威騰電子 (WDC) 2019 Q3 法說會逐字稿

Good afternoon and thank you for standing by.

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

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

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

You may begin.

Thank you and good afternoon.

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

We assume the 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 references 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 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.

Also with us today is Bob Eulau, newly appointed Executive Vice President and Chief Financial Officer, who will formally take over the CFO role from Mark Long on May 9, 2019.

Bob has more than 30 years' experience in various financial and operational leadership roles in the technology industry, and I'm pleased to welcome him to the Western Digital team.

Thanks, Steve.

I'm excited to be part of the Western Digital team, and I look forward to working with all of our investors and analysts in the future.

Turning to our business.

Market conditions have generally been consistent with our expectations described in January.

That being said, we are encouraged to see demand incrementally improve in certain areas such as capacity enterprise and client compute.

Our expectation for the demand environment to further improve for both flash and hard drive products for the balance of calendar 2019 is largely unchanged.

And looking at the quarter from a product perspective, sales of hard drives were a bit stronger than expected.

Higher demand for capacity enterprise products drove most of the upside, and we saw a nice rebound at higher capacity points.

We are experiencing a very smooth product ramp with our 14-terabyte product.

In addition, we also realized a significant expansion of our presence in the midrange.

Demand for flash-based products was slightly better than expectations.

However, prices declined more than we anticipated.

We are executing well on the plans we laid out last quarter in terms of enhancing our product lineup, driving technology advancements, rightsizing our factory production levels and lowering our cost and expense structures.

We continue to make excellent progress toward commercializing our internally developed NVMe-based platforms.

I'm pleased to report the commencement of initial revenue shipments of our enterprise NVMe SSD, and we are on track to accelerate the volume ramp of this product over the remainder of calendar 2019.

We also commenced shipments of our NVMe client SSDs based on 96-layer 3D flash BiCS4 technology.

The manufacturing ramp and commercialization of BiCS4, which we believe is the industry's lowest-cost technology, is progressing well.

In the second half of calendar 2019, BiCS4 will become our highest volume runner in terms of flash output.

In HDDs, we continue to lead the industry in areal density through technologies such as SMR and energy-assisted recording and are on track to ship our first energy-assisted capacity enterprise drives later this year.

From a flash supply perspective, we are continuing our previously announced wafer output reductions without compromising our cost leadership position.

We made substantial progress on realigning our cost and expense structure, the Kuala Lumpur manufacturing facility has largely ceased operations, and combined with other actions, we are already realizing incremental cost savings.

Our focus is on driving long-term value creation for Western Digital and its stakeholders while prudently navigating near-term business conditions.

Emerging technologies such as 5G open up new applications and use cases, creating immense amounts of data.

Further, applications such as artificial intelligence, machine learning, autonomous vehicles, mobility and IoT will continue to generate growing data volumes that need to be captured, preserved, accessed and transformed.

The relentless pace of innovation bodes well for the long-term growth in data storage requirements and the resulting opportunities for our company.

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

I would also like to thank Mark Long for his service to Western Digital.

Mark has been instrumental in the transformation of the company, and I wish him the very best in his future endeavors.

With that, Mike will now share our business highlights.

Thank you, Steve, and good afternoon.

Starting with our highlights for the March quarter.

Within Data Center Devices and Solutions, our capacity enterprise drive category performed better than expected with demand strengthening as we moved through the quarter.

For the first half of calendar 2019, we now expect total exabyte shipments in capacity enterprise be flat to slightly up on a year-over-year basis.

Additionally, if current demand trends continue, we see an opportunity for the category to reach 30% year-over-year growth in exabytes for calendar year 2019.

This updated industry forecast is a significant upward revision from our prior estimate that was in the low 20% range.

Our 14-terabyte capacity enterprise drive qualification and adoption have been seamless, and we are leading the industry in this transition.

We are in the midst of a significant ramp of this product in the current quarter.

We are on track to introduce our first energy-assisted 16-terabyte CMR and 18-terabyte SMR hard drives later this calendar year.

Compared to competitive solutions, our new products will be cost optimized offerings, containing fewer disks and heads, showcasing our significant areal density advantage.

Our refreshed product line in the midrange capacities also did very well, and in total, we have meaningfully increased our presence in the capacity enterprise category.

