Micron Technology Inc. (NASDAQ:MU) UBS Global Technology Conference Call November 12, 2018 11:45 AM ET
Sumit Sadana - EVP and Chief Business Officer
Tim Arcuri - UBS
Good morning. Thank you. We're going to get started. I'm Tim Arcuri, I'm the Semiconductor Analyst here at UBS. And we're very pleased to have Micron Technology with us, we have Sumit Sadana, who is the EVP and the Chief Business Officer at Micron. So thank you, Sumit.
So, I think, that -- I think, you wanted to say something, maybe a Safe Harbor initially.
Yes, definitely. So you have -- I'll be making some statements that could be forward looking in nature. So please consult our SEC filings for our risk disclosures. With that we can proceed.
Q - Tim Arcuri
Great, thank you. So why don't we start by just giving sort of an overview. You're in a very unique seat, because you're sort of at the center of the business strategy for the company and you talk to other customers all the time. Can you sort of give us an overview of sort of where the company is today?
You came over, there was a bunch of folks came over to Micron, of course, Sanjay. So, can you sort of give your perspective having been at SanDisk and sort of what you've learned and what the team has done since being at Micron. Thank you.
Definitely. So, yes, I mean, I have been with Micron since June of last year. Sanjay, was announced as the CEO in May have 2017. And we have had good influx of talent into Micron from many different companies, and we are very proud of the leadership team and the talent in the company in the engineering ranks, and other functions is just outstanding. So we're continuing to build on that.
I think our priorities that we have made clear relate to sort of this next chapter that we believe we are able to drive in a very significant way, which is the new Micron story that we talk about. Micron is exceptionally well positioned, we’re at the nexus of multiple secular demand drivers that are long-term in nature. The DRAM industry is pretty well consolidated. We have 70% of our revenue and much more than that of our gross profit percentage coming from DRAM. And besides the growth drivers and the consolidated position in DRAM, we have some phenomenally exciting technologies in the hopper, like 3D cross point, as an example.
We have also been making tremendous progress on the profitability improvement initiatives that include getting our costs to become extremely competitive. And we have outlined our significant progress on that front. We've spoken about $9 billion profitability improvement measure at our Investor Day in May and we are -- we’ve made dramatic progress towards that goal over two-thirds of the way there already.
So our core profitability as a company at any part in any market condition is dramatically higher today than it has been at any time in the past. And then we have other initiatives, like improving the mix of our high value solutions. We have been working hard at that and that is certainly a journey, but we have been making really good progress on that front. The -- for example, the percent of our total NAND output that we sell as solutions versus selling them as components has been steadily rising through the last several quarters and we expect to continue to make progress on that effort as well over the next couple of years as new products come out.
So very excited about the -- both the market opportunity, the way we work with our customers, and how we are changing that relationship to be lot more strategic in nature. Dramatic improvements in our core profitability and underlying financial performance and couldn’t be more excited about the future.
Great, thank you for that. Can we talk a little bit about China and maybe some of the recent U.S. government actions against China that relate to DRAM. There have been recent restrictions on Gin Wa [ph] which is a company that you've been in dispute with around IP. There's a lot there to potentially unpack. I know you can't say too much, but I thought it was interesting that that DOJ has clearly been involved in this case, but the action came from the U.S. Department of Commerce much like ZTE.
And I also thought it was interesting that the Chinese Government Official release was a bit measured, it didn't carry the normal retaliatory language in it and it was more about promoting normal business. So China is a big end market for you. So what can you say about this, and sort of where does this leave Micron in the industry?
Yes. Sure, happy to discuss that. So, China is certainly an important market for us. If you see our SEC filings, we have said more than 50% of our revenue comes from China. But I think the important thing is that revenue includes a lot of revenue from customers around the world who have their supply chains in China. So our revenue that comes from companies that are headquartered in China is substantially lower.
But nevertheless, we have a very important supplier to the China ecosystem. We have strong customer relationships there. We have good business there. And we also have R&D facilities, manufacturing facilities there, and we employ thousands of people in China. So we are pretty well embedded in the ecosystem in China and it's a great market for us.
In terms of our technology, we have always advocated fair competition. And so we are not afraid to compete with anyone, but the competition has to be fair, which means that we have -- we are going to always defend our intellectual property. We have spent decades and many, many billions of dollars creating this intellectual property. So we take it very serious to ensure that it's not infringed. And in this particular case, it was IP theft situation, which we were very much interested in resolving.
