© 2024 Alpha Spread Limited. All Rights Reserved.
www.alphaspread.com is not operated by a broker, a dealer, or a registered investment adviser. Under no circumstances does any information posted on www.alphaspread.com represent a recommendation to buy or sell a security. The information on this site, and in its related newsletters, is not intended to be, nor does it constitute, investment advice or recommendations. The information on this site is in no way guaranteed for completeness, accuracy or in any other way. In no event shall Alpha Spread Limited be liable to any member, guest or third party for any damages of any kind arising out of the use of any content or other material published or available on www.alphaspread.com, or relating to the use of, or inability to use, www.alphaspread.com or any content, including, without limitation, any investment losses, lost profits, lost opportunity, special, incidental, indirect, consequential or punitive damages. Past performance is a poor indicator of future performance.
You don't have any saved screeners yet
Intrinsic Value is all-important and is the only logical way to evaluate the relative attractiveness of investments and businesses.
Okay, great. So good morning again. Thank you for attending J.P. Morgan's 16th Annual Technology Investor Forum here at CES. Again my name is Harlan Sur, Semiconductor and Semiconductor Capital Equipment Analyst here at the firm. Very pleased to have Ernie Maddock, CFO of Micron.
Since Micron is such a broad based supplier of solutions to many different applications, I think their products are pervasive in all of the products that you are going to see here at CES. And so rather than have him talking about a snapshot of what they’re showcasing because they are in everything, I think we’re just going to dive into the Q&A, and I think that the first question will be the announcement that you put out yesterday about the -- maybe give us a recap of the Intel-Micron announcement yesterday, after a 12 year JV partnership, very successful partnership. Both will be working independently on future generations of 3D NAND post development of the third-generation NAND family in 2019, so love for you to start us off by giving us a recap of that.
Sure. So I think one of the things important is to be clear about what was impacted by the announcement and what wasn't impacted by the announcement. And we used the term JV very broadly, but specifically what was impacted by the announcement was the joint development agreement that the two companies had relative to the development of NAND.
What was not impacted by the agreement was the JV fab in Utah, which predominately focuses around the production of 3D XPoint which is known as IMFT. That is the physical manufacturing joint venture. We also continue to do joint development with Intel on 3D XPoint. So it really is a fairly focused change relative to the ongoing development of NAND.
And as was stated in the release, basically as our technical teams continue to look and evolve the roadmaps for 3D NAND, we determined that based on the markets that we serve and Intel jointly determined that based on the markets that they serve, that the roadmaps or the technical roadmap would need to diverge in order for both companies to leverage the maximum opportunity that exists for our respective markets.
So that was the essence of the announcement. It is -- the reason it was a joint announcement is it was very thoughtful and very carefully considered and we do not believe this changes any other dimension of the relationship that we have relative to those other products.
When I think about the relationship, and I do view it as a successful relationship and I thought that -- historically I think the Micron team has had a track record of best and stellar sort of execution, especially on planar NAND. As soon as the team hit 3D NAND, the execution improved, competitiveness from a cost per bit perspective, technology perspective, design perspective improved. Now you guys have a little bit of a tailwind that’s helping you guys to maintain that leadership position.
So I think potentially doing this now maybe is a good decision for the Micron team. There are some puts and takes though, and the takes would be that the R&D has always been shared on the 3D NAND side between Intel and Micron. So you are going to have the Intel part of the R&D contribution going away. So how does that impact Micron’s R&D momentum on a go-forward basis.
And then on the positive side, you have been obviously selling your bits to Intel at near cost, and you’re going to be able to take all of that now and sell it at fair market value, which is a tailwind for your revenues and a tailwind for your margin. So help us understand the two dynamics there?
Sure, those two things are actually quite separate. We do have these arm's length agreements where we sell product to our partner, and those have a natural expiration in 2018 and 2019 and have nothing to do with the decision the parties made to diverge on the technical path. So those are agreements that we’ll be rolling-off here. And the company will be doing what it thinks is in its best interest relative to redeploying those bits, and that could be another agreement, that could be selling more SSDs, more novel managed NAND. But it is important to know that that has nothing to do with the decision on the technical path.
As you noted, we will be bearing more R&D expense relative to the ongoing development of NAND in the grand scheme of the company’s expense structure. This somewhere in the range of $10 million to $15 million a quarter, and we would expect to be able to sustain at that level for some time. I wouldn’t expect that there would be any sudden or dramatic increase beyond those levels that might actually as we go through and rationalize things in this new world, be able to do a little bit better than that. But that’s -- I know everyone is interested in sort of framing what the rough order of magnitude is, that would be a good rough order of magnitude to be thinking about there.