Just a few weeks ago, we were delighted to hear that the Western Digital storage solutions played a key role in unveiling of the first-ever image of a supermassive black hole.

The Event Horizon Telescope project utilized our high-capacity helium drives which perform reliably in harsh, high-altitude and the remote environments.

This unique event highlights the fundamental role we play in the capture and transformation of data.

In enterprise SSDs, NVMe product qualifications at hyperscale customers are progressing to plan.

This product, built on our internally developed controller and firmware, complements our other enterprise SSD solutions that are already shipping.

Consistent with the platform approach we discussed last quarter, we expect to expand our enterprise SSD product portfolio throughout 2019 in a cost-effective and predictable manner, including a version that incorporates BiCS4.

In Client Devices, demand for hard drives for the PC market was slightly better than our expectations.

In client SSDs, our exabyte shipments more than doubled from the year ago quarter driven by strong price elasticity.

Additionally, we have begun shipping our mainstream client SSD products based on BiCS4 technology.

In Client Solutions, we are very pleased with the success of our external SSDs sold through retail, and we have continued to expand our presence in this category over the last several quarters.

In the March quarter, we also launched the world's fastest 1-terabyte microSD card establishing an industry first in the removable products category.

Average capacity per unit for flash devices grew 44% year-over-year, reflecting significant price elasticity.

From a flash supply perspective, we remain on track to achieve an overall reduction of 10% to 15% of our bit output in calendar year 2019.

Despite seasonal softness in the March quarter, our flash inventory level was essentially unchanged from the prior quarter.

As we look to our fourth fiscal quarter, we expect to see a full quarter impact from our reduction in wafer starts which will help to further reduce our flash inventory levels.

This combination of stabilized inventory and lower supply growth in flash will allow us to be more selective in how we pursue flash revenue opportunities going forward.

In terms of longer-term planning for flash, the construction of the building show in Iwate, Japan is on track with meaningful output from this fab expected in fiscal 2021.

Based on the schedule and the planned slowdown in our capital deployments, we expect lower flash-related capital spending for fiscal 2020.

The ramp of our 96-layer-based products continues as expected.

We are pleased to be shipping the industry's leading technology into retail products and client SSDs.

In the June quarter, we estimate BiCS4 to represent more than 25% of our total shippable flash bits.

We are on track to implement BiCS4 across our portfolio, including mobile embedded and enterprise later this calendar year.

As Steve mentioned earlier, we have largely stopped production activities at the KL facility.

We have also consolidated our head manufacturing operations from Thailand to the Philippines.

These continuing actions are allowing us to lower our cost structure and rightsize our manufacturing footprint to meet long-term demand.

Turning to our outlook.

We see a few additional data points indicating incremental improvement in current demand conditions compared to a very tough fourth calendar quarter.

Flash pricing conditions remain challenging, but we anticipate the rate of price decline will moderate as the year progresses due to a slowing rate of industry supply growth, elasticity driving demand for higher capacity points and seasonal strength in the back half of the calendar year.

Based on recent industry announcements, we estimate flash indices supply growth to be slightly more than 30% in calendar 2019, somewhat lower from our prior forecast.

To summarize, our current product portfolio is the best in our history.

In HDDs, the migration to higher capacity drives plays to our strength in technology and manufacturing.

In flash, our product portfolio in 2019 has been significantly enhanced with the expansion of our NVMe product line for both client and enterprise SSDs, and we continue to have brand leadership in retail.

We are taking appropriate steps in a challenging market environment to position ourselves for ongoing success.

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

Thank you, Mike, and good afternoon, everyone.

Revenue in the March quarter was $3.7 billion, in the middle of our guidance range.

Flash revenue was $1.6 billion with a sequential bit decline of 5% and a sequential average selling price per gigabyte decline of 23%.

The sequential decline in flash revenue was primarily due to price, seasonality and weaker sales of embedded mobile products.

Hard drive revenue was $2.1 billion which was similar to the prior quarter.

Non-GAAP gross margin in the quarter was 25.3%, below our guidance of approximately 28% due to a $110 million charge or a 300 basis point impact incurred for lower of cost or market LCM reserves.

