威騰電子 (WDC) 2019 Q4 法說會逐字稿

Good afternoon, and thank you for standing by.

Welcome to Western Digital's Fourth Quarter of Fiscal 2019 Conference Call.

(Operator Instructions) As a reminder, this call is being recorded.

Now I would like to turn the call over to Mr. Peter Andrew.

You may begin.

Thank you, and good afternoon.

Before we begin, I would like to remind everyone that today's discussion contains forward-looking statements, including product development expectations, business plans, trends and financial outlook based upon management's current assumptions and expectations, and as such, does include risks and uncertainties.

We assume no obligation to update these statements.

Please refer to our most recent financial report on Form 10-Q filed with the SEC for more information on the risks and uncertainties that could cause actual results to differ materially.

We will also make reference to non-GAAP financial measures today.

Reconciliations between the non-GAAP and comparable GAAP financial measures are included in the press release and other materials that are being posted in the Investor Relations section of our website.

With that, I'll now turn the call over to Steve Milligan, our CEO.

Thank you, Peter, and good afternoon, everyone.

With me today are Mike Cordano, President and Chief Operating Officer; and Bob Eulau, Chief Financial Officer.

We ended fiscal 2019 with a significantly expanded market presence in both capacity enterprise and client SSDs and began building momentum for our enterprise SSDs.

The Western Digital team maintained its focus on execution, further improved operational efficiency and delivered the strongest product portfolio in the company's history.

We accomplished all of this in the midst of various challenges on both the market and geopolitical fronts.

In hard drives, we continue to focus on the higher growth, higher margin portion of the market, namely capacity enterprise.

Our technology leadership, demonstrated by our 14-terabyte drive, combined with the breadth of our portfolio across a range of capacities, enabled us to gain market share.

In flash, we doubled our client SSD exabyte shipments due to the strength of our portfolio.

I'm also pleased that our latest enterprise NVMe product is commencing its revenue ramp in the current quarter at both the major hyperscale and OEM customers, positioning us for strong growth in this area in fiscal 2020 and beyond.

In May, we announced a formal agreement with our flash JV partner, Toshiba Memory Corporation, to jointly invest in the K1 fab.

This new facility will provide needed space for the continued transition to future 3D flash technologies.

On the cost front, we accelerated the streamlining of our hard drive manufacturing footprint.

We also made significant progress in lowering our operating expenses, and Bob will provide additional details.

In terms of our June quarter performance, overall results were within the expectations we provided in April.

In hard drives, the ongoing customer transition to our 14-terabyte capacity enterprise drive drove meaningful sequential revenue growth.

In flash, revenue declined sequentially as an improved product mix was more than offset by weak pricing.

Our expectation for a stronger demand environment for the second half of calendar year 2019 remains intact for both our flash and hard drive products.

Before I hand the call over to Mike, I want to provide some comments on 3 additional topics: Huawei, the Yokkaichi JV fabs and flash inventory.

Effective May 16, the U.S. Commerce Department's Bureau of Industry and Security, or BIS, added Huawei to its entity list.

Western Digital stopped shipments to Huawei at that time.

After discussions with the Commerce Department on the relevant laws and regulations, we determined that the entity list regulations do not apply to products that were previously qualified for Huawei and we started resuming shipments of those products to Huawei in mid-June.

Western Digital's business activities that are subject to the entity list restrictions remain on hold, but we have applied for the BIS for a license to resume those activities.

Turning to the Yokkaichi fabs, an unexpected power outage occurred on June 15, effective -- affecting production operations at the flash fabrication facilities operated by TMC.

Since that event, the Western Digital and TMC teams have worked diligently on recovery activities, and as of now, nearly all of the equipment in the fabs has returned to normal operations.

Consistent with our prior comments, the disruption to fab operations caused a loss of approximately 6 exabytes of wafer output for Western Digital, with most of the wafer output impact expected to be contained in the September quarter.

As Mike will explain, we anticipate we will be able to largely mitigate any revenue impact for the September and December quarters.

With respect to flash inventory, based on recent industry announcements and a stronger demand environment, we estimate that any excess inventory will be substantially reduced by the end of this calendar year.

It is still too early to characterize it as a trend, but we are seeing a more stable pricing environment for our flash business.