And you are right that the companies -- I think the important thing is relative to the current partnership, relative to the third-generation, I can speak for Micron, we have great products that we’re going to be introducing on that third-generation NAND. We’re just now ramping second-generation NAND. And so this really is something from a timeframe perspective that is from a -- that will determine products in 2020-2021 relative to the roadmap that the company has, which is based on that JV development for both second and third-generation. We continue to be very, very enthusiastic about the product portfolio. We’re going to have the expansion of our SSD portfolio, in the context of that, and certainly our managed NAND portfolio, certainly both the mobile and the embedded market.
So we are not letting off the gas for one second anytime over the course of the next several years, as we finish productizing our third-generation product and then begin this exciting new development path for us relative to fourth-generation.
And again I think from my perspective, timing wise it comes at a relatively good time where as you just mentioned, you talked a little bit about the enterprise and cloud traction that the team is getting on the NAND side, and Micron historically has had a small position in the enterprise and cloud SSD market, but this is a big market, right, a billion dollar market growing at a 15% CAGR. And the Micron team has recently introduced some new platforms, like the 5100 SATA SSD platforms for cloud, 9200 NVMe based SSDs for enterprise and cloud. Where is your enterprise and cloud SSD mix as a percent of your overall storage business unit? And do you expect this to increase as we move through 2018?
We do expect that to increase. We’ve just completed in our November quarter a record quarter for the company relative to our SSD sales, and that really extended across both the client market as well as the cloud and enterprise space, that really was just great performance. We think our share in aggregate increased from the mid-single digits to upper-single digits. And we can certainly foresee a future where that number moves into the double digits, do your job well, execute on the roadmaps that we have.
So we're very pleased about that. We do plan to introduce a broader portfolio of products that will underpin our ambitions relative to this expansion of share. So this isn't just a pipe dream. This is a very, very specific product roadmap that we think is going to enable the company to have even more meaningful solutions for our customers, and that of course should enable the expansion of our share, because Micron has great relationships with its customers. They like to do business with us, and if we can provide products at competitive pricing, we tend to find that they're very receptive to buying from us.
Let's turn to just kind of industry fundamentals. It's been a particular concern of the market, especially on the NAND front. There continues to be a concern by investors about the potential for oversupply as you and your competitors ramp 64-layer 3D this year. The market continues to be focused on spot pricing, which has come down since the peak in November and is sitting slightly above contract pricing. We know that demand was held back because of the strong pricing environment for NAND in 2017. Help us understand, how you see the NAND market unfolding in 2018. And what is it do you think investors are missing as it relates to their concerns on big oversupply and significant ASP declines?
Sure. So I think all of the participants in the NAND space have come out and talked about their view of 2018 bit supply growth. I think the low end of the range is in the realm of 40%, Micron is at the high end of the range. We said approaching 50%, so it's something maybe slightly under, but for ease of math let's just say that ranges between 40% and 50% for 2018. At that level of bit supply, our belief is there is ample demand to absorb that level of bit supply.
Remember, it is healthy for pricing to come down and follow the cost curve, and that is the way markets expand, that is the way that we will penetrate more of the client SSD market. That's the way the densities will increase. So this is a -- has the set up to be an incredibly healthy environment at that level of bit growth, so 40% to 50%, the industry in aggregate will have significant cost reductions. They would likely be as an industry, well in excess of 20% maybe even ranging into the middle part of the 20s.
And so that leaves ample room for some price reduction if that is what the market bears out without being terribly detrimental for the margin structure of the industry. So the simple fact of the matter is if you have a point of view that says pricing in aggregate over the course of the year is going to decline north of 30%, that's probably going to cause margin compression. If you were on this pool that says with all the demand drivers and all the supply variability that we're seeing here, that, that pricing will come down generally less than that then you're probably going to see stable to potentially expanding margins.
Yes, I think that's maybe one thing the market doesn't understand is that if the industry is successful where you are actually improving yields and driving 40% to 50% bit supply output, yields must be pretty good. And so, for example, for Micron if you're driving towards the upper end of your target range closer to 50% bit supply growth, should we be assuming that from a cost curve perspective that you're also driving closer to the 30% type of cost per bit declines?