Flash non-GAAP gross margin was 21% due to the rate of price reductions and the aforementioned $110 million LCM charge primarily related to an inventory write-down of multichip packages that contain DRAM.

Hard drive gross margin on a non-GAAP basis rose to 29% compared to the prior quarter driven by an increased mix of capacity enterprise drives.

Excluded from the non-GAAP cost of revenue is a $148 million charge related to the underutilization of our portion of the flash joint venture fabs.

Non-GAAP operating expenses were $742 million, below our guidance as a result of better progress towards our expense reduction target.

As a reminder, for non-GAAP cost of goods sold, we expect the full benefit of our $100 million per quarter cost reduction efforts to be reflected by the end of the December quarter of 2019.

And for non-GAAP operating expenses, we expect to see the full results of our $100 million per quarter expense reduction efforts to be reflected within the September quarter of 2019.

Our non-GAAP tax expense was $49 million which was higher than estimated due to a quarterly effective tax rate true-up and the fact that our tax expense is a relatively fixed dollar amount at this profitability level.

Non-GAAP EPS was $0.17.

The LCM charge impacted diluted non-GAAP EPS by approximately $0.37.

Operating cash flow for the March quarter was $204 million, and free cash flow was a negative $110 million primarily due to lower operating income.

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

At quarter end, we had $3.8 billion in cash, cash equivalents and available-for-sale securities, and our principal debt outstanding was $10.8 billion.

Earlier today, we announced the successful execution of an amendment to the existing financial covenants under our credit agreements.

The amendment provides Western Digital with significant additional financial flexibility to navigate market cycles.

Turning to inventory.

On a dollar basis, hard drive inventory decreased, and flash inventory was essentially unchanged from the prior quarter after the impact of the LCM reserves.

Looking into the June quarter, we expect both flash and HDD inventory to decline on a sequential basis.

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

We expect revenue in the range of $3.6 billion to $3.8 billion; gross margin of approximately 24% to 25%; operating expenses between $720 million and $740 million; interest and other expense of approximately $100 million; tax expense between $20 million and $30 million; diluted shares of approximately 295 million.

As a result, we expect the non-GAAP earnings per share of $0.10 to $0.30.

Finally, for modeling purposes, please note that the September 2019 quarter will have 14 weeks instead of the normal 13 weeks.

In closing, before I turn the call to the operator for the Q&A, I'd like to thank Steve, Mike and the entire Western Digital team for the opportunity to serve as the company's CFO.

It's been a great experience working with all of you.

With that, operator, please begin the Q&A session.

(Operator Instructions) And our first question comes from Wamsi Mohan with Bank of America.

Steve or Mark, this amendment to the credit facility is clearly prudent.

But would you say this is just preemptive versus your expectations of EBITDA and free cash flow generation over the next few quarters?

Has that changed in a material fashion?

Yes.

We took the opportunity to amend our credit agreements because the markets were favorable, and this does give us a significant amount of additional flexibility, and we were able to do it with very low cost.

And do you expect -- sort of your expectation of just sort of the EBITDA generation as you look over the next few quarters, how would you characterize that?

This was not in response to any near-term concerns.

This was just a prudent move that gives us good long-term flexibility.

I appreciate the color there.

And if I could really quick.

Steve, you mentioned demand trends were largely unchanged.

Can you just talk about sort of what has changed?

And that capacity enterprise uptick, where do you think that is coming from, from a customer base standpoint and maybe even regionally, if there is some color there?

Sure.

I'll have Mike comment and provide a little bit more detail on capacity enterprise.

But as I indicated, capacity enterprise was a bit stronger than we expected which we were very pleased to see that.

Client compute, as you all know, PC volumes continued to decline, but they declined at a more moderate rate than what at least we were expecting.

One area of -- a bit of weakness from a demand perspective, which I think is largely known in the investment community, is handset volumes were a bit lower than expected, so we did see a little bit of drawdown as a result of that.

But Mike can comment a little bit more on the capacity enterprise upside that we saw.

Yes, Wamsi, I think the capacity enterprise, it was fairly broad-based.

At a highest capacity, I think it would be viewed as sort of domestic hyperscale providers.

We saw strength of demand both in our 12- and 14-terabyte products.