In summary, we are successfully navigating dynamic market conditions.

Our hard drive business is performing well, and we believe that the flash market has reached a cyclical trough.

The secular trends of data growth and its increasing value, key drivers of our business opportunities, are strong.

Our continued transformation into a data infrastructure company is on solid footing, and we look forward to reporting on our progress.

I want to thank the Western Digital team and our partners for their ongoing support.

With that, I will now ask Mike to share our business highlights.

Thank you, Steve, and good afternoon.

I'd like to begin by amplifying Steve's perspectives on fiscal 2019.

We made significant strides in broadening and strengthening our product portfolio, particularly in capacity enterprise drives, enterprise SSDs, client SSDs and embedded products while maintaining our strength in our retail offerings.

In fiscal 2019, we expanded our lead in areal density and established the 14-terabyte capacity enterprise drive as an industry standard.

We were the first to announce the 16- and 18-terabyte products based on our energy-assisted recording technology.

In flash, we're the first to successfully commercialize our 96-layer 3D technology across a range of products, and we will continue to implement it across the rest of our portfolio in the remainder of calendar 2019.

Moving to highlights for the June quarter.

In data center devices and solutions, a broad adoption of our new capacity enterprise products drove a mid-single-digit year-over-year exabyte shipment growth for the first half of calendar 2019, outpacing industry growth resulting in exabyte share gain.

Coupled with our continuing expectation for stronger demand in the second half of calendar 2019, we have high confidence that our capacity enterprise exabyte shipment growth will meaningfully exceed 30% for this calendar year.

With this strong demand, we are seeing some signs of supply constraints within our capacity enterprise product portfolio.

Our 14-terabyte capacity enterprise drive qualification activities are complete at virtually all hyperscale and OEM customers, and the accelerating adoption of this higher capacity drive further extended our leadership in this category.

There have been some market rumors suggesting that customers are skipping the 14-terabyte generation to adopt the 16-terabyte capacity point, but our demonstrated performance with our 14-terabyte offerings should invalidate those claims.

In addition, we are on track to introduce our first energy-assisted recording 16-terabyte CMR and 18-terabyte SMR hard drives later this calendar year.

In enterprise SSDs, I'm pleased that our internally developed NVMe platform is commencing revenue ramp at hyperscale and OEM customers in the current quarter.

Next week, at the Flash Memory Summit, we will be announcing additional enterprise NVMe products implemented with X4.

We expect a significant acceleration of our enterprise NVMe product revenues over the course of fiscal 2020 and beyond.

Within client devices, surveillance continues to be a long-term driver for client HDDs.

We are leading the transition in this area to drive managed SMR.

In the September quarter, we expect to see a strong revenue ramp of our drive-managed SMR products, allowing us to meaningfully participate with compelling surveillance portfolio.

For client SSDs, shipments of our NVMe-based products grew significantly, representing nearly 50% of our client SSD revenue.

As price points have declined, we are seeing customers migrating to higher capacity points, and average capacity per unit of our client SSDs increased 50% on a year-over-year basis.

In client solutions, we are very pleased with our continuing success in the external SSD category with further expansion in our market share.

Average capacity per unit for flash devices...

You skipped a page.

Okay, I'm sorry.

In mobile and embedded, we continue to strengthen our product portfolio, including the announcement of a 3D NAND automotive grade product, broadening our automotive offering with unique capacities and capabilities.

Due to market conditions and geopolitical dynamics, we reduced our participation in mobile during the June quarter and expect to see immediate recovery of our participation in mobile in the September quarter.

I would now like to make a comment on Huawei.

As Steve described previously, we stopped shipments to Huawei for about a month, having subsequently resumed shipments.

As a result of the stoppage, we estimate there was approximately $100 million revenue impact in the June quarter.

In addition, given the ongoing uncertainties, our near-term opportunities at Huawei have been reduced.

In Client Solutions, we are very pleased with our continuing success in the external SSD category with further expansion in our market share.

Average capacity per unit for flash devices grew 36% year-over-year.

The power of our brand and our customer's preference for the performance and reliability of our solution allowed us to maintain our leadership position.

From a flash supply perspective, as Steve mentioned, the JV fabs have essentially resumed normal operations.

I will share a few additional details on the impact of the fab power outage and Bob will provide the financial implications.

