威騰電子 (WDC) 2019 Q1 法說會逐字稿

Good afternoon, and thank you for standing by.

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

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

Now I will turn the call over to Mr. Peter Andrew, you may begin.

Thank you, and good afternoon, everyone.

This conference call will contain forward-looking statements within the meaning of the federal securities laws, including business plans and strategies, industry trends and business and financial outlook.

These forward-looking statements are based on management's current assumptions and expectations, and we assume no obligation to update them to reflect new information or events.

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

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

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

During the Q&A, we ask that you limit yourselves to 1 question.

Before I pass the line to Steve, I want to alert everyone that in an effort to provide investors greater transparency into our business and make our earnings related information easier to find, we've made a number of significant changes to the information in the earnings-related materials posted on our website.

All of these materials can be found in the Investor Relations Sections of our website under quarterly results and earnings documents.

Please take a few minutes to check out this new material, and we look forward to hearing your feedback.

With that, I will 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 Mark Long, Chief Financial Officer.

After my opening remarks, Mike will provide a summary of recent business highlights, and Mark will cover the fiscal first quarter financials and wrap up with our second quarter guidance.

We will then take your questions.

For the first quarter of fiscal 2019, we reported revenue of $5 billion and non-GAAP gross margin of 38%.

Non-GAAP operating expenses were at the year-ago level and with the help of a capital structure that has significantly reduced interest expense, we delivered $3.04 in non-GAAP earnings per share and over $700 million in operating cash flow during the quarter.

While our first quarter results were impacted by declines in flash average selling prices, we experienced strong performance in Capacity Enterprise, surveillance hard drives and embedded flash solutions with each growing revenue over 30% on a year-over-year basis.

Current geopolitical and industry dynamics are creating a more challenging global business environment.

For example, trade tensions with China, changes in monetary policy, foreign exchange volatility and the corresponding economic impacts are causing our customers to be more conservative, resulting in a demand slowdown for our products.

This softening demand, in combination with increased flash supply, has led to a market imbalance resulting in a deteriorating near-term flash pricing environment.

In response to these conditions, we are making an immediate reduction to wafer starts and delaying deployment of capital equipment.

These actions will reduce our wafer output beginning in fiscal Q3 2019.

The goal of these actions is to better align our output with the projected global demand for flash.

The duration of the planned output reduction will depend upon market conditions and will not impact our ability to meet customer commitments nor will it impede our ability to deliver the most innovative and cost competitive solutions to the market.

Longer term, we remain focused on technology leadership, the broadening of our product portfolio and continued operational improvements.

Highlights of our recent successes in these areas include: leadership in bringing the highest capacity and lowest total cost of ownership solutions to cloud customers with our recently introduced 15-terabyte SMR-based hard drive; successful introduction of 96-layer BiCS4 solutions, including our latest SATA-based client SSD and the industry's first UFS 2.1 embedded flash drive for high-end smartphones; ongoing growth in datacenter systems; and industry accolades for our ActiveScale object storage systems, Gartner has recognized ActiveScale as a challenger in the 2018 Gartner Magic Quadrant; and operational improvements that include the rightsizing of our hard drive manufacturing footprint.

We remain well positioned to capitalize on the long-term opportunities associated with the rapid growth in the volume and value of data.

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

We look forward to hosting our 2018 Investor Day on December 4 to provide further insight about our plans to deliver long-term shareholder value through ongoing growth and leadership in today's data-driven world.

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

Thank you, Steve, and good afternoon, everyone.

As Steve mentioned, we are taking aggressive actions to align our supply of flash to current end market demand.

I'd like to provide some additional insight into our actions announced today.

Our immediate reduction in wafer starts and delayed deployment of capital equipment is a reflection of the adjustment to bit supply growth needed to improve the supply/demand dynamic by mid-year 2019.

The magnitude of these actions is a reduction of 10% to 15% of our planned bit output in calendar year 2019.

With these adjustments to our supply, we expect our bit growth in calendar year 2019 to be in line with our view of end market demand growth of between 36% to 38%.