So we have filed some cases against the companies in question regarding these theft of intellectual IP, intellectual property. And we are very glad to say that the U.S. Government has also taken its own independent action around this particular case. They have looked at the details of the case, they've analyzed it in depth, and broad their own actions, which we feel good about and we are thank for. At the end of the day, we are very hopeful that the two countries U.S. and China will have improved trade relationships and I think for the good of the global economy that's an important goals for two countries to have.
But as far as Micron is concerned, we will continue to defend our intellectual property and continue to advocate for fair competition across the board and continue to work with both our customers and the government at all levels in all the countries where we do business, including in China for continuing our business there in a positive way.
Great, thank you. Can we sort of click a bit down into DRAM, and maybe you can summarize the current environment in terms of both supply and demand. Obviously we've seen some demand slowdown, some digestion from the hyperscalers on the demand side, on the supply side, your biggest competitor is pushing out wafer capacity addition, somewhat a typical behavior before pricing even goes down to see that much wafer capacity pushed out. So can you sort of talk about the dynamic, both on the supply side and the demand side?
Sure. Yes, that is true, I mean, we have been monitoring the developments, both from the semiconductor equipment companies announcements about what they are seeing about orders, as well as from our competitors. And of course we do our own internal planning for our own capacity. So I would say that it is definitely interesting to note that so much cut back has been taking place on the DRAM side from a CapEx perspective coming at it from all of the announcements in the marketplace, from competitors and equipment vendors.
So, I think, the supply side of the situation is in a pretty good place. From a demand perspective, there are some near-term issues that we had highlighted in our last earnings call that relates to some inventory adjustments that are taking place. We think those inventory adjustments are continuing. And besides the inventory adjustments -- and of course, the inventory adjustments relate a lot to a fairy lengthy period of price increases that were happening quarter-on-quarter in DRAM, so customers felt that it would be good to carry higher levels of inventory in that environment as understandable. So certainly the inventory adjustment period in underway.
And then we had also spoken about the fact that certain types of processors and client computing are in short supply from one of the important suppliers and that is also causing some near-term and medium term impact to the overall market demand. We view these factors as relatively temporary and the longer term trend on demand, whether it is cloud demand. If you see the number of datacenters that are being built out by some of the top cloud companies, some of their two year, four year, sort of, growth trajectory that they are planning for. These cloud companies have very, very strong roadmaps for continuing growth that will require ongoing increases in CapEx.
And certainly we have spoken in the past how DRAM CapEx -- DRAM spending and NAND Flash spending constitutes pretty significant part of the server cost today. These servers that are sold in cloud have 60% to 70% of their cost in just DRAM and NAND Flash. So, as the CapEx for spending in the data centers by cloud companies increases over the next few years, we see ongoing robust trend in demand there.
So, there are some short-term factors that are playing themselves out. And with the supply situation I mentioned, we are pretty optimistic that the DRAM market will continue to be very healthy.
Thank you. So, I wanted to follow-up on that. So, just more a question about how you run the business longer term. It used to be the margins across the cycle were very low. I think that your gross margin in DRAM was about 20% in the 15 years leading up to 2015. It sort of averaged 20%. So, at the op-level it was basically zero across the cycles. And of course in 2016, gross margin went to 56%, or rather 2017 gross margin went to 56% and it’s going to be almost 70% this year. So, obviously something has changed. And we all know that cost cutting is getting more challenging with the shrinks.
So, obviously the idea seems to be to pass on the cost reductions to your customers. But the question is, what is the right margin level off of which to actually do that? Is it 40%, is it 50%, is it 60%? How do you sort of think about that?
Yes, that’s a great question. There is no doubt that the gross margins have expanded quite significantly. One important thing I’ll mention is that we had also shared some exclusive data in our Investor Day a few months ago, but we have had a cost structure in DRAM several years ago, which was 2x the cost of the leading competitor. And from that point, we have steadily brought that gap down. And we are in striking distance now of the world-class competitive cost of the leading-edge technology node that is out there in the industry.
So, we have made dramatic progress on the cost front in DRAM. And certainly, the industry in no small part, thanks to Micron’s own strategy of consolidation is now down to three players, which constitute 96% of the bit output in DRAM. And so, with those dynamics, we continued to believe that the market is going to be very healthy, the sustained gross margins in the DRAM business over the next several years will be significantly higher than they have been over the past decade. And with the cost improvements, we had made and with the structural improvements in some of the demand trends.