We anticipate the incident will result in a reduction of Western Digital's flash wafer availability of approximately 6 exabytes, consistent with our prior estimate.

For the balance of calendar 2019, given constrained supply, we are working closely with our customers to align to their increasing demand.

For calendar 2019, we estimate industry flash bit supply growth in the low 20% range.

Turning to flash technology.

The ramp of BiCS4, our 96-layer 3D flash technology, is progressing well, and we expect bit crossover exiting the September quarter.

We expect to ship BiCS4 technology offerings in each of our product lines before the end of calendar 2019.

Looking beyond the normal course of the business, we're driving an important industry level initiative called zone storage.

This approach is an open source, standards-based initiative which was formally announced in June and will be further discussed and demonstrated at the Flash Memory Summit next week.

From a customer perspective, the key benefits of zone storage include lower total cost of ownership; intelligent data placement; and greater economies of scale for applications such as video, artificial intelligence, machine learning and IoT.

Zone storage will utilize a combination of SMR hard drives and NVMe SSDs, including QLC offering.

The zone storage initiative brings together our deep expertise in both the underlying storage and system-level technologies and is a great example of what the power of our portfolio can deliver for our customers.

In summary, we are entering the new fiscal year with a significantly enhanced product portfolio, leaner cost and expense structure and a persistent drive to address and capture the opportunities presented by the growth in and increasing value of data.

The Western Digital platform continues to strengthen and, combined with our ongoing focus on execution, we are positioned to perform well in a variety of market conditions.

I will now turn the call over to Bob for the financial discussion.

Thanks, Mike, and good afternoon, everyone.

I'm pleased to announce that results for the June quarter were in line with our expectations provided in April.

Revenue for the June quarter was $3.6 billion, down slightly from the March quarter.

On a sequential basis, data center devices and solutions revenue increased 2% as we experienced continued strength in capacity enterprise drives.

Client devices revenue declined slightly as growth in the client SSD market was offset by declines in the client HDD and flash mobile applications.

Client solutions revenue decreased 6% primarily due to lower flash pricing.

In the June quarter, flash revenue was $1.5 billion with a 1% sequential decline in bits and a sequential average selling price per gigabyte decline of 6%.

Hard drive revenue of $2.1 billion was up 3% sequentially, led by strong demand for capacity enterprise drives.

Moving on to cost and expenses.

All of my comments will be related to the non-GAAP results, unless stated otherwise.

Gross margin for the June quarter was 24.2%, with a flash gross margin of 19% and a hard drive gross margin of 28%.

Hard drive gross margins were slightly below expectations due to surveillance and desktop margin declines offsetting improved margin in capacity enterprise drives.

Hard drive gross margins will show the full benefits in the December quarter from our previously discussed closure of the Kuala Lumpur facility.

Excluded from the cost of revenue was a final $67 million charge related to the planned underutilization of our portion of the flash joint venture fab, bringing an end to this program that we announced in late calendar 2018.

This quarter, we took a GAAP-only charge of $145 million to cost the revenue for the unexpected power outage that impacted equipment and operations in the Yokkaichi flash manufacturing facility.

We expect to incur an additional charge in the range of $170 million to $190 million in the September quarter related to this event.

We and our partner will be vigorously pursuing recovery of our losses associated with this event.

Operating expenses were $722 million, at the low end of our guidance as a result of continued progress on our expense reduction targets.

Operating cash flow for the June quarter was $169 million, and free cash flow was negative $179 million.

In the June quarter, we paid $146 million in dividends to our shareholders.

At the end of the quarter, we have $3.45 billion in cash and cash equivalents and our gross debt outstanding was $10.7 billion.

Our liquidity position continues to be very strong with the cash and cash equivalents I just mentioned and the $2.25 billion of undrawn revolver capacity.

I have no concern regarding our debt covenants, especially given the amendments made to the credit agreement in April.

As we generate free cash flow going forward, our first priority will be to reinvest in the business to maximize long-term shareholder value.

After paying our dividend, our next priority with our free cash flow will be to reduce our gross leverage.

Inventory dollars were down approximately $160 million sequentially to $3.3 billion.

Hard drive inventory decreased due to the drawdown of inventory built to enable the Kuala Lumpur facility closure.