We will continue to monitor market conditions and make additional adjustments either up or down as the situation dictates.

As we enter the second fiscal quarter, we are faced with a number of market factors that will impact our near-term results.

Beyond the flash supply dynamic, other challenges include, we expect to be negatively impacted by the widely publicized CPU shortage.

Industry analysts have estimated the PC unit growth will be constrained between 5% to 10% for the current quarter, with the situation potentially lingering into early calendar year 2019.

We are experiencing a temporary slowdown in Datacenter capital spending, particularly by large cloud service providers after several quarters of growth above the expected long-term exabyte growth rate of 40% for Capacity Enterprise.

We are in the midst of adjusting to a more normal growth rate.

For the calendar year 2018, we now project Capacity Enterprise exabyte growth of approximately 55%.

Based on discussions with our customers, we expect stronger growth to resume in the second half of calendar year 2019.

Steve has already talked about how the global business environment is contributing to a risk-averse orientation with our end markets.

These dynamics are having a disproportionate impact on certain end markets outside the U.S., particularly China.

As a management team, we are well aware of the current macroeconomic and industry conditions and are taking appropriate actions to respond, while ensuring we invest for future growth and leadership across the markets we serve.

Examples of our progress in the first quarter were: we maintained our leading position across all the brands in our Client Solutions portfolio, including G-Tech, SanDisk and WD.

Client Devices experienced healthy demand in both surveillance hard drives and embedded mobile flash solutions versus a year ago.

Leading smartphone providers continued to increase the average capacity per device, which is driving ongoing growth for our mobile products.

Data center devices and solutions revenue grew year-over-year, driven primarily by continued demand for our Capacity Enterprise hard drives.

In addition, we experienced record revenue in our emerging Datacenter systems products, a strategic growth area we will discuss in more detail at our upcoming Investor Day in December.

I will allow turn the call over to Mark for the financial overview.

Thank you, Mike, and good afternoon, everyone.

As Peter mentioned earlier, we've updated our disclosures to provide investors with additional transparency into our business.

In particular, we are providing a breakdown of our total revenue and non-GAAP gross margins for flash products and hard drives.

These disclosures are included within the earnings presentation available on our IR website.

I will now review the financial performance for the September quarter.

Revenue for the September quarter was $5 billion, a decrease of 3% on a year-over-year basis and below our original outlook for the quarter.

The shortfall was primarily driven by weaker-than-expected flash pricing.

Specifically, our flash-based products had revenue bit growth of 28% sequentially, while the average selling price was down 16%.

The September quarter revenue for Data Center Devices and Solutions was $1.4 billion, an increase of 6% year-over-year.

Our Datacenter revenue growth continues to be driven by cloud-related storage.

Client Devices revenue was $2.7 billion, which was essentially flat year-over-year.

We had significant growth in embedded flash and surveillance products, offset by client-compute hard drives.

Client Solutions revenue was $932 million, a decrease of 18% year-over-year, driven by the normalization trends in flash market pricing.

Non-GAAP gross margin for the September quarter was 38%, down 430 basis points year-over-year.

Gross margin declined primarily due to a reduction in flash ASPs.

Non-GAAP OpEx for the September quarter totaled $820 million, consistent with the prior quarter.

Given the current environment, we continue to align our spending with our top priorities.

On a non-GAAP basis, net income for the September quarter was $906 million or $3.04 per share.

In the September quarter, we generated $705 million of operating cash flow.

We continued to reinvest in our business with $248 million in capital investments, resulting in free cash flow of $457 million.

In the September quarter, we had a sequential increase in inventory, primarily driven by a return to normal operating levels for Capacity Enterprise products.

Flash inventory grew slightly contrary to normal seasonal trends as a result of the market factors that Mike identified.

In the September quarter, we returned $711 million to shareholders, of which $148 million was in dividends and $563 million was through share repurchases.

These share repurchases put us on target of our goal to repurchase $1.5 billion of our common stock during fiscal year 2019, depending on market conditions.