And one notable example there I will give you is that, many years ago the PC demand used to be two-thirds of the overall demand in DRAM, and now the PC demand is just a little bit over 10% of the whole market in the teens. And so, that is how much that PC segment has come down in terms of overall share of the DRAM business. And the business has become a lot more diversified between servers and the cloud, servers and enterprise, mobile, et cetera.
So with that dynamic and our cost structure and the supply situation, we think the sustained gross profits would significantly higher. We think our ability to generate free cash flow, which is something that in the new Micron, management team is extremely focused on is maximizing our free cash flow, and maximizing that throughout the cycle in different market conditions. That is going to be at a level that I don't believe our investors have fully comprehended those structural changes yet. And so, we are very optimistic and very excited about demonstrating that in our results over the next few years.
Great. I had one more DRAM question, and then we can shift to NAND, but it's sort of interesting this year, it's that it's not really about market share so much anymore. If you look at -- you look at supply, everyone is adding about the same amount of bits this year, bit growth is pretty similar this year between you and the others as well. However, from a CapEx perspective your CapEx is guided up and your peers is CapEx looks like it will be down. So can you talk a little bit about your capacity planning?
Sure, I think it's a great question. Micron stats has been marked by periods where we have not been able to invest adequately in ensuring competitive ramps of new generation technologies and that has certainly been one of the factors if you look back the past three, five, seven years, which has impacted our competitiveness of our cost structure. And so one of the important things that we are trying to do is to focus on the investments that allow us to ramp new technologies on time and create a very, very cost effective operations structure inside the company, compared to the best of breed in the industry.
And one of the things that we find ourselves in is been a space constrained environment. So new generations of DRAM technology and even NAND technology for that matter, require a lot more tools and these tools require a lot more space. So new space, new cleanroom space is needed simply to keep the wafer capacity at flat levels. So as new cleanroom space is not added, wafer capacity would start to degrade quite substantially.
So we have had in our fiscal 2019 budget set aside, more than $2 billion incrementally over fiscal 2018, simply to add cleanroom space and manufacturing space just in order to aid technology transitions without increasing any wafer capacity, or minimal increases from time-to-time. So, at least the capability for future additions would be there, when this cleanroom spaces built out and we are not going to find ourselves in an environment where we need to make techno transitions and we don't have space to do that and we consequently end up impacting our cost structure.
So, if you look at our equipment CapEx, it is actually down in NAND year-on-year in fiscal 2019. If you look at our total CapEx, we did about $8.2 billion in fiscal 2018. We have guided to roughly $10.5 billion for fiscal 2019, but more than $2 billion incrementally year-on-year is just the building aspect of it, the cleanroom space and other manufacturing CapEx increase not related to equipment.
And so if you look at the overall CapEx, it's very flattish on an equipment basis, when you back out some of these adjustments. So we feel very good about where we are on the CapEx side related to the cost structure and we are focused on getting substantially all of our bit growth happening in the form of technology and node transitions rather than wafer increases.
So that is to say overall WFE flat, but NAND down, DRAM probably still up a little bit?
Very small changes, but again, mainly for largely what you're saying is directionally correct. The focus of our CapEx just remains on the tech transition side. So, yes, that we're just focused on facilitating enough space. So we have flexibility for the future and being responsible with the CapEx and managing the business with an eye towards end market demand.
And, of course, one thing I will also say is that we are continuously monitoring the market environment and we always aim to be very flexible and very agile with our decisions around CapEx. Because like we said, we are extremely focused on ensuring a healthy market environment and maximizing our free cash flow.
Great, thank you. Maybe we can shift to NAND. NAND is interesting because it seems like a much different dynamic structurally sort of in terms of the market. Whereas in DRAM you’re seeing consolidation, NAND is the opposite. You’re seeing these former joint ventures are breaking up and so two are becoming four. And so, you’re actually seeing the number of competitors in NAND seems to be going up not down, yet the elasticity in NAND is obviously much higher.
So, sort of in terms of how you plan CapEx in NAND, do you keep spending because elasticity always will be there, because it seems like a less favorable competitive dynamic in NAND than it is in DRAM?
Yes, I mean, it is true that there are six important players in NAND who are manufacturing NAND at decent scale versus three main players in DRAM. So, it is -- what you are saying is true that there more players in NAND. We are certainly very focused in ensuring that our overall cost structure in NAND is extremely competitive. And one important thing that we have done there is get to a die size with our latest generation 3D NAND that we believe is industry-leading, which gives us phenomenal cost structure.