Flash inventory also decreased, primarily due to scrap of work in process inventory at the fab as a result of the power outage.

We expect both flash and hard drive inventories to decline further in the September quarter.

Now I will provide our guidance for the first quarter of fiscal 2020 on a non-GAAP basis.

Please note that the September quarter will be a 14-week period.

We expect revenue to be in the range of $3.8 billion to $4 billion.

We expect gross margin to be approximately 24% to 25%.

Please note that this range includes approximately $50 million of start-up costs associated with our investment in the K1 fab.

We're on track with our previously announced efforts to lower our cost of revenue by $100 million per quarter by the end of the December quarter.

Operating expenses are expected to be between $750 million and $770 million on a 14-week basis.

We're on track with our original target to reach $740 million in the September quarter on a 13-week basis.

The 14th week will add approximately $45 million of operating expenses in the September quarter.

We expect interest and other expense of approximately $95 million, and we expect the tax rate to be 27%, plus or minus 2 points.

As a result of this detailed guidance, we expect earnings per share between $0.15 and $0.35, assuming approximately 300 million in fully diluted shares.

For those of you who've worked with me in the past, you know that prudent cash management has always been one of my key areas of focus.

Total cash capital expenditures at Western Digital include the cash CapEx on our balance sheet along with a net change in notes receivable for our flash joint ventures.

For fiscal 2020, our preliminary estimate for total cash capital expenditures is less than $500 million, which is a significant reduction from the $1.36 billion spent in fiscal 2019.

The key drivers for the reduction of total cash capital expenditures are the previously announced planned slowdown in our capital deployments, the timing on the payment for capital equipment and an anticipated increase in lease financing for the flash JVs.

Before I turn over the call to the operator, I would like to note that I'm very excited about the company's position as we enter the new fiscal year.

For me personally, it is a great time to be at Western Digital.

We clearly have momentum in the capacity enterprise part of our HDD business, and in flash, we are successfully ramping our NVMe platforms for client and enterprise applications while maintaining our strong retail presence.

There are signs that market conditions in flash have reached a cyclical trough, and we are now entering the seasonally stronger part of the calendar year.

My goal is to ensure that we maintain financial discipline to deliver the best business results as we look into the future.

With that, I will now turn the call over to the operator to begin the Q&A session.

Operator?

(Operator Instructions) Our first question comes from the line of Aaron Rakers with Wells Fargo.

Two if I can, real quick.

So first of all, underpinning your gross margin assumption for the current quarter, I'm just hoping that you could provide us with a little bit more granularity of how we should think about the trajectory of the hard disk drive gross margin given that, that was relatively weak, I think, versus expectations in this last quarter.

And then how do we think about the gross margin on the flash business not just for this current quarter, but as we start to see kind of the cyclical trough story play out, how we think about the gross margin moving higher over the next couple of quarters?

And then I do have a follow-up.

Yes, so Aaron, I'll address that and then Mike and Bob can add a little bit of color.

In terms of HDD gross margins, if you go back historically, I know you're well familiar with this, Aaron, but we would normally expect our gross margins to be in the low kind of 30% range, say, 31%, 32%.

So we're operating below where we'd like to be.

And we would expect that as we move through the balance of this year and certainly into 2020, we would see our gross margins begin to move more back into that range.

And so that's kind of our expectation from that standpoint.

Flash, I would say that will be similar performance, maybe slightly down.

Kind of hard, but range, let's say, roughly similar to where we're at right now.

A lot of that's going to depend upon the pricing environment, and we're seeing a more stable pricing environment from what we've seen over the last several quarters.

And we would expect that, as we move through the balance of fiscal 2020, we will see an improving gross margin profile on our flash business as well.

Yes.

And Aaron, let me just add, and Bob sort of mentioned this, so in our desktop category, so the drag in the quarter just reported is really we're in the midst of a product transition on desktop and surveillance.

So that was a drag to HDD margin.

And just to reiterate what Steve said, we expect to see trending up sequential HDD margin in the back half of this year and beyond.

Yes, I think the only -- I'll just add one thing.

The other thing I'd mention is we -- on the flash side, we are expecting improvement, but we do have a headwind of the K1 startup costs, which we saw some in the fourth quarter.

We'll see again in the first quarter and throughout this fiscal year.

Very helpful.