And then in the mid-range, we saw strength in Asia.

So we saw fairly broad-based strength across the capacity enterprise category.

And our next question comes from Aaron Rakers with Wells Fargo.

Two, if I can, real quick.

So first of all, just kind of thinking about how you rolled up to the guidance number that you've given, particularly around gross margin.

Can you help us understand, assuming that the Kuala Lumpur facility closure kind of keeps a positive trend on gross margin and also coupled with the commentary on the demand front for the high-capacity drives, if we were to assume hard disk drive gross margin kind of stays flattish here, it would appear like your flash gross margin looks to be maybe in that 17%, 18%, 19% range.

So first of all, am I kind of thinking about your guidance in the context of flash gross margin correctly?

And underneath of that, what assumption are you making in terms of pricing and kind of cost-down discussion as we go into the June quarter?

So I'll give some overall comment, Aaron, and then either Mike or Mark can add a little bit more specifics.

One of the things that I will highlight, if you go back particularly to fiscal Q2, our gross margin levels for the hard drive business had gotten kind of below what we normally expected.

27%, I believe, is the number we saw, a nice rebound this quarter as we begin to see some of those cost improvements that you noted as well as a better mix in terms of capacity enterprise.

We would expect that, that would continue, and we'll work our way back up into a -- let's call it, a more normalized gross margin level in the hard drive space which clearly implies that given that we're guiding essentially to flat gross margins, that we'll continue to see a downdraft in terms of flash gross margins in the current quarter as we see cost -- price pressures continue.

As Mike indicated in his commentary, we expect, although, of course, we don't know for sure that, that price decline will moderate as we move through the calendar year.

So it'll take a little bit of time for that to get into a more acceptable range of price declines.

So I don't know if that helps a bit, Aaron, in terms of high-level commentary.

Yes.

The other thing I'll add, Aaron, is on the flash cost declines, we've kind of talked about being at the low end of the long-term range at 15%-ish, so you got to think about it that way on an annualized basis.

Okay.

And then just a real quick follow-up, if I can.

If I think about how you kind of consider a longer-term gross margin, what -- as we kind of get back to a more normalized trend and the product portfolio kicks in, what do you guys think is a good way to think about what you would consider as a normalized gross margin in the flash business?

And I'll end it at that.

So the margin range, I think is really...

Yes.

I mean I think on an overall basis, we are not changing our long-term model at this point.

So we just need to see the industry work off the inventory and the price declines to moderate as we've talked about.

And then as flash normalizes, and Steve talked about the improvements in the hard drive gross margin through the back half, we should see the kinds of dynamics we talked about at the last Investor Day.

Yes.

And to put a finer point on it, Aaron, if you just look at our model, that's high 30s, low 40s for flash.

So you think you could get there in the back half of the calendar year?

We're not saying that, Aaron.

No.

We -- we're talking about the long-term model.

You said in the long term.

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

I just want to go back to margins, if I could for a second.

So your margins in NAND would suggest that you're losing money on an operating level, but how much of your margin decline was due to the near-term cost ramp -- of ramping BiCS4?

And your NVMe SSD, that may reverse as volumes ramp for the balance of 2019.

And secondarily to that, I think you initially announced total underutilization cost would be $250 million to $300 million.

We've got probably another $75 million to go.

And so may you remind us how much of the $400 million annualized benefit in COGS from your restructuring actions will be realized in June?

And then taking those two as a whole, how do we think about the margin trajectory, at least qualitatively, for the second half in NAND?

I'll talk to sort of flash cost-downs, and given the ramp of BiCS4, I just talked about the annualized rate, that is more back half loaded.

So we see a shallower cost decline in the first half, more to come in the back half.

And Mark, do you want to...

Well, I think you were exactly right in terms of our underutilization charge, with the majority of it occurring in the prior quarter and through this quarter.

So we haven't given a breakdown in terms of the amount of our Kuala Lumpur shutdown benefit that we're receiving in the Q4.

But as we said, we will be getting that full $400 million a year of cost benefit exiting calendar Q4 this year.

The other thing I'll comment on relative to our ability to navigate the market, our inventory position that we discussed being sort of more in check and improving sequentially allows us to be more selective.