We continue to believe this is an attractive capital allocation opportunity and demonstrates the confidence we have in our long-term outlook.

We also declared a dividend in the amount of $0.50 per share.

We closed the quarter with cash, cash equivalents and available-for-sale securities totaling approximately $4.8 billion.

Before I provide our guidance, I would like to describe the financial impact of our decision to reduce our flash output through reduced wafer starts and delay deployment of capital equipment.

We expect to reduce our planned flash output by approximately 10% to 15% for calendar year 2019.

The goal of these actions is to bring our supply more in line with the demand environment, with the majority of the supply reduction occurring by the middle of calendar year 2019.

Based on our current plan to temporarily reduce our flash output, we expect to take a GAAP-only charge in the range of $250 million to $300 million spread over the remaining quarters of fiscal year 2019.

This charge is driven by our fab capacity decisions and does not reflect an incremental cash payment.

As Mike indicated, we may adjust our plans based on market conditions and this could impact the extent of any charges.

I will now provide our guidance for the second fiscal quarter of 2019 on a non-GAAP basis.

We expect revenue in the range of $4.2 billion to $4.4 billion, gross margin in the range of 32% to 33%, operating expenses between $760 million and $780 million, interest and other expense of approximately $105 million, an effective tax rate of 10% to 12%, diluted shares of approximately 295 million.

As a result, we expect non-GAAP earnings per share of $1.45 to $1.65.

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

Operator?

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

I just want to try and dissect a little bit the guidance that you just laid out.

Help us understand the variables to consider between the hard disk drive revenue assumptions and that of the flash assumptions.

And then also, on the gross margin line, it looked like just simple math is that you're able to take your cost down by a high-teens percentage rate on a year-over-year basis this last quarter.

How are you thinking about the cost down equation?

And what's the assumption you're making for flash gross margin in this current quarter?

Well, why don't I let Mike talk about the product side and the cost side first and then I'll finish with the modeling question.

Yes.

So on the cost down, just as we think about cost, we've talked about the long-term range of between 15 and 25.

We've been operating and expect to be operating near the low end of that.

We continue to see that as our expectation and, obviously, we'll update this further with more perspective at the Analyst Day in December.

Relative to the product impacts, it's roughly impacted equally on both sides between the HDD and the flash business relative to the impact in question.

Yes.

So from a modeling standpoint, when you think about the HDD side, the Capacity Enterprise and client compute contributes to the decrease in revenue.

And then on the flash side, we have the pricing environment reducing revenue as well as mobility declines.

The margin changes are largely a function of the flash pricing environment and then the lower mix of Capacity Enterprise on the hard drive side.

Our next question comes from Wamsi Mohan of Bank of America.

Steve, I appreciate the disclosure on the flash side.

That was very helpful.

Clearly, you're taking some actions here to curtail flash output to be in line with demand.

But can you give us some sense of how long do you think that it might take to get the supply-demand back into balance?

Sounds like in the next year, possibly, in the way that you're currently thinking about it.

And why do you expect this action will be followed by other competitors versus a strategy of, potentially, share gains?

That is, they keep pricing down and try to gain bit growth as opposed to sort of reducing output similar to what you're doing.

Yes.

So let me add color on that.

Our -- the efforts that we're taking are intended to get our supply/demand equation in balance by the middle of calendar 2019.

We don't know what the rest of the industry's going to do.

And so your comment, in terms of why do we think that the industry's going to follow that, that's not a correct characterization.

That's going to depend upon the behavior of our competitors.

And so really what we're trying to do is, if you want to call it that, get our own house in order so that we are better positioned to take advantage of, let's call it, the long-term growth opportunities in the market.

And in particular, to kind of put a finer point on it, we would expect that demand, from a seasonal perspective, would pick up in the second half of calendar 2019 and getting our inventory situation in better balance will position us better than we otherwise would have been if we had not taken these actions.

Yes.

And just to add a little color on that.

Our cuts are -- bring us in line with what we expect the end-market demand to be.