And then just as a real quick follow-up on the flash business.

I think one of the notable things that differs in your results here tonight is kind of the capacity shipment number down 1% on a sequential basis, first, I think Hynix and Samsung were like 30% or 40%.

So I'm curious of how you think about that differential.

Is it Huawei-driven?

Is it your lack of participating in kind of the discrete market?

And what is your expectation as far as bit growth into the September quarter?

Yes, so as I talked about in my script, Aaron, this is really an impact to the mobile category and it's both some impact to the geopolitical activity, but also some choices around business that we didn't see was favorable from a profitability standpoint or lower ASP.

So the 2 of those drove the bit shipment delta on our side, and obviously, we think we got more interesting places to spend our bits as we progress through the calendar year.

Yes.

And Aaron, I mean you can -- you'll notice this as well.

Our sequential ASP decline in the flash area was meaningfully lower than others' reported numbers, which -- there's a direct correlation in terms of that data point and our participation in the mobile segment.

And our next question comes from the line of Wamsi Mohan with Bank of America.

A quick one for Bob.

The tax rate that you are forecasting here in Q1 is quite high.

How should we -- what is driving that tax rate for the quarter?

And how should we think about modeling that for the rest of the year?

And I have a follow-up.

Yes, thanks, Wamsi, for the question.

So -- actually, to begin the year, what we do is we forecast the tax rate that we're going to experience for the fiscal year.

And we actually have a very good set up here as our profit improves.

But right now, given where our profitability is at, we have certain minimum tax payments that we have to make in some jurisdictions around the world.

So that's why the rate is somewhat inflated.

As our profitability improves as we move forward, we're expecting the rate to come down.

Okay.

And then more broadly, you noted a sort of more positive tone around the demand environment.

Can you address a little bit more specifically what specific areas you're seeing signs of improved demand for you?

And just a follow-up, a quick one on Huawei as well.

You said there might've been about $100 million of lost sales here for the stoppage of shipments for a month.

But as you resume shipments, it sounded like you're not doing that at your full rate.

So at what rate are you expecting to be shipping in the September quarter relative to maybe the anticipated reduction that you're foreseeing here?

Okay.

On the overall basis, I think the way we look at end market growth, certainly we see strength in the data center across both HDD and flash as we get into the back half of the year.

We think there's growth possibilities both at the market level and specifically for us.

We also see continued resilience in normal cyclical demand flowing through the historical PC category.

So we see that trend continuing to be good both in terms of unit demand, but also in terms of capacity mix of that demand.

And generally speaking, mobile is, on a relative basis, stronger in the second half than it was in the first half.

So I think that's the general theme.

So we do see normal seasonality across markets that tend to behave that way.

And then obviously the data center is continuing to show robust capital spending expectations moving forward.

Yes, and then relative to Huawei, we have resumed shipment, as we noted in the prepared remarks.

But during the time when we see shipment to Huawei, they moved their sourcing requirements to some of our international competitors that were not impacted by the entity list.

And so when we subsequently resumed shipment, it was at a reduced rate that contemplated them sourcing components from other competitors.

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

If I'm doing the math correctly, so your full year industry outlook for capacity enterprise HDD exabyte shipment growth of 30%, that implies about 60% year-over-year, 60% petabyte consumption in the second half versus the first half.

That's very strong acceleration in the second half.

First of all, is this about right?

And this implies strong capacity enterprise shipments in the second half.

Secondly, is the cloud data center strong uptake pretty broad-based across most of the large U.S. cloud and hyperscale guys or limited to just a few?

Yes, so I think you're about right, and that's really reflective of our growth.

We do see general strength in the market, but I think we're growing at a faster rate than the rest of the market.

Yes, so the 30% that we were talking about was our growth, not necessarily industry.

Correct.

But your math on -- if you look at half to half comparison, your math is right.

60-40 is consistent with our math.

Got it.

Okay.

I appreciate that.

Good traction on the 14-terabyte.

On enterprise NVMe commencing revenue ramp this quarter, good to see the initial design win traction.

When will the higher margin profile of these products start to show up as a mix-related gross margin impact?

Is it December quarter you think?

Is it more first half of next calendar year?

Yes, I think there's some positive impact in the December quarter and it will sort of accelerate as we move forward.

Yes.

