see the recovery path in storage, are there other strategic alternatives that need to be considered?

Okay.

Got it.

Yes, so I think at a high level, it's difficult to give you a precise answer, so I think it's probably what you said, which is it's somewhere in between.

I mean obviously, we're taking this 1 quarter at a time.

I think the positives we see certainly in the short term is that this business appears to have stabilized after having dropped fairly precipitously for us as well as the rest of the industry, and obviously, guiding up sequentially in Q3 is a positive sign.

So I think I'd limit my commentary to probably the short term, which is, at least, we're seeing some positive signs, but -- and we do expect, by the way, that as we look into next year when you sort of asked for catalysts, I think one is obviously we're going through this inventory rationalization at the customer base.

But we also have new design wins and new programs that we're in.

I alluded to the traction in nearline as an example, which are programs that are currently in production as well as next-generation, I think that's going to benefit us heading into next year as well as the new opportunities in SSD that come out of this do-it-yourself framework that we've discussed as well.

So I think there's some sort of just general recovery, plus we do have new design wins and traction.

And as a result, I think on your third question, the -- storage remains to be an extremely important business for us.

It's a profitable business.

It's one where we have strong technology leadership and market position, and we have a lot of customers that are counting on us to keep delivering the solutions that we are.

So we're investing in this area, and we do see it as an important part of the Marvell story, having really compelling offering of storage, networking and compute or processor assets that, from a system point of view, are very important to our customer and offers a breadth that really very few to none of our competitors can provide.

And our next question will come from the line of Karl Ackerman with Cowen.

Two questions for me as well, please.

First, you indicated in your prepared remarks that you received some earlier shipments of 5G products in the October quarter or in the upcoming October quarter.

How much 5G revenue do you expect in the October quarter?

And what is the expected ramp within the January quarter?

And I guess, from a qualitative perspective, what's the right way to think about your 5G opportunity over the next few quarters?

Karl, I'll answer the near-term question.

Matt can add about the future of 5G.

So we just started first [two] production shipment, so you shouldn't expect -- it's a very -- it's a small revenue in Q3.

Certainly, we expect that Q4, you will see the sequential ramp from the 5G product line.

In the longer-term, and Matt, you probably can give more longer-term, we talk about $600 million, in a few years, opportunity.

Of course, it would take time from the initial ramp to get to that, the $600 million annual run rate.

So that's -- hopefully, that gives you some color about the near-term and the longer-term view.

Yes.

Sure.

And maybe I'll add to that, Karl, as you think about -- as you said qualitatively, how do you think about this business heading into next year.

I think the first is the base station per percentage of shipments, as I highlighted, even the external analysts who have got a range of estimates, but they're all kind of converging, that's going to grow significantly next year.

It's probably 10% of base stations this year.

It will be north of 30%, maybe 40% next year.

So that's one positive.

Deployments for us with our lead customer are going to really be for new base station shipments deployed in Korea.

Obviously, Japan's going to deploy.

There's a lot of activity there, the Olympics being one of those catalysts, and also the United States.

Second, there are shipments that are going to ramp up as well, where there's going to be a replacement of existing line cards that are installed today that are based on discrete solutions that are going to get replaced [with their] optimized solutions.

So that's also going to be kicking in.

And then, of course, you have starting at the end of next year our second customer coming online.

So we see it as a layering effect.

And the timing of all of these pieces, really hard to nail them down exactly.

This is a fast-moving and dynamic market, but you should expect that in the calendar year -- next calendar year, this business should ramp up very quickly.

That's helpful.

For a last question, if I may.

I'm curious what your own view is on the campus networking upgrade cycle for your core switch and PHY business.

I mean should we assume that the growth opportunity is more nascent?

Or is the October quarter lull a transitory event?

Yes.

I think we should separate it into 2 pieces.

I think one is what's the business conditions today, and there's clearly a pause and a slowdown going on driven by a lot of effects, which we've seen both from the effects that -- with -- of Huawei, obviously, as well as recent announcements from large U.S. networking OEMs as well.

So I think that's sort of out there.

That's the current environment.

We remain very bullish on our design win capture that we've had, that business.