And when we talk about selectivity, that's -- think about it as quality of the business which would be profitability.

We have principal measurement and all that.

Yes, I mean if you look at -- I'll provide a little bit of additional commentary, which will be largely consistent with what I talked about at the last earnings call, is that as we move through the balance of the calendar year, one, starting from a top line perspective, we expect our top line to improve at a meaningful rate as we see demand pick up, capacity enterprise as well as seasonal pickup and demand from both a flash and from a hard drive perspective.

That will allow us -- obviously, we'll have higher revenue.

We will also have more of the benefit of the cost and expense reductions as we particularly move into the September quarter and then into December.

And so that will allow us to see some lift in terms of our earnings as we move through the back half of the year.

The big question, which is the question that you're trying to get at is what is our margin level going to look like?

Let me provide a little bit more color on that in the sense that we continue to expect that our hard drive margins will improve as we move through the balance of the calendar year.

Flash becomes the wild card.

Now I said this last call and I'm going to say it again because it really is the same answer.

What are we assuming?

We are assuming that we're going to continue to see pressure in terms of our flash gross margins as we move through the balance of the calendar year.

And the reason that we're doing that principally is, is that we want to plan for the worst, and if you want to call it hope for the best.

Now we don't know exactly what's going to happen to flash gross margins because a lot of that is dependent upon other -- factors that are outside of our control: what do our competitors do, production levels, demand levels and all that.

So we don't know exactly how it's going to play out.

But from our standpoint, it is safest to assume that we're going to continue to see our flash gross margins remain under some degree of pressure as we move through the calendar year, albeit maybe at a moderating level from a pricing perspective as we move through to the back half of the year.

That's very helpful.

If I may just ask more of a question on the hard drive business.

Steve or Mike, would you endorse that part of the reason why nearline demand has been soft, at least for the last 2 quarters, may be due to a pushout in storage arrays until these higher-capacity nearline drives progress through qualification.

And secondarily, I'd appreciate hearing from -- your thoughts on how you view the competitive landscape changing, if at all, from nearline drives as the industry providers offer slightly different to nearline technologies in the back half.

Yes.

So I think the principal driver of growth was actually broader inventory levels within the broad-based hyperscale levels.

So yes, I think there is some modest waiting for the next capacity point that played, let's call it, a secondary role.

But the primary role for calendar Q4 and the drop we saw was really inventory levels and then our customers making an inventory correction.

Relative to sort of competitiveness in the back half of the year, I talked about that.

I think there's multiple approaches.

We're quite comfortable with our approach.

It gets there at the next viable capacity point with less heads and disks which will give us a cost advantage vis-à-vis our competition.

So that's the application of our leading technology in a way that gives us an advantage.

So from our standpoint, we've talked about energy-assist being important to us as a technology enabler.

We continue to be confident that it will give us the benefits we expect, and we will see that in the next generation of product.

And our next question comes from the line of Mark Delaney with Goldman Sachs.

The first question was on the NAND underutilization expenses.

And I was hoping you could be a bit more explicit about whether or not you still expect to be taking underutilization in the second half of the calendar year.

And then related to that topic, my second question was -- I'll ask them both now.

At whatever point those underutilization expenses do stop getting excluded, where do you think you'll fall within that 15% to 25% annual cost per bit target that the company has for NAND cost downs?

Okay.

Well, in terms of the underutilization charge, as you'd heard me say earlier, the majority of that will be taken and reflected in the first half of this calendar year, and then a very small portion is left for the remainder.

And that is the full extent of our current plan for utilizing the fabs.

As it relates to any changes to that, Mike, that's...

Yes, relative to any changes, as you would expect, we'll continue to monitor market conditions, but at this point, no additional plans beyond what we've already announced.

And then relative to sort of the cost implication once, of course, we're through that period, that's reflected in our estimate of around 15% annualized cost down.

And our next question person the line of Mehdi Hosseini with SIG.

Two follow-ups.

Mike, you said nearline (inaudible) growth of 30% '19 versus '18, is that for Western Digital or is that the industry guide?

Yes, Mehdi, we think that's an industry opportunity.

And obviously, we would hope to do a bit better than that.

Got you.

And then with regards to NAND bit shipment growth of low 30%, what is your expectation for NAND production bit growth?