We have said that we're actually planning to be above -- somewhat above the industry. So by implication, above 50% for the company in our fiscal '18. And yes, I think it's reasonable to assume you're going to see the cost reductions that will correspond to that level of bit growth. So we are very excited about the opportunity to continue to reduce costs for having great yield ramp on our 64-layer product.
We talked about introducing our third-generation product toward the end of the year, that's still on track. And so we feel that we will be performing fairly well in the context of the broad industry, but we also think it's going to be performing well, relative to the ramping in general of 64-layer and cost reduction. I think Micron is comfortable that we should be positively differentiated there.
On the DRAM side, little bit of a different story, right? Pricing continues to hold up relatively low, in fact, pricing remains relatively strong in all segments. We just checked the pricing data recently, server/cloud, mobile, PC client, all pretty strong pricing here in what normally is the seasonally weakest part of the year. How do you see the fundamental environment as we move through the first half of the calendar year, what are going to be the drivers of continued strength in fundamentals for DRAM?
Yes, I don't think that I would say any difference between the front half and the back half. But we think the strongest drivers for DRAM bit growth this year are going to continue to be the expansion in the server space, both from the hyperscale guys, as well as classic enterprise servers. Mobile is certainly on track to have another good year of growth. And that would be complemented by growth in this autonomous driving embedded business, which would encompass automotive as it's sort of its marquee piece of that market.
But also you have this increasing trend toward Edge Computing, whether it be your refrigerator or security camera or any other number of devices that today have very limited functionality, but in the future will have expanding functionality. And as a result of that would be driving many, many endpoints that require more and more memory. And those in turn will require more memory compute capability in the datacenter.
So I think one of the most often overlooked thoughts as we look at this market is how interconnected everything is. It's not as if something that happens in the mobile market ultimately doesn't affect the server market, because most of your content on your mobile phone is delivered to you from a server somewhere. So the more content you are demanding on your mobile phone, the more that's placing demands on servers.
If you look at an autonomous vehicle that is generating significant amounts of data, not only is that data being used in that vehicle, and we don't have autonomous vehicles today, but we have vehicles that have significant driver safety systems and sensoring and all sorts of things. That data is also uploaded into the cloud. I had a recent experience, where I drove one. And I could actually get -- download a report every night telling me how I could drive better, whether my intention was to go faster or to save more fuel. Well, that data is going from the car up to the cloud and then down to my mobile device.
So this interconnectedness is really, really an important concept that is broadly driving demand for both storage and for processing capabilities. And of course those are very dependent on DRAM for processing and SSDs and NAND based storage on the storage cloud.
I'm going to back track a little bit, because this has been something that I have been asked about by investors recently with the U.S. tax reform bill being signed off, just after you guys actually reported the November quarter results. Wanted to see if you had any updates for us, I know you mentioned on the call, no impact for fiscal year '18 non-GAAP tax rate. How is this going to impact your fiscal '19 and beyond tax rate?
Sure, so we reiterated there is no impact to our fiscal year '18 tax rate. We still think it's going to be in that mid-single digit rate. As we go forward, I think the most impacted thing from the new tax legislation is this idea that offshore earnings are taxed at some minimum level. And that's the fundamental difference for Micron and I think for a large number of other tech companies as well that generate significant portions of their earnings offshore.
And so if you were to look at 2019 and beyond, and I want to make sure we're clear about this, so we avoid confusion, so if you take 2018 consensus earnings, I presume this occurs in 2019 and beyond, an apples-to-apples comparison for the company would be our tax rate increasing actually from the mid-single-digits to the low teens.
And I think that we'd be able to sustain the tax rate at that level for an indefinite period of time. It’s not a case where we’re going to be in the mid-teens for a couple of years and then skyrocket up to 20% or 25%. Based on our belief of the tax law, of our business and how those earnings will be distributed, that feels like the most reasonable estimate that we can provide for 2019 and beyond.
And of course, I think, it should be clear to all of us that there are many dimensions of that tax code that are yet to be fleshed out. And so we would reserve the right to revise this, if in fact there are significant revisions as these details are fleshed out. But that's the best perspective we can provide at the moment.
So low-teens tax -- effective tax rate on a go-forward basis.
And that's on a non-GAAP basis.
Non-GAAP basis. And -- but you get a benefit under repatriation...
We have NOLs that will -- the reason we are able to maintain our fiscal year ‘18 tax rate where it was prior to the tax law change, is the fact that a lot of the expense associated with the repatriation and/or the total tax rate because not all offshore earnings will be repatriated, is that we have NOLs and our belief is that those NOLs can be used to offset the liability created by that.