So straight calculation there is it's obviously, intending to maintain share.

Yes.

And the other thing that I would add is that there is no intent or implied shipped-bit share loss in our numbers.

In other words, as I indicated in my prepared remarks, we still intend to have enough bits or supply to demand our customer commitments.

This is really a matter of supply has been outstripping demand for a variety of different reasons.

We want to get that excess out of the system and better position us going forward.

If I could, really quickly, when you think about the gross margin trajectory here, obviously, the flash margins have been compressing here a few quarters and going to continue.

But as you think about these capacity reductions coming ahead, should we expect gross margins to compress beyond the December quarter here, like at least for the next couple of quarters at this point?

Well, it's great question.

And the reality is, is that, I'll be sort of literal, we don't know for sure.

We're going to have to wait and see, but we would expect that we would continue to face a challenging flash market through the first half of calendar 2019.

And so that implies that there could be further margin compression but at what rate, we will have to see.

But we don't -- these reductions that we're making from a wafer start perspective will not begin to hit, just because of lead time for us, until the first quarter of 2019, it will primarily allow us to reduce our inventory and then, of course, we've got the demand environment that we're dealing with and whatever competitive dynamics we'll be dealing with.

But we do expect the first half of 2019 to continue to be challenging for us and for the industry.

(Operator Instructions) Our next question comes from C.J. Muse of Evercore ISI.

I guess, in light of the wafer start change and delay in CapEx, I would love to hear your thoughts on how you think about scale and other factors to enable best-of-breed cost-down in this highly competitive market.

So are there other avenues to lift that sort of low end of the range of 15% that you're focused on, perhaps pulling in 96-layer or above or a faster move to QLC?

Would love to hear your thoughts, particularly in light of 6 players plus YMTC entering the fold.

Yes.

Relative to the cost declines, you're on the point.

It's tech transition, so rate in which we convert to the next node and the implementation of QLC over the longer horizon, that's the bigger lever.

The thing we are highlighting in our projections as of now is just the realities of the 3D era, right?

They're more capital intensive and the cost declines are going to certainly be less than they were in the 2D era.

So that's all part of how the market and the industry is having to plan differently.

Yes.

And we continue -- because of the joint venture structure and that we continue to expect to benefit from scale despite the reduction in wafer starts because it's only our capacity that we're reducing in those wafer starts.

But if you look at the total output of the joint venture, we continue to have the broader benefit of a scale semiconductor manufacturing operations.

Our next question comes from Amit Daryanani of RBC Capital Markets.

This is Amitesh Bajad for Amit Daryanani.

Just on the NAND ASP decline, do you, at this point, see any risk of you having to cut pricing on the HDD side as well?

Or actually, at what point do you think that NAND ASP declines could pressure you to cut HDD pricing to slow down the substitution of SSDs versus HDD?

Yes.

I think, in the case of relative pricing, that's not the direct impact.

It really is the rate of cannibalization and we have seen that accelerate.

We would expect we exit the year about 60% in terms of SSD versus HDD in the PC space.

Yes.

And just to comment on that because the -- as painful as it is, the announced closure of our Kuala Lumpur manufacturing facility anticipated continued decline, particularly in terms of client hard drives.

And so that expected uptick in terms of substitution of SSDs for hard drives has been factored into our plans, and we'll continue to look at opportunities at how do we optimize our manufacturing footprint in light of that.

But the KL activity, that closure, not to take it lightly, but was done in direct response to that anticipated uptick.

And if I could just have 1 follow-up.

The cost downs of 15% to 25% range, do you think that, that could have like some kind of a downside bias next year as you ramp up 96 layer?

So if your question is a downside bias to the lower end of that range, I think, we'll give you more commentary in December what we think next year will look like.

Our next question comes from Mehdi Hosseini of Susquehanna.

This is for the entire team.

Inventories have been going up since earlier this year.

And I'm just curious how you're thinking about the debt covenants, especially given your guide and your comment about the margin erosion.

Is there any of the covenants that would be at the risk of default?