That is production bit growth, not demand.

Supply growth.

I see.

Supply growth, okay.

That's right.

And then given your comment on inventory declining for both NAND and HDD into the June quarter, should we assume that days of inventory has finally peaked and is going to decline looking forward?

We would expect days of inventory to be roughly in the same range for Q4 and then to decline and -- begin to normalize in the back half.

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

I had one on capacity enterprise.

You saw better demand this quarter.

Have you seen any indications that the demand for second half could be stronger than what you initially thought?

Or do you see a risk that we can run into a similar situation as last year where first half is stronger and then second half could be softer?

No, our view and the reason we commented on our expectations for calendar year growth is that we'll see some additional strength in the first half, and we believe the second half will remain strong.

So that drove our update and our forecast from the low 20s year-on-year total bit growth to approximately 30% year-on-year bit growth.

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

I guess I was hoping first question to focus on gross margins.

I believe before you were thinking that you would take the cash charges most heavily in the March quarter and it would be more bell shaped.

And now it appears like June quarter, based on your guide, is similar at a roughly $185 million, and you're now taking further utilization plan charges into the back half of '19.

So could you share with us...

No, no.

Sure.

So just to clarify -- let's clarify.

Our underutilized -- underutilization charge, the largest portion was in the March quarter, then we are expecting between $55 million and $75 million in the June quarter.

And then that will account for virtually all of the -- they'll be just a very small amount that will go beyond the June quarter.

Yes.

C.J., just to reiterate, we have not modified our production plan at all from the original announcement, so hence what Mark just said.

Okay, that's helpful.

And I guess as you think about the adjustment on the leverage ratio amendment, that was really all about the higher rate that goes into January '20, right?

So you still have the 4.25x into September, December.

That's unchanged.

So curious, as you look at adjusted EBITDA on an LTM basis and as you project into the back half, is that something that you're going to need to negotiate in your view today?

No.

So that was the benefit of the amendment.

The amendment really had 2 great features.

One is what you talked about, which is we push out a step-down for our total leverage maintenance covenant from 4.25x to 4x.

We push out that step-down by 1 year.

So it just provides that additional headroom.

Now the second important piece is we changed the definition of adjusted EBITDA to increase adjusted EBITDA by the amount of the depreciation for our portion of the JV CapEx, basically, that is...

That's billed to us, in effect, by TMC.

Exactly.

But it was previously not included in the definition of adjusted EBITDA.

So that was a -- that provided us a significant benefit in the calculation of that leverage ratio.

And our next question comes from the line of Tristan Gerra with Baird.

If market conditions warranted, basically how much room do you have to further reduce expenses beyond the $800 million and your reduction that you're targeting?

Well, we have not dimensioned that, certainly not externally.

The one thing that I will add to that is, is that -- and we commented on this previously is that when we look at -- let's just talk about our expense structure.

Forget about the cost structure for the time being.

But one of the things is that we were attempting to not do was to disrupt or harm our future product roadmap.

If market conditions -- I mean, it's hard to dimension this because you don't know.

If market conditions got more challenging, first off, you'd have to evaluate why they got more challenging and sort of understand that.

But if they got more challenging, we would have to take a harder look at our go-forward product roadmap.

Right now, we don't think that's necessary based upon our expectations both from a market perspective and from a financial perspective, but that would be the next place that we would go look from an expenses standpoint.

And like I said, we have not dimensioned any potential opportunity as it relates to that.

Okay, that's useful.

And any commentary that you could provide in terms of NAND flash inventory that you see in the channel?

So we know it's flat or near, but what's your view in terms of what's sitting out there in the channel?

Well, I think we've seen the channel is getting in better shape.

I think it's a little bit of a different story with certain manufacturers, and there's a different story with each of them, obviously.

But our case has been that we wanted to get ourselves in a better position relative to flash inventory.

We feel comfortable about the actions we took.

We think it's put us in that position.

So we see things sort of trending down from here.

And our next question comes from the line of Sidney Ho with Deutsche Bank.

Last quarter, you talked about that previously NAND wafer starts will reduce its supply by 10% to 15%, which you kind of reiterate today.

But then you also said you will continue to assess the situation.