And again, prior to the downturn, we were growing that enterprise and campus switching and PHY business fairly robustly.

So while we see a short-term pause mostly driven by the macro issues, the pipeline and the design win momentum has actually continued to accelerate.

And whether it's our new 12.8 switch platform, which is a brand-new market that we're servicing, the fact that we've gone from basically announcing that product earlier this year to now being able to say we've actually secured design wins that are going to go into production, I think that's very positive.

On the switch side and on the PHY side, I highlighted a number of examples in my prepared remarks.

So I would just kind of separate the 2. Currently, there's clearly a lot of weakness in the market and a lot of choppiness, but we believe when that settles down, our design win position is going to be -- continue to be very strong and drive longer-term growth once we're back to a normal situation.

And our next question will come from the line of Quinn Bolton with Needham & Company.

I just wanted to ask first on the 5G market.

When you announced your second customer win, you've talked about the possibility of winning at additional customers beyond that.

And just wondering if you could give us any update on wins beyond your first 2 customers.

Is that opportunity still out there?

Has it gone away?

Just any update you could provide on sort of options beyond your first 2 customers.

Sure.

So nothing formal to announce today or I would have called it out in my prepared remarks.

But what I would say is the activity level, not only with our 2 kind of lead customers, but also beyond that, is quite robust right now.

I mean the fact that we've got strong offerings now in baseband, in transport, in massive MIMO, which was something that really wasn't there sort of earlier this year, fronthaul solutions.

And then obviously, with Avera coming in, the DFE capability and radio solutions, I think are -- and then, add on top of that, request to continue to improve and kind of optimize our Ethernet switch products is that part of the base station.

The connectivity continues to evolve.

We have a lot of design activity, and it's not limited just to 1 or 2 customers.

It's really limited -- I mean it's really much broader than that across a lot of different technologies.

So design win opportunity is -- the pipeline is very large right now and there's a lot of activity, and that's how we would characterize it.

So hopefully, we continue to be successful here, and we can announce sort of the realization of some of these wins as we go forward.

But it's -- we're pretty excited about our technology lineup because we really have all the key pieces that our customers are looking for.

Great.

And then just on the storage side at Flash Memory Summit a few weeks back, you guys announced your new, I think, Gen 4 PCIe controller that was DRAM-less and allowed a fewer number of channels, so smaller die size.

And I'm just kind of wondering, as you look at the launch of that new controller, does that have a negative ASP effect or negative margin effect as we're thinking about the SSD business going forward.

Or do you think that that new product comes in at similar margins to the business historically?

Hi.

This is Jean.

So the new product launch we're very excited about.

The gross margin, it's actually very similar.

Our team has done tremendous work to really try to design, to optimize the design, the performance, to make sure we continue to drive to get to the value of our IP for our design, for our expertise.

So I would say, in general, you should expect us to have a very similar gross margin in our SSD, HDD market.

Yes.

And I'd just add, I mean, we actually highlighted that this product is in 12 nanometer, so we actually are leveraging both of our -- both our architectural experience as well as process technology leadership, which we've had for some time in the SSD area, and we kind of continue to push the envelope there.

So that obviously, coupled with time-to-market advantage, ends up being a good situation with respect to margins and the kind of value you can deliver.

So...

And our next question will come from the line of John Pitzer with Crédit Suisse.

Matt, my first question is on the networking side.

In your prepared comments, you kind of called out 2 things, one, China's strength in the data center that you might think is inventory-related, which I think is reasonable if we try to stay on the right side of conservative.

I'm just kind of curious if you could kind of size what you think that might be and as you looked out to the fiscal third quarter guidance, whether or not you're still embedding strength there.

And then secondly, you also mentioned within the guide a single networking customer where you saw sort of a downward revision in their orders.

I think you don't want to get too specific about the customer's name, but any more color you can give around what's going on there?

Is that something that you feel is pretty transitory?

Is it a reflection of just macro uncertainty, or is there something else going on?

And then I have a follow-up.

Yes.

No problem, John.

So I think, yes, I think you captured it right on the dynamics in networking.

High level, obviously very choppy here.