And is there anything other than the cost down, manufacturing cost down, that you could do to improve cash flows and minimize the downside risk?

So in terms of our covenants, through our refinancings and the optimization of our balance sheet earlier in this calendar year, we put ourselves in a good position to manage through volatility, and we remain in good standing with respect to our covenants.

And our current forecast keeps us in a good position with respect to compliance.

As it relates to managing cash flow, we are looking at all the levers, as Steve pointed out, it's part of our overall review.

We look at our manufacturing footprint, we look at how we manage working capital, we look at our OpEx levels.

And we are constantly pushing to both manage our business as efficiently as we can with respect to the current market environment and then make sure that we are also investing in the right ways for what we see as some very good long-term opportunities and some very good growth potential on the Capacity Enterprise hard drive side and in the flash business.

Yes.

And Mehdi, just to add color to that because I think it's important to emphasize, when we looked at the decision to cut capacity, not cut capacity, I mean, ultimately what we were solving for was what did we think was a better answer from a cash perspective.

And that action, in and of itself, just as a for instance, provides us with a meaningful net cash positive.

And so that is something that we are constantly looking at mechanisms to optimize not only on a short-term basis but on a long-term basis.

But what is it that inventory, and it ties into my question, the days of inventory are up 30% year-over-year and up 2% Q-on-Q, and this is kind of the repeat pattern that happens.

And I'm just -- understand as you focus on cost down, is there something with the mix or is there something with your inventory hubs that still needs to be looked at?

Well, let me give some color on the inventory situation.

The inventory is up, it's up more than we want it to be, okay?

Let's put that out there first.

But it's up for 3 principal reasons.

One is, is that we have more flash inventory than we'd like, right?

That's why we're taking the wafer start cuts so that we get that inventory back down to a more acceptable level, better matched in terms of supply and demand.

So that will be corrected barring there being other changes in demand and that kind of thing.

That will be corrected over the course of the next -- well, through the middle of calendar 2019 because of those actions.

The other thing is, is that we have built up a bit of more inventory, not excess inventory but more inventory.

As we look to close our Kuala Lumpur facility and transition that manufacturing, we need some buffer inventory to enable that transition because you can't unplug it from one factory and plug it into another factor and not lose any manufacturing.

You have manufacturing downtime as a consequence of that.

That's the second reason.

The third reason is that we have largely been running, over the course of the previous 9 months, let's just say, at very high demand levels in terms of Capacity Enterprise.

And so we've been a little bit, let's call it, hand-to-mouth without much buffer inventory in terms of capacity inventory.

As that market -- well, not only begins to soften but as it softens, we'll see a returning to a more normal level of capacity inventory for capacity hard drives because the demand predictability on the part of our partners or our customers is not necessarily that great.

They don't want to go down in terms of having drives, and so we tend to hold a little bit more extra inventory for them in order to maintain high levels of service.

So that's, in brief, the explanation of what's going on with our inventory balance.

Our next question comes from Joe Moore of Morgan Stanley.

Sorry if I missed this, but could you give us a little bit more color on the noncash -- or the non-GAAP charge -- or the GAAP charge, I should say, that you're taking to take the utilization down?

I guess, my understanding of the way the JV tab works is that when it's underloaded that you still pay the cash -- the higher cash cost for those wafers reflecting the underloading.

So can you just help me understand what that charge actually is?

So the technical structure of the JV is not as you describe, it's actually that we make cash payments to our -- the JV, but they're not -- it's not for the purpose of purchasing wafers.

As a result, it's what I referred to as the -- we have no incremental cash payments, but we do continue to make cash payments in connection with our JV agreement for -- to basically pay for the CapEx, depreciation, et cetera.

The CapEx that we have not funded directly.

Exactly.

They own the buildings, they do that.

They have the labor, we have to pay for the labor, things like that.

Right.

So these are -- this is a true underutilization charge that we are taking, not a -- it doesn't have anything to do with the purchasing of wafers.

Our next question comes from Munjal Shah of UBS.