With 3 months -- more months of data, what are your thoughts today with in terms of further lowering production?

And maybe remind us what portion of your bit supply is in 2D NAND right now.

So at this point, we don't see any need to reduce our output schedule.

We talked about the industry supply growth rate being around 30% or -- at or slightly below that.

We're comfortable with that position relative to our production plan given our current view of market outlook.

We have not commented on the percentage of output on 2D planar.

I'll give you this color that remaining percentage of output tends to be in long life cycle businesses that have quite good economics.

So we're very comfortable with the remaining percentage of our output that's on 2D planar.

Okay.

Maybe just a follow-up on that.

I think you mentioned CapEx will be down in fiscal 2020.

Are there any changes for this calendar year?

I might be mixing calendar year versus fiscal year there.

So I think just to dimension it, we've previously said for fiscal '19, our expectation for cash CapEx would be in the $1.5 billion to $1.9 billion range.

And we believe we'll come in at the low end, around $1.5 billion for fiscal '19.

For fiscal '20, as Mike indicated, we expect our cash CapEx to be lower, and it should be in the $1 billion range at this point.

And our next question comes in the line of Joe Moore with Morgan Stanley.

I wondered in terms of the utilization comments, you said that the utilization charges are winding down now.

Does that mean that the supply comes back on in the third quarter?

And should we be modeling for a step function increase in production in Q3 -- in calendar Q3?

Yes.

I think the way you should think about it is our industry growth rate for the year will be at or slightly below the 30% we talked about.

We chose to implement it in such a way.

We took pressure off of our inventory position in the first half of the year which is seasonally slow.

So I think generally speaking, we think we're well aligned with our view of market demand, and our inventory position is in a reasonably good position as we sort of go through this quarter into the back of the year.

Yes, and let me add to that because when we announced these cuts in terms of wafer starts, we had a simple goal which was to get our inventory levels down to a range that we felt comfortable with.

With the planned cuts that we have made we are moving in that direction.

Our intent was not to eliminate all pockets of inventory within the system.

That will require other activities.

So we are in a position to be where we thought we wanted to be.

And frankly, we were talking about it exiting the June quarter, so we're kind of in that spot or moving in that spot.

But I just want to make sure that everybody's clear in terms of what we were trying to accomplish as a consequence of the wafer starts that we made to our production levels.

That makes sense.

And then separately on the lower of cost or market inventory charge.

Can you talk about -- I expect DRAM's generally moving around.

What makes a move big enough to sort of trigger a charge like that?

And do you expect it to repeat in the coming quarters?

Sure.

So with respect to the driver, it has to do with our multichip package product that goes into smartphones, and as we said, these carry the DRAM that we purchase.

And it turned out one of the reasons it was a higher charge than you might have expected was just a function of the amount of DRAM we had in inventory.

And as it relates to future quarters, this is something we evaluate every quarter.

And this one just was higher than -- much higher than typical, so we spent time explaining it.

And our next question comes the line of Jim Suva with Citigroup.

As we look back at the consumption and digestion, the cloud, what they've procured, purchased and then used, can we look back and kind of help us understand whether -- do they just pre-buy a lot of memory and storage or do they find better efficiencies with cloud to use it?

And is there still more like compression that's really helping them out?

But what exactly was the disconnect of why it's taking so long to digest the inventory situation now that we're kind of coming out of it at some point?

Yes.

So listen, those -- all those factors are at play, but the principal one that occurred in the tailwind of calendar 2018 if you remember, lots of components were on constrained supply, so DRAM, flash and even capacity enterprise.

So in that environment, the hyperscale players are very concerned about availability, about the ability to continue to build out their infrastructure, so they bought ahead of demand to secure that supply line.

So once the supply -- overall supply environment began to change, they took upon themselves to normalize inventory positions.

So that was the primary driver.

Yes, there's always an efficiency effort going on within the hyperscale, but that's generally comprehended, well within our growth rate expectations.

The thing that impacted at least us in calendar Q4 was what I just described relative to inventory adjustment.

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

I would now like to turn the call back over to the CEO for any further remarks.

All right.

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

Have a great rest of the day.

Ladies and gentlemen, thank you for participating in today's conference.

This does conclude today's program.

You may all disconnect.