On the first part of the question, on the China strength, I mean, we're fairly detailed in how we look at our demand, and in particular, a lot of the China business gets realized through distribution.

So we did see very strong POS.

And as we've double-clicked through it, a number of these customers seem to be ordering a lot more than they normally would.

And so a couple of things there.

One is it's not a huge number, so just to size it, it's not an enormous number, but still, we felt it was appropriate to call that out.

I would say also, the way we're managing our channel is very much with that in mind.

So just to give you a sense, I mean, we've -- actually, we're running our distribution days of inventory about 2 weeks below our target levels right now, just bearing in mind that if there is some inflation in the POS, that we're going to be on the right side of that.

So that's happening there.

Again, hard to pinpoint, but we think that's what's going on.

And then on the second one, that -- the situation we called out, I mean, it's definitely something that's kind of evolved recently, so we've seen this sort of situation develop over the last several weeks.

And obviously preparing for our guidance, we've had to really assess that situation.

So it's hard to say.

I mean I think you have to look at the broader commentary out there in the market from what our end customers are saying.

But if you look back, I'll just make one final comment, when -- if you go back to the beginning of this downcycle, we were fairly early there because of the way we reported and also, our exposure to storage to see a lot of this weakness fairly early.

And so that first leg down, if you remember earlier this year, was in storage.

That was followed by a compute leg down, I'm talking about industry-wide, where servers have weakened.

And I think a lot of investors had asked us are we going to see that in networking, and while we don't predict that far out, we definitely see, from our lens, the networking business also now being weak.

So that's sort of the deal.

I think, yes, I think I've answered the China and the large customer one.

That's really helpful, Matt.

And then just as my follow-up.

I was a little confused by your answer to Vivek about the recovery in storage.

Not predicting when you get back to next cycle peak, but if you exclude the vagrancies of what goes -- typical vagrancies of the storage market, especially given some of your new product launches and repositioning, refocusing of the business, is there any reason why, at some point, you don't get back to where you were in the last cycle peak?

And has your view of the growth rate of this business longer-term changed at all?

Yes.

Sure.

No, look -- yes, I think it's worth clarifying then, John.

So the simple way to think about it is, and again, understanding we have momentum in that business from a new design win point of view on the markets I mentioned, but if you look at the PC exposed portion, we did a pretty good job on the kind of the first transition, I would say, when HDD was converting to SSD.

And if you remember, over the last few years, we've sort of managed that HDD going down by winning new designs in SSD.

We've said, even going back to our Analyst Day that that remaining HDD to SSD conversion in notebooks, which is typically at the lower end, when that conversion happens, we're probably not going to benefit from that.

We really haven't put a lot of R&D into that area.

That's kind of a lower-end client opportunity, and our mission has really been around enterprise and data center.

So I think that's -- when I look at sort of the gap of it, hey, can you get back to peak and when, I think we've got that headwind and that's sort of we've been very transparent about that one.

Again though, being offset by even next year, where we're very confident in our ability to ramp up our new do-it-yourself solutions in SSD, some of those will be forthcoming.

And then in the nearline programs we won, which we believe we're well positioned on both at 16 terabyte and then even going to 20 and beyond, our design win footprint is actually quite strong.

So that's the puts and takes.

Where -- and again, I am being very careful to not call a complete sort of trough to peak on this one or peak to trough, but we do see that business clearly recovering.

How high it gets back up to, we need to kind of take this 1 quarter at a time.

And our next question will come from the line of Blayne Curtis with Barclays.

Maybe just 2 on 5G actually.

You mentioned that you sampled your chip, and it's actually in trials now.

I was wondering if you could just walk us through that time line.

Is there anything else you have to do from -- on your side from a silicon perspective?

And then you mentioned a follow-on generation, if you could just -- when would that start to kick in?

And then I just wanted to revisit, you said greater than $600 million.

I think most people kind of thought about a 4x of content, like you talked about, and if you just take the 2 customers to kind of get there and you've talked about some incremental wins.

So maybe just kind of walk through that $600 million.

And I know you said greater than $600 million, but I was just kind of curious, it seems like you're getting every call, added wins, and there are still opportunities [with] incremental customers.