Yes.

I had a question on the capacity drives.

How broad based is the softness from your end customers?

You mentioned outside of U.S. macro impact, but is it more broad based on the Capacity Enterprise side?

Yes.

So on the Capacity Enterprise side, it is quite broad base.

So we're seeing it sort of across all.

Our comments relative to China were not specific to the cloud providers, it's really more broadly our assessment of the market in China.

So just to separate it distinctly, broad impact on cloud service providers on a worldwide basis and then the China comment was more general to broad market slowness in China.

Yes.

Just to add to that.

Really what we're seeing, and I don't want this to sound alarmist, but we've seen other people talk about it in terms of reports and that.

The simplest way of characterizing it in terms of what's happening from a broader demand perspective is there's a bit of a risk-off kind of mentality in terms of our customers' end markets, all that.

So everybody is kind of leaning out, in effect, their supply chains.

So if you look at the hyperscale guys, they're going to push up their utilization rates to higher levels than maybe what they were operating at.

If you look at, say, it's a PC cut, they're going to thin out their inventories, our channel partners are going to [chin up].

So there's a -- but it's generally under this guise of kind of just taking a little bit of air out of the balloon from a risk standpoint.

We are doing the same, we are absolutely doing the same and taking actions from a cost, from a capacity perspective, from an expense perspective that is reflective of that environment.

That's great help.

And just related, do you think that the cloud customers might have like double order?

I know you mentioned that you keep extra inventory to kind of provide service, but do you think that they might have given higher forecasts or ordered more and that's why you have inventory?

No.

I don't think we have any specific indications of that.

But I think, as Steve said, they are taking a more conservative posture.

I think in some indications, they are continuing to consume some inventory, but that's not the primary driver.

(Operator Instructions) Our next question comes from Jim Suva of Citigroup.

Given what the share price has done recently plus your outlook to earnings, any changes to what you think about capital deployment for your stock buyback or the early cadence of it?

Or do you need to keep that cash a little more handy for your closing of Kuala Lumpur and you're shifting in other businesses?

Just wondering if there's any change in the tempo.

I know what you've done historically, that's for sure, but just kind of thinking -- getting a pulse and the sense currently.

Sure.

There's no change to our capital allocation strategy.

As we have said, we have a balanced strategy, we allocate the capital among our top priorities.

So investing in the business, our CapEx and then we look at, of course, the dividend and our share repurchases.

We had targeted $1,5 billion for repurchases in fiscal '19.

And as we said, we remain on track for that.

So there's really no change other than we do agree that, with the current share price, it is an attractive use of capital, and we will factor that into our balanced strategy going forward.

And then as a quick follow-up.

You mentioned the word dividend, that was my follow-up is how should we think about dividend going forward?

Do you use it as a percent of cash flow, as a rate of change in earnings?

Could the dividend be reduced since the yield is going higher?

How should we think about the dividend?

Well, dividend is a critical part.

We're not that formulaic, I would say, #1.

But dividend is a key part of our capital allocation program, if you want to call it that, and we continue to expect it to be that way going forward.

And we'll look at different factors and see if we want to modulate it.

But clearly, in the a short term, we're not looking at anything that would -- we're not looking at cutting the dividend, I can assure you of that.

Our next question comes from Karl Ackerman of Cowen & Company.

I had 2 questions, if I may.

First question is just really perhaps trying to level set some things on what is your comfort level about hyperscale end demand beyond the December quarter?

I think one of your primary memory peers spoke about some challenges in server demand -- server memory demand for the first half of '19.

I'm just kind of curious how should we think about the impact on your nearline hard drive business?

Yes.

Let me answer that.

So yes, as I said in my prepared remarks, we expect more robust growth to return to that sector in the second half of calendar 2019.

So that's the way we're thinking about that.

So the first half will continue to be certainly weak compared to what we saw last year.

But that was -- the growth was above the longer-term exabyte growth rate of 40%.

But we expect this, if you want to call it, stall, from a buying perspective, to persist through the first half and then resume to more normal patterns in the back half of 2019.