Great. Do we have any questions from the audience? So this is CES and it's all about new and emerging areas of potential growth. Last year -- last couple of years the focus has been on automotive. You mentioned a little bit about this as well. At our forum this year we’ve got GM, Wrangler, and Continental and other discussing their R&D efforts on the fully autonomous, semi and fully autonomous driving. These guys were talking about 20 or 40 gigabytes of DRAM and 1 terabyte SSDs per car for level four and level five autonomous vehicles. That's a pretty significant step up in most DRAM and NAND content per vehicle. Is that consistent with how the team sees the content trends in automotive over the next kind of three or five years?
Yes, this is some data that we also shared on a call, the 40 gigabytes of NAND and a terabyte of -- 40 gigabytes of DRAM and a terabyte of storage in the car feels about right to us and I think those are big headline sort of numbers. But I think the other thing that’s important is that, that trend trickles down to something other than those non-autonomous -- fully autonomous driving cars. That trend will be present in every single car that’s made and you are going to move from simplest application, a rear view camera to increasing number of safety sensors, digital dashboards.
So these are all things that in addition to that marquee fully autonomous driving content provide very healthy growth opportunities for us in the automotive segment. And of course we have very, very strong automotive share. So we are benefiting from this market trend and we believe we're helping to enable it by bringing innovative solutions like automotive grade GDDR product into the market, helping automotive customers move more quickly through the technology transition, so that they can take advantage of the capabilities of latest generation memory and latest generation storage.
I think we all know that qual times for automotive are very long and they tend to be qualified by highly rigorous standards. But we're helping our customers ensure that we can get a short a time as possible so that they get the benefit of that latest and greatest generation technology. But this is a great opportunity for the company and those autonomous cars are pretty fun to drive, not fully autonomous, but I can tell it was a revelation from a driving experience for me.
On the earnings call the team actually brought up the areas of home automation, home assistance and other forms of edge computing, Amazon Echo, Google Home, video surveillance, all of these edge devices, which are quite intelligent and these devices carry actually a pretty reasonable memory content, 0.5 to 1 gigabyte of DRAM and actually a few gigabytes of NAND, but the volume opportunity is significant. How do you see this market unfolding, where is Micron seeing the most success, for example, I know that you guys are in the Amazon Echo platform. So how do you see this segment of the market evolving?
I think the most important trend in that market is what -- typically what starts out as a simple device requiring relatively less memory. Ultimately we as consumers want more sophistication from all of those devices. Law enforcement wants more sophistication from those security cameras and the ability to react quickly. We want whether it's the Echo or whatever your choice of home assistant is, you eventually want that to do more.
You want your refrigerator to do things that you never wanted a refrigerator to do before and pretty soon your dishwasher or the entire lighting or entertainment system of your home, all of those things will become more pervasive with the believe that we want to make our lives easier. We don't want to think about when we open and close our shades, we don't want to have to be bound to some problematic thing. We want to be able to go on vacation and say I think I forgot to do this and tell our home assistant to make that happen for us.
So I think the important trend is that what starts out as something simple like in cars, usually ends up being far more sophisticated and that usually spells more memory content and again this is an area of strength for the company, it's an area where we have focus and we have every intention of helping our customers with innovative products that allow us to maintain that lead in the marketplace.
On the cloud and data center side you have got Intel ramping Skylake, AMD ramping EPYC on the server CPU side, cloud and hyperscale CapEx spending is going to grow 30% this year, cloud servers are driving 50% more DRAM content versus enterprise servers, given all of this, and I believe that the cloud server mix is already big, the biggest portion for Micron. If you look at it for the industry mobile DRAM is still the biggest, but given all of these trends in data center, do you anticipate that over the next two to three years server and cloud memory is going to overtake mobile to be the biggest consumption segment for DRAM?
It’s too close to call. Definitely that gap will narrow and depending on whose forecast we believe and whose growth rate we tweak a little bit we are certainly within spinning distance of the markets being fairly equivalent and if any of those growth estimates are off by a tiny amount on a compounded basis, you could actually have that scenario. But right now it feels like they’re going to be neck and neck at minimally.
Any questions from the audience?
We have one over here, we have a microphone.
I don’t think so just we’re going to have to repeat the questions.
It’s okay.