So if you could just comment on opportunities above that $600 million, it would be helpful.

Sure.

So on the first one, which is really 2 parts, with respect to ramping on 5G and what's still needed there, yes, those products are -- they're done.

We're going to be able to ship them this quarter.

There is certainty a lot of work there with respect to getting the yields where we want them, the production ramp.

I mean, I'm saying this because I don't want my team to come back to me and say you made it sound so easy.

This is a very challenging ramp for our company.

But from a silicon perspective, for the first shipments, we're -- that part's done.

So now, it comes down to execution and how fast we can ramp and matching the demand signals from our customer.

So that's positive.

The second one is we did win the next-generation product, which we're very pleased about.

That takes us down into the advanced node process area, which is a very big positive, I think, for our customer as well as just for us.

But that's a design that's going to be kicking off now.

So think of it as we've got very sort of solid runway on our current generation of products to last us at least for the next several years.

And so again, this is just really a testament to the fact that we've continued to engage well with our lead customer, and we've got the road map in place with them.

The $600 million, just to kind of set the stage, I think on -- first, we felt it was appropriate to just talk about it because we get a lot of questions from investors on this.

And so I think the baseline case is really you can sort of take the lead customer, you can use the 4x increase in content, right, and that would get you to roughly, call it, $400 million-or-so.

You would add the potential contribution from our second customer onto that, and then obviously, there's more that could come later.

And then the timing of all these things are still to be determined.

So we're trying to establish what a base case would look like, Blayne, but certainly, for the broader investor community, we get this question all the time, and we felt it would be better to just get that out now.

But you should assume these other wins that were -- if we achieve them, would layer on top of that number.

And our next question will come from the line of Gary Mobley with Wells Fargo.

I have a question about Aquantia.

Jean, I think you mentioned Aquantia's revenue run rate normalized at the time of the acquisition announcement.

It was around $100 million.

Most recently, per the 10-Q filing from Aquantia, that revenue was around $9 million.

Is that just inventory digestion?

Or do you feel any differently about that business being above or below or at the $100 million mark?

Hey.

Thanks, Gary, for the question.

So as you probably have seen from Aquantia's 10-Q filing, they are key customers of Aquantia's.

They have been reducing inventory, so that's why it has a significant impact on Aquantia's revenue.

Frankly, we think it's a great operational discipline to really balance the inventory and the supply demand to make sure you're really [lean] on the inventory at the customers.

So that has always been Marvell's practice.

We typically really don't have inventory at the customers unless they need it.

So that process is ongoing.

I would say, the Aquantia team continues to win designs, and also, they had been ramping up new designs with the key networking customers.

So we actually really are quite comfortable.

Once this inventory adjustment period ends and going into next fiscal year, which is our fiscal '21, they should be able to ramp back to the $100 million per year run rate we articulated during the deal announcement.

And so we feel pretty good about that.

And our team, they're doing the integration planning, and everything we have seen is quite consistent with what our deal assumption has been.

So we'll give you more updates when we get close to the deal closure.

Okay.

And in the other category, the products within that, I believe you've communicated are in a long-term decline, but it certainly had a good first half of fiscal year 2020.

And so can you speak about the strength and what's driving the growth and maybe any sort of update on your long-term expectations for that business?

Oh, yes.

So it's about our other revenue category.

So let me start by saying, right, this product category includes our printer business and some of the other consumer-related business.

Those businesses actually have a very long tail.

It will last for multiple years, but it could be very lumpy because sometimes, the customer wants to have a long-term buy, so that's what happened in Q2.

You actually saw that product increase in revenue.

That's because of a long-term buy.

In general, the way you should think about it or model it is, in the longer-term, you should see the high single-digit year-over-year decline last for many years going forward.

So that's how you should think about modeling this business going forward.

And our next question will come from the line of Ross Seymore with Deutsche Bank.

Our next question will come from the line of Harlan Sur with JPMorgan.

Good to see the quarter-on-quarter resumption in your storage business, but it still looks like you guys are shipping about 13% to 15% below normal consumption levels.

I know that more of the headwind for Q3 is client storage.