Understood.

I guess, with that demand backdrop, you obviously run a fairly lean operation today, but how should we think about the trajectory of OpEx as the macro drives some lower revenue near-term but you're also going to rationalize your hard drive facility?

Any commentary there would be helpful.

Yes.

So as, I think, both Steve and Mike mentioned, we are very focused on optimizing our cost and expense structure given the market environment.

As we indicated, our OpEx will be trending down in the current quarter.

And we will look for opportunities to continue to improve that and reduce that going forward, so.

While also making sure that we're not cutting our nose off in spite of our face in terms of enabling our technology and product competitiveness over a longer threshold.

Our next question comes from Ananda Baruah of Loop Capital.

Just sticking right there, Steve, if we could, with nearline.

Should we think -- I mean, what do you think is the wisest for us to do as you think about first half of '19 exabyte?

Should we think of it as flattish?

And I know the term growth, even though muted, has been used, so do you actually think we can get back to a little growth?

Or should we think of it as just more flattish?

Or actually just how should we think of it?

And then...

Yes.

I think the best way to think about it, year-over-year is very flattish.

Got it.

For first half.

And then the thinking you'll get back to...

In exabyte terms.

On exabyte terms.

Correct, on exabyte terms.

Got it.

So then, as, Steve, you had mentioned sort of the risk off broadly, is there any way to get a sense of -- or do you have a view on, for nearline, how much of sort of the exabyte softening is risk off versus just sort of cycle winding down?

I'm not sure I'm smart enough to dimension that.

I'll turn it -- see if Mike's got a sense for that.

Yes.

I think, it's a combination of the two.

I think, obviously, there's been a rapid buildout that has been occurring.

Different providers have had different drivers of their buildout.

So in some instances, they are onetime things that are sort of behind them.

In other instances, as you know, in addition to the risk off, once they do a large buildout, they'll move into an optimization mode, so there's some of that going on as well.

So it's a combination of those things that are really creating the demand environment we're articulating here.

Our next question comes from Kevin Cassidy of Stifel.

This is John Donnelly on for Kevin.

In terms of the CPU shortage, can you describe how you see your customers positioned?

Are they waiting to buy the storage components?

Or is there a potential for some excess inventory to build up there?

Our view of it would be this.

First of all, depending on which part of our market, it's not, let's call it, proportionally impacting the market.

So it's less in the first tier and more elsewhere.

And there's -- we actually see it as a constraint to shipments in the period.

So that's how we are evaluating it.

Our next question comes from Vijay Rakesh of Mizuho.

Just looking at the 2019 guide on NAND for 36% to 38% bit growth, just wondering how you split it between 64 and 96 layer?

We haven't commented on that.

We'll give you more insight on that in December.

I will add that we are making good progress.

I mean, very pleased with our ramp of 64 and 96-layer technology.

So we feel that we are very competitively positioned in that regard.

Got it.

And on the hard disk drive side, you talked about [near] landing a little flattish year-on-year first half.

But do you see -- what kind of cost difference could you get -- or density increases could you get next year in terms of when you compare to the cost declines on the NAND side?

Yes.

So relative to the datacenter side of the business, Capacity Enterprise, in particular, we think that cost decline runs roughly at that same rate, so in that sort of 15% to 25%.

So it tracks nicely.

And our last question comes from Nehal Chokshi of Maxim Group.

I wanted to get a sense as to what you felt the new entry, Yangtze Memory Technologies' cost competitiveness is relative to their joint venture with Toshiba?

And whether or not they are absorbing some of the share, given the trade tensions that you've talked about?

They're having minimal impact at this point, minimal to no impact.

And this does conclude our question-and-answer session.

I would now like to turn the call back over to Steve Milligan for any closing remarks.

All right.

Thank you for joining us and we look forward to seeing many of you on December 4 at our Investor Day.

Have a good rest of the day.

Thank you.

This concludes today's conference call.

Thank you for joining.

You may now disconnect.