And so, we believe that the combination of the cost structure we have in NAND, which is world-class. And the leadership and technology with QLC is going to allow us to really get ahead of the competition when it comes to our solutions that we are putting out in NAND. So, we have made announcements on the industry’s first QLC SSD in the enterprise. We’ve just recently announced the consumer NVMe drive, SSD drive that’s also QLC.
And so, we feel like we have certain advantages that we are pressing. And the other aspect is we are also growing our mobile solution business with NAND. So, even though there are six players as we mentioned in NAND versus far fewer in DRAM, areas like MCPs, multi-chip modules that go into mobile phones that combined DRAM and NAND in a single package are ones where there are only three main players. Because you need DRAM to be able to supply a multi-chip module effectively -- cost effectively to a mobile phone company.
So, in areas like that we are growing our business very, very rapidly. And you can see some of the impact of that mixed shift that is happening in our NAND business in our financial results as well. So, when you look at our reported gross margins, there has been a lot of angst that I have heard from investors about what’s happening to pricing in the NAND business. And if you see our reported results, our NAND gross margins are doing very well. And part of it is the ongoing improvements in our cost, driven by the ramp up of our newer technologies. And the other aspect is driving these solutions, where we have advantages.
So, we feel very good about our NAND business. And in spite of the competition, we feel that the market itself is going to become more healthy at some point in time in the next few quarters, because there has been several announcements of cutback in CapEx. And there is dramatic improvement taking shape on the demand side, with the average capacity increases that are happening due to elasticity. The significant improvement in volumes that are happening in many, many client-based devices on the SSD front.
So we are very optimistic about the long-term health of that market, in spite of the larger number of competitors, because of the elasticity of demand and because of enough [ph] competitive advantages we have.
Just to that point, I mean pricing, obviously has been weak 64-gig MOC is down I think 40% of the peak. But you did mention that you're seeing some elasticity or some signs of it. Can you talk about that, is that a recent development? And, sort of, typically how long after you begin to see elasticity does pricing bottom? Are we talking about two more quarters, we could see calendar Q2 bottom in pricing, can you just sort of talk about what you’ve seen in the past?
Sure. So we don't make explicit projections about pricing in the future. But I can certainly provide some color on what we are seeing. So, when we look at our discussions that we're having with customers around the world in the mobile smartphone space. We are seeing very significant increases in the average capacities of NAND going into phones. With the newer phone models that are designed in, a lot of phones were 32-gig are going to 64 gigabytes of capacity. A lot of phones were 64 gigabytes are getting 128 and 256 gigabyte capacity points.
At the high end, we see a significant move by some of the tier 1 companies in pushing 256 gigabyte and even 512 gigabytes SKUs as greater percentage of their mix and they're using the lower prices that are taking shape in the market to actually make that transition happen.
The other aspect that is happening is the average capacities in client devices, SSDs. Again if you look at the number of platforms that the laptop and PC companies are designing in with SSDs versus hard drives, there is pretty noticeable shift in the number of platforms that will be natively designed to SSDs instead of hard drives. So the hard drives volumes will be coming down, SSD volumes will be going up.
And certainly the average capacities in those SSDs, which were long clinging to 128 gigabytes and 256 gigabyte capacity points, which many consumers didn't find to be adequate, and they had some kind of storage anxiety on will this be enough that those capacity points are now moving to 256 gigabyte and 512 gigabyte quite aggressively. And the pricing in the market is allowing them to make those shifts without material changes to their borne cost.
So both in terms of the number of platforms, as well as the average capacities on the consumer side are increasing, and we are seeing very healthy increases on the enterprise side as well. The average capacities are going up significantly on the enterprise side, as these datacenters appetite for NAND Flash is very, very high and more workloads are shifting from high end hard drives to Flash, because it is becoming so attractive.
We just introduced, very short while ago QLC drive, that has the same kind of cost per gigabyte as 10K RPM hard drive. So these prices are -- in the enterprise, so these prices are becoming incredibly attractive.
And if you look at all the other benefits that Flash brings with it compared to hard drives, dramatically lower power consumption, order of magnitude, lower defectivity and failure rate. And the ability to have dramatically higher performance that can actually shrink the amount of server capacity you need. Because there are several workloads that are storage bound, meaning you need to buy more servers in order to just be able to hookup the amount of storage that you need. You can actually do with fewer servers if you deploy more Flash instead of hard drive.
So Flash actually becomes a pretty important deflationary force in the enterprise. And the cloud companies who have a very, very strong handle on their infrastructure know that very well, which is why their consumption of Flash as a percent of the total is so much higher than the enterprise -- in traditional enterprise.