But on nearline, it also seems that the mix shift in the second half of this year is more towards 14-terabyte platforms, where your competitor has more controller share and less so maybe to the 16-terabyte platforms, where you guys have a strong position.

So has the slight mix shift also moderated your view on the slope of the recovery of the storage business kind of second half of this year?

Yes.

So I appreciate your question.

So the way to think about our storage business, you're right, is from a -- we have been undershipping the market.

And our storage business, when you look at it overall, I appreciate the question about the 14 terabyte versus 16 terabyte, and our position going forward on 16 terabyte and above, certainly, is much stronger.

But overall, when you look at the dynamics of our storage business, it's actually we continue to see just a stabilization, not like a significant pickup.

So we are looking at it quarter-by-quarter, like Matt said earlier.

And overall, the way to look at it is, actually, the data center, which -- and the enterprise, which include both HDD and SSD solutions, you should see that revenue continue to go up in the second half and also going into next year.

But as far as the dynamics of 16 terabyte, 14 terabyte, actually, right now, at this point, it's -- the revenue is relatively lower percentage.

So it's actually not impacting the overall revenue dynamics.

Okay.

And then a question for you, Matt.

Can you just give us an update on the cloud hyperscale qualifications on ThunderX2, your ARM-based server platform?

And are you still on track to ramp some of these customers in the first half of next calendar year?

And then on ThunderX3, your 7-nanometer platform, I know you guys taped this chip out.

Is it back from the fab?

Have you started sampling this product?

Any early performance metrics you can share with us?

Yes.

Great.

Thanks for the question.

So yes, no update on X2 from what we've said before other than those quals that we referenced before are proceeding, as we had outlined before, right?

So that's something that we do expect to ramp up next year.

And with regards to X3 and performance data, these products actually have quite a long fab cycle time, and it's going to take us some time to really provide that information.

So I don't have an update there for you at this time.

And our next question will come from the line of Tim Arcuri with United Bank of Switzerland.

This is John on -- calling in for Tim.

That's UBS Securities.

Just kind of wondering about the China strength that you see from customers not being Huawei.

Kind of wondering, was that isolated to this past quarter, or when did you start seeing that?

I'm just trying to gain some kind of knowledge on when that might have started.

Were they starting to pull in orders a lot earlier than last quarter?

Yes.

Hey, John.

So yes, it's kind of interesting.

I mean if you actually go all the way back and you look at even the end of last year, when the tariffs were announced, there was unusual customer behavior in China that even started then, right?

And you may remember this, but we had 1 quarter in our Q3 a year ago where we said our networking business was up like -- Marvell core networking was up like 29% year-over-year, which we had said basically looks like there's people pulling in.

So there's been a lot of disruptive activity going on given all the trade tensions and macro issues in China.

Specifically, this latest go-around, I would say, we saw that during the second quarter.

I'd say we saw the POS accelerate and the customer pulls.

But you should just kind of assume right around the time that Huawei was put on the entity list, I think that triggered some additional ordering from the customer base there in terms of being worried about supply chain disruptions.

So that's probably the time frame, but I just wanted to bound it.

It wasn't like, hey, it just sort of showed up out of -- there's just been a lot of noise in the China market as a result of the trade tensions that are going on, which are obviously widely publicized and documented on a daily basis.

So John, just as Matt mentioned earlier, right, is the firstly the exposure for us, this is a channel business and a China channel business, and so it's a very small percentage of our revenue.

And also, as Matt mentioned earlier, we have been really prudent to manage the channel inventory and took down channel inventory significantly, really to think through this kind of end-market behavior.

So I think it's a small exposure for us, but it is an interesting phenomenon.

Okay.

Great.

My follow-up.

Going back to the 5G business or what you expect coming into 2020.

I know in the last 4G generation, I mean, order rates were always pretty lumpy.

I'm sure it's going to be like that for the 5G.

But in the initial ramp into 2020, what are you guys seeing, or what can you guys guess going into 2020?

Yes.

It's -- I mean, John, it's too early to call that slope.

I think the -- another caller was asking sort of the same thing.

At this point, we just need to stick to more qualitative because, as I mentioned, there's a dynamic where there's multiple countries deploying.