And the other thing is, is that -- putting aside what it'll due to margins, it also allows us to diversify and derisk our placement of bits going forward.

Our next question comes from the line of Karl Ackerman with Cowen.

Two, if I may.

First, Steve, Bob, or Mike, you made great progress on transitioning to 96-layer 3D NAND.

And you indicated in your prepared remarks that client SSD capacity rose 50% year-over-year.

I guess I'd like to ask how you think philosophically about your consumer hard disk drive businesses relative to your client SSD businesses now that NAND prices have effectively been cut in half on a year-over-year basis and NAND demand elasticity, as you called out, is making SSDs much more economical today.

Yes, so I think the way we think about that, and I think it's an essential to why we combine these assets, there will be reasonably long tail on client HDD assets, both 2.5 and 3.5, but what we've seen as we move back into this pricing regime is this sort of continued predicted penetration of SSDs into what was historically the HDD part of the PC market.

So that continues sort of as planned.

Obviously with the adjustment we made in our manufacturing footprint was in recognition of that.

So I think from our standpoint, things are generally progressing as planned and we're well situated to take advantage of that shift from one to the other, maintaining sort of the profit delivery of the HDD business as it moves into its later stages within the PC segment but grow within the SSD transition on the other side.

That's helpful.

I guess as my follow-up question, regarding your NAND industry bit supply assumption of low 20s, I assume that's a production number, but how should we think about your own bit supply number which I guess I would imagine would be mid-30s or higher?

And as a -- in relation to that, I'm curious how we think about idling wafer capacity and going after -- or exceeding market share in certain areas, mobile, as we described earlier, when some of the South Korean peers appear to be going after share near term?

I guess how should we think about the longer-term implications of that in regards to our cost structure?

Yes, so I think one of the things to think about is we focus very carefully on capacity -- relative capacity share.

I think we remain at a quite competitive position.

The short-term decisions we'll make based upon sort of economic and current market conditions we see is more tertiary, right?

So ultimately, we were comfortable with the decision we made several quarters ago to run underutilized.

That brought us into the middle of the year with a better overall inventory position than we would have otherwise had.

And ultimately, we're able to place our bits in a way that gives us a better overall economic outcome.

Yes.

And I would just -- to quibble a little bit with your choice of wording, I would not say that we ceded market share in the mobile market.

We determined that at the prices that were in effect, it did not make economic sense to take that business.

So it was an economic decision that went into that as opposed to the ceding of market share.

Our next question comes from the line of C.J. Muse with Evercore.

I guess a bit of a follow-up question.

I'm not sure I completely understood.

I believe on the NAND bit growth outlook for calendar '19 across the board, we've heard 30 plus percent from your 3 major competitors and you guys are talking low 20s.

So is the low 20s what you expect to ship this year that's production?

And as part of that, if in fact you are growing less than the industry, can you speak to how you're going to focus those bits?

And what are the implications we should be thinking about to your NAND gross margins?

Yes, so let me clarify both the 20% and the 30%.

If we look at the consumption of bits by our customers, and this is an industry statement, it would be in that 30% range.

While we're talking about the 20%, that is our estimate of industry growth and supply.

The difference between the 30% and -- the 30% and the 20% is consumption of industry inventory which we estimate will be largely back in balance as we exit this calendar year.

And you're talking industry inventory downstream?

Or on bit maker books?

Both.

Industry in the supply chain, whether they -- whether it's inventory on ours and our competitors' books or it's inventory that was built up either in the channel or at our customers.

So I guess as we model you guys through the second half of the year, we should be thinking about your bit shipments to revenue in the low 30s range or the low 20s range?

Yes.

So you should think about us in the low 30s range.

Okay.

Great.

And then I guess as my follow-up, as you think about just short term for the Q3 guide, I'm coming to roughly 60% of the incremental revenue growth coming from NAND.

Is that fairly accurate?

Yes, we're going to see growth in both hard drive and in NAND.

I don't know that we're going to get too specific on the markets, but we're going to -- it's going to show good growth in both areas.

Our next question comes from the line of Mehdi Hosseini with SIG.

I have a couple of follow-ups.

It's actually encouraging to hear the cycle bottoming, and I think demand for [casual] is going to be volatile.

But it seems to me that you have better control of the cost.