So Flash is seeing a resurgence of demand, driven by the lower prices and tremendous impact of elasticity taking shape in the funnel. And just as all of that kicks in over the next couple of quarters, you will see the impact of the cutback in supply. So, I think, the industry dynamics will become healthier substantially better is what I expect over the course of the next few quarters.
Just on that point, do you think also that there is an added effect in that, now that we sort of push into 96-layer that the transition to 96 is going to be fundamentally more challenging than the 64-layer transition was, so that you could have an additional sort of limiting factor on the supply side?
I think that's a terrific point in going from 32-layer to 64-layer, there was a very substantial increase in the bits per wafer and going from 64-layer to 96-layer, the increase in bits per wafer is dramatically smaller. And then looking ahead to the next generation beyond that: A, it is going to get more difficult; and B, the bit growth from going from one technology to the next will continue to erode, will continue to fall. And that will have a natural moderating effect on the longer term bit growth rate in the industry.
And what we are going through now is sort of an artifact of the step function increase in supply that came from the transition of the industry from 2D to 3D NAND and from 32-layer to 64-layer, which were both pretty substantial increases of bits per wafer that a lot of it -- lot of that is behind us, and we do expect the bit growth rate in the industry to fall quite a bit from 2018 going into 2019. And that's part of what gives us optimism about the long-term health of the industry.
I have many, many more questions, but maybe we have anything from the audience, I want to open it up for a couple questions from the audience.
Okay. Can we talk about your networking business and your embedded business? As a reported segment it's about $3.5 billion worth of revenue in fiscal 2018, but I believe 10% to 15% percent of CNBU is also embedded. So that's about 20% total of the company's revenue. So that's about a $5 billion to $6 billion business that has proven to be much less cyclical than the rest of your business.
I could argue that the entire valuation of the company is supported by that business alone. So the rest of Micron basically is free today. Can you talk about that business and sort of why it is so different than the rest of the company?
Yes, a substantial part of our EBU business or embedded business is automotive and industrial and there is a small portion -- I mean, there is a decent portion of consumer business there today. But the automotive and industrial portion of the embedded business is expected to continue to grow over the long-term. And that automotive and industrial business is a business that has very different dynamics than the rest of the business of the company.
The networking business inside of CNBU, which you mentioned also has very similar dynamics to the automotive and industrial business, which is that the business is very steady it's more predictable and the peak to trough declines in pricing from one part of the cycle to the other are far, far more muted. And so the business generates very healthy free cash flows and also relies on older technology nodes that ship for long periods of time through these long product cycles that last anywhere from 7 to 10 years.
And so that business is exceptionally good business and we have demonstrated our commitment to that business, with announcements of increases and our investment both on the R&D side, but more notably, we recently announced the increase in our Manassas, Virginia site for long leaved assets. And we really like the predictability in that business, the free cash flow aspect of that business and the much, much lower volatility in that business.
So we think it's a great business and the fundamental trends in the industry when it comes to artificial intelligence and machine learning and autonomous driving are really going to drive the long-term growth in that market and that too delivered in a very steady and far less volatile way in terms of financial results. So very optimistic about that business.
Thank you for that. Maybe just a question about the divorce from Intel, can you maybe spend a little time in terms of what the financial impact is? Is it a net inflator to CapEx; is it a net deflator to CapEx? And the same question about OpEx as well.
Yes, so in terms of what we announced with our separation with Intel in times of our joint development first on the 3D XPoint side, and then without joint manufacturing as well. We have mentioned that we do intent to call the IMFT JV assets, after the January 1, date when we are able to make that determination. And when that happens, there is a period of time over which, Intel and Micron will decide when the closing of that transaction occurs. We don't expect, I mean, there is some chance that it could happen towards the tail end of the fiscal year, but we expect more likely than not it will happen in the next fiscal year, fiscal year 2020.
When that does happen, the transaction will take place based on the contractual agreements between the companies related to the book value of the asset. That asset has $5 billion or $6 billion of capital that has gone into it overtime and we'll have the opportunity to purchase it for just over $1.5 billion or something like that. So it's a very good asset for us to have. And it has a terrific team that comes with it that very much focused on doing effective development, as well as manufacturing in the technology.
And we have an intent that our long-term financial model as it relates to CapEx is not going to change with all of the moving parts that we have related to different aspects of how we manage our manufacturing. So there could be some short-term fluctuations, but longer term, we expect to stay within the low-30s as a percent of revenue guidance on sort of expectation on CapEx.
Great, thank you. We are out of time. Thank you, Sumit.
Thank you, Tim.