There's our customer's ability to get share in those countries.

There's some replacement opportunity we're going to get.

We got a second customer coming in.

And you're right, I mean, that the carrier business historically for all suppliers is a lumpier business.

So you have to layer some increased volatility on top of that.

So as a result, it's difficult to give you this ramp at this point, John.

But certainly, as we get through Q4 and we're looking out into the next fiscal year, I think we'll be in a better position certainly lining up with our customer and the market on what's happening.

But just to be clear, I mean, calendar '20 for us is going to be a very strong year.

There's no question.

When it fully ramps and what quarter it -- by quarter, I think we're too early to call that one.

And our next question will come from the line of Ambrish Srivastava with Bank of Montreal.

I apologize for any background noise.

I had a question on the networking business on the reported quarter.

You talked about the 4G pause as customers transition to 5G.

So the question really is, is there a bottom to that 4G that we should be thinking about?

What is the mix of the 4G within the networking?

And so we're all focused on the $600 million, but is there cannibalization going on, off the 4G, so the net-net is not really $600 million and it's $600 million minus whatever that business is, which has yet to stabilize?

And then I had a follow-up.

Sure.

Yes.

So I think the way to think about it is kind of point one is, and we sort of called this out at the Analyst Day last year, today, the current 4G carrier business for us is not a big business, right, and that's one reason we're excited about 5G because obviously we have a big content growth and then all the dynamics associated with that.

So today, that's not a terribly large business, and I don't think the way to do it is try to figure out how much 4G we're still shipping versus how much 5G because the 4G is almost gone if you think about it because all the new systems we're in at our lead customer -- pretty much all their new shipments once they're ready to go, even if they're not ready to light up 5G yet, they will ship as a 4G base station.

So that's why we've just sort of gone to this model to say, look, the business run rate was x amount before, at kind of its normal 4-quarter average, let's call it, and then you just take the average content increase and that's where you get the $400 million from.

So I don't know if that's helpful, but it's not one that I don't think we spend a lot of time on in terms of -- because, one, it wasn't a big business for us, and two, most of that has really slowed down in anticipation of this new ramp-up, which will cover 4G and 5G in the next year.

Got it.

No, that's helpful.

And I asked because you brought it up again about the slowdown for the reported quarter.

But it's clear.

I understand.

And then my follow-up is a little bit more architectural maybe.

If you could just maybe share your insights.

So you talked about FPGA replacement.

Can you just help us understand what exactly you meant and what exactly -- what is going on on the architectural side?

Is it as simple as these products are now going into production, so the prototyping was in FPGA and now we are moving to standard products created by you?

No.

I think, yes, I think you just said it.

I think that what we meant by that really was that in -- and this isn't just for 5G.

I mean if you go dust it off and go all the way back to the 2 to 3G transition and the 3 to 4G transition, there's always been this wave of very strong FPGA deployments that, in some cases, cover baseband and cover the radio head and they cover all the major applications because the standards are changing and the OEMs are focused on time-to-market.

And then obviously, once optimized solutions become available and the standards are set, then ASICs take over.

And so what we've seen is just simply that I think one is kind of qualitatively, from 2 to 3G, 3 to 4G and 4 to 5G, the percentage of conversion from FPGA to ASIC appears to be increasing, certainly because the processing density is increasing a lot and therefore, just to optimize on power and capacity and all the things that the customers and the operators care about are optimized, there's clearly a need to move to these custom solutions.

But that being said, FPGAs have always been an important part of the wireless communications network design, but we are seeing very robust activity where anybody can to basically leverage the IP set, we've got to go target, not just the baseband, but these remote radio head applications.

And obviously, we have a very strong transport business, which is just our embedded processor.

So I hope that's helpful.

It's not anything that's terribly different from the prior generations.

It's just that Marvell today has a very strong offering in this area and therefore, we think we can capture a lot more share than we had, say, in 4G.

And our last question will come from the line of Ross Seymore with Deutsche Bank.

Can you hear me this time?

We got you, Ross.

Yes.

Sorry about that last time.

Lots of great questions have been asked and answered already.

So I'll try to hit on 2 that haven't.