So if I were to hypothesize and assume that ASPs would just be flat, this is NAND ASPs flat into 2020, how should I think about your ability to expand NAND margin?

I'm not asking for a specific margin target, I'm just trying to understand, a, the mix of business that you're pursuing, especially given your preference on higher margin business; and b, how the 96-layer 3D NAND is going to impact your cost?

So any color on margin expansion, assuming that prices will be flat in '20, will be great.

And I have a follow-up.

Yes.

So Mehdi, let me -- this is Steve.

Let me address that in a broader sense.

We expect, as we move through fiscal 2020, that our gross margins for both hard drive and flash will improve, resulting in increasing earnings as we move through fiscal 2020.

We just have to wait and see how the supply and demand dynamics would impact pricing to have a bit of sense of gross margin, would that be fair?

Yes, in a literal sense.

But as said, what the specific dynamics are, we expect that our gross margins will continue to improve for both hard drives and flash as we move through fiscal 2020.

Also with the expense reductions and other things that we would see an increasing and improving EPS performance as we move through fiscal 2020.

Okay.

Fair.

And just a follow-up question for Bob.

We have had 2 consecutive quarter of negative free cash flow.

How does it evolve into the second half of the calendar year?

And given your conservative CapEx and improving margin profile, should we expect free cash flow growth to accelerate into the new calendar year?

Yes.

So good question, and we're very focused on cash generation right now, and I do expect we'll see improvement.

We're not going to give a forecast for free cash flow for next year, but we're definitely expecting to generate cash.

And as I mentioned, there are some timing benefits for us with respect to CapEx next year, so I think it'll be a pretty good cash generation year.

Would it be kind of breakeven in the worst case into the second half of the year?

I would say definitely a worst case.

Our next question comes from the line of Steven Fox with Cross Research.

Two questions for me as well.

First of all, just getting back to the mix issue.

Is it reasonable to assume, given some of the positive product development you talked about on the opening, that you start to see gross margin benefits from mix as you get maybe into the second or third fiscal quarter this year and where we might see that in HDDs and NANDs?

And then I have a quick follow-up.

Yes.

So let me talk about flash.

So I think it is reasonable to assume that on a normalized basis, where flash margin mix will help us, on a margin rate, obviously to the comment Steve made, we expect the overall margin rate on flash and HDD to continue to migrate in a positive direction through the year.

Same goes for HDD.

As the capacity enterprise becomes an increasingly large part of our output, we see an upward trend on mix as it affects margin.

That's helpful.

And then just following on that answer on the HDD side with the 14TB ramp now going pretty strong.

Do we think about average capacity per drive maybe reaccelerating, especially as you get into 16 and 18TBs?

Or do we look back at how it's been over the last 18 to 24 months as a guide for going forward?

No, I think we're going to see sort of -- our long-term sort of rate is 40% year-over-year exabyte growth.

We obviously have seen a down year this year, but we expect to return, over time, to roughly that 40% annualized growth.

And that's capacity enterprise, Steven.

And that's including, based on what you're seeing in terms of your own product development, say, beyond the year?

Yes.

And that is an industry number.

And obviously our performance, at least in the short term, we're outperforming.

Our next question comes from line of Vijay Rakesh with Mizuho.

I was wondering on your NAND side, what's the split of 64- and 96-layer and -- especially exiting the year?

Yes, so we talked about that in my prepared remarks.

Bit crossover happens as we exit this quarter in September.

That's the 96-layer.

So the total bit output crossover point happens in September.

Got it.

And just to clarify, I think you mentioned CapEx, the cash CapEx grows to $500 million for fiscal '20 from $1.5 billion in fiscal '19.

Is that correct?

And do you think that will impact any of your bit supply or transitions as you go into fiscal '20?

Yes, I don't think it's going to impact our production volumes at all.

As I mentioned, we are going to benefit from the timing of the capital deployment next year.

So we'll have the capital deployed and we're not expecting to pay for all of it within the year.

Our next question comes from the line of Munjal Shah with UBS.

Just 2 real quick.

First, on the data center's position business, if you look at it sequentially, revenues were up only 2% while capacity enterprise was really strong.

I was just wondering if there were any offsets in that business where revenue could be stronger.

Yes, so I think the offset is, to some modest degree, enterprise SSD which, as we discussed, will resume growth in the back half of the calendar year, first half of the fiscal year.

And then second question on the NAND side.

As your supply comes on from the Yokkaichi fab and it comes back fully, do you think you'll be equally -- you'll be able to place bits everywhere?

Or we would you still be...

We expect that we will have sufficient demand for the bits that we will be generating this quarter and beyond in the near term.

We expect to be fully sold out in that regard.

Our next question comes from Ananda Baruah with Loop Capital.

Two if I could, Steve and Mike.

I would just love to get your updated thoughts on how you're viewing the potential for this hyperscale cycle.

I think on the last call, Steve, you said you thought -- well, you guys made remarks that you thought it would go until at least the June quarter of fiscal '20, but that potentially could go through the calendar year.

So I'd love to get your thoughts there.

And then also thoughts on maybe potentially strengthen it as well.

And then I have a quick follow-up.

Yes, so I think what we've talked about was we see core strength through the end of this year and as best we can tell into the second half of our fiscal year, so first half of calendar year.

Sort of anything beyond that is a little outside of our normal visibility.

Okay, great.

And then any context on how -- I guess on degree of strength?

Yes, I think degree of strength, we sort of talked to in our guide.

So we think we're going to be above the 30% growth rate this year.

So it's been strengthening as the years progressed.

And I really think it's heading back towards that average of 40% year-on-year sort of growth.

Yes, and I don't know if it helps a little bit, but we have been talking about second half strength in terms of hyperscale demand, I mean frankly since the beginning of the year.

I would say that our level of conviction regarding that has remained pretty constant as we move through this calendar year and remains at high level.

So we've got pretty good confidence that previous comments that we've made in that regard remain in effect.

The last question comes from Mitch Steves with RBC Capital before we end with a short statement by our CEO.

Very good color on the NAND bottoming here.

But I just had a quick question just in terms of your cost control.

I think that one of the big questions I have is, if you have a demand environment that's similar to, let's say, about 1.5 years ago, do you think that you have the ability to reduce OpEx even further?

Meaning that the operating margin could actually exceed the prior levels when NAND pricing improves?

Well, I mean the team's done a lot of work in terms of getting the cost structure in place this year.

And we're going to -- as I mentioned, going to have the cost structure we want from an OpEx standpoint in the December quarter.

Probably would've had it in September had it not been for the 14th week.

I don't know that we're going to see a lot of opportunity to reduce from there and continue to execute the road map which we're pretty pleased with.

Yes.

And one of the things, I'll add a little bit of color on that.

I've been around a little bit longer than Bob, so I can probably add a little bit more.

But one of the things we learned having gone through the up-and-down cycle of the flash cycle is, is that -- because we're expecting that things are going to improve.

We talked about a cyclical trough in terms of the flash market.

What we want to do is we want to drive for efficiency as we are going through an up-cycle, and not add back expenses that were previously taken out through the down cycle.

That's not necessarily a promise that we will drive below the $740 million, but we're going to continue to drive for efficiency as the demand and pricing environment improves.

Got it.

I understand.

And then just one quick follow-up related to kind of the hard disk drive business.

So one of the things I understood is that the larger 16-terabyte drive takes a little bit longer to build essentially, maybe an extra 30 to 60 days.

So your comment about no change to kind of your data center expectations, do you guys have better visibility than you had when you were selling, let's say, 8-terabyte drives or smaller sized drive or is that incorrect?

Well, in general, we have increasingly good visibility because of the evolving sort of maturity of the cloud providers.

So just the sort of operational engagement with them gives us a better view into their plans.

And of course in some instances, we've even got commercial arrangements that give us a very good feel for their intentions.

We have one particular case where we actually are paid to carry some inventory for continuity of supply reasons.

And that gives us a very good feel for their deployment plans.

And so it's really that, that allows us to do it.

And as you say, as capacities grow, there is a lengthening lead time, so that's important for us to have.

But obviously over time, we want to counter that lead time lengthening with test time improvements that help shorten our cycle time.

All right.

Well, thank you all for joining us, and we look forward to continuing our dialogue.

Have a great rest of the day.

Thank you.

Ladies and gentlemen, this concludes today's conference.

Thank you for joining.

You may now disconnect.

Everyone, have a wonderful day.