The enterprise weakness, the networking side that you talked about, is that localized to one customer, or is it something broader?

And Matt, do you believe it's something that's just beginning, or is this kind of a macro trend that is generally along the same timing we've seen across the weakness in most of the sectors?

Yes.

So I would say, point one is it is broader, and broader means, obviously, a number of customers, some of which are either flat or they're slightly down or they're just not sort of meeting their forecast.

And then we did also call out that one large customer.

It was quite pronounced in terms of the change in the demand signals and the outlook that we saw heading into this call we had today.

So it's not -- we're not signaling it's just one customer, we're saying there's broader weakness certainly from our lens.

But clearly, our results and our guidance are driven mostly by the large customer, just given their contribution to our overall revenue.

Got it.

That's helpful.

And then maybe one for either you or Jean.

You guys have been busy in this last quarter, lots of moving parts, acquisitions, divestitures, et cetera.

If we look forward, whether it's a year or 2 years down the road, you guys before had a long-term target, I think, of the gross margin to be at above 66% and operating margin, I think, both of those on a pro forma basis, about 35%.

If we think about the Marvell of tomorrow after these 3 deals, are those targets still in the right ballpark?

Does it expedite or delay the ability to achieve those?

I just kind of wanted to see longer-term what the company's going to look like after these deals versus at the time that you gave those long-term targets at the last analyst meeting.

Hi, Ross.

And thanks for the question.

I think if you take a level higher to think about how we think about our business model, it continued to be the same driving principle is we really want to maximize the returns for our stakeholders, which means generate a strong free cash flow, strong earnings and strong operating profit.

So when we think about our long-term target model, which was set up like almost like more than a year ago, what we said is, hey, the top line revenue growth will be in the range of 6% to 8%.

Gross margin is going to be around 66% and the operating margin, 35%, as you mentioned.

So right now, when we look at the progress the company has made during the last year or more than a year, we have gained tremendous more momentum than the Analyst Day in 5G in the infrastructure market.

Matt talked about the opportunity and the upside, which was not part of the original Analyst Day.

And of course, both Aquantia, Avera and the Wi-Fi divestiture were not part of the original business model.

So the way we're thinking about it, we can give you some perspective, but we're not updating the model.

The perspective is basically, on the top line, if we can drive the opportunity where we have design wins and everything else, in a normalized macro environment, we should see our revenue growing more toward the high-end range of our model.

On the gross margin side, I think the way -- first, I want to say, as a company, we are laser-focused on gross margin improvement and that you have seen us, we have had a really strong track record to improve gross margin.

Going forward, there are a few takes and puts of this gross margin dynamic.

I think that, certainly, the storage business is stabilizing.

The nearline, the data center will help us increase the gross margin.

And also, secondly, on the networking side, our automotive business, our switch, PHY refreshed portfolio, they're all going to help to contribute positively to the gross margin.

On the 5G side, it's going to be a significant revenue driver, but some of our 5G products, actually right now, currently, have a lower than corporate average gross margin.

So the way you should expect us to do is, going forward, we're going to focus on improving the yield, the test time because it's the early ramp.

There's certainly a maturity of production, and on the yield side and test time side, which will improve gross margin.

So we continue to think when we balance everything, our gross margin should be still very high, maybe in the 63% to 65% range.

But overall, the way to think about it is because our top line revenue growing is going to be much faster, the gross margin dollars are actually going to be much higher, which on the operating margin side, we definitely are very committed to the 35% operating margin.

We do think, structurally, if we can grow the revenue, our model would leverage quickly to deliver that kind of operating model.

So hopefully, that helps you to think through the longer-term model.

Ladies and gentlemen, this concludes our question-and-answer session for today.

So now it is my pleasure to hand the conference back over to Ashish Saran, Head of Investor Relations at Marvell, for any closing comments or remarks.

Thank you, everyone, for joining us today, and we look forward to seeing you at the upcoming Citi Technology Conference in New York City and the Deutsche Bank Technology Conference in Las Vegas.

Thanks.

Bye.

Ladies and gentlemen, thank you for your participation on today's conference.

This does conclude our program, and we may all disconnect.

Everybody have a wonderful day.