[Question Inaudible]
So the question was around how much -- what our DRAM bit growth would be. We have actually talked about that.
[Question Inaudible]
Memory companies go all out all the time to be very honest with you. We do, we operate 7 by 24 and in any given limited time period. So your question was sort of limited to this year, you are bound by the productive capacity that you have, you couldn’t possibly order more equipment and get it installed then you are [ph] starting now. So those plans and that bit growth that we talked about which is at being at the lower end of the industry range, which was centered at 20% that’s pretty much what we’re going to get for our fiscal 2018 and that we’re making decisions now about how we think about DRAM bit growth from an outlook perspective in 2019.
[Question Inaudible]
You might get a point or two in the margin, but our fiscal and calendar offset by about a quarter.
Any other questions? The team is on track turning to the financials to hit the lower end of your deleveraging target, $8 billion of debt remaining exiting this fiscal year, net cash positive, you just generated $1.7 billion in free cash flow in the November quarter and that’s going to go up here in the February quarter, at least by our estimates. How are you thinking about uses of cash beyond fiscal year 2018? Does the team -- is the team motivated by potentially putting in place a sustainable capital return program as a part of that thought process?
Yes, so I want to be clear about this. We have a CEO who has experience in companies that have had both share repurchases and dividends. I personally am oriented around the fact that at the right time it’s probably reasonable for the company to think about some sort of sustainable shareholder return program. That's certainly not in 2018 and we are a long way away from the end of 2018. So it’s not rationale or reasonable for us right now to have a formulated and full opinion of exactly when that will be other than to say we are not -- it’s not in the short-term here.
And when we get to the end of that and we sort of talk about what our plans are for next year, we will certainly factor that into the decision making process and conversation, but we are very, very focused on executing and delivering on all those things. We are on track to do every single thing that we said and we’d like to get a couple of more quarters under our belt of making progress there before we then talk about what’s next.
But certainly in terms of the framework of our thinking, and by our, I’ll say the higher management team, we are absolutely thinking about the fact that it will be appropriate at the right time to include shareholder returns in the scheme of capital deployment for the company. Because we really do believe the business is becoming scaled enough and that our capabilities are strong enough to begin thinking about those things.
On the manufacturing footprint, team is keeping its wafer start profile flat in both DRAM and NAND this fiscal year relying on technology migration to drive bit supply growth. But even to keep capacity flat you do need to backfill lost capacity on technology migrations with some new capacity add so that the net number ends up being zero. The team has had a great approach to modularly adding capacity, but at some point -- in both NAND and DRAM, but at some point I would think that the team would need to consider building a new DRAM or a new NAND fab to continue to drive bit supply growth. Help us understand the considerations and potential timing of any new fab initiatives?
Sure, so we talked a little bit about this on the last call, where we said we are beginning to explore for NAND particularly, but the same is true for DRAM. What would best suit the company over the course of the next two to three years relative to, A, keeping its wafer capacity flat. And I agree with you, there is some additional equipment and essentially capabilities necessary to do that. But also with respect to two markets that look very different, one, which we believe has a CAGR of 40% to 45%, one of which we believe has a CAGR of 20%, how we want to provide the company flexibility to ensure that it can maintain its bid share.
We've talked on previous calls about the fact that we have no strong ambition to try and overtake a competitor. We have a very strong desire to at least maintain our bit share and maximize the value that we get for those bits. So we are approaching both within that framework. These are usually two year sort of cycle times from the time you announce something to the time that something is brought online in any meaningful amount of capacity. It's shorter than that from just the construction project. But that's sort of our framework. We did talk a little bit about it on call and I suspect we'll be talking about more about it specifically for NAND as we look forward to the call coming up in March.
The team is going to spend $7.5 million [ph] in CapEx this fiscal year. Obviously a big part of that is to drive your bit supply requirements, but also to continue to drive and close the gap the cost gap with major competitors. How should we think about the current cost performance versus Samsung, some of your other competitors in DRAM and against Samsung, Toshiba, Western Digital on the NAND front? And in addition to expenditures on CapEx, what are some of the other initiatives Micron is undertaking to further close the gap between its nearest competitors from a technology perspective?
Sure, so I joined the company about two and half years ago and I'm just providing that as a frame of reference here. At the time the company was just beginning to transition to 20-nanometer DRAM production whereas our competitors had been on that node for some time. And you would roughly size the gap there as two to two and half years. So we were about two to two and half years in terms of how we were implementing leading edge productive capability.
Today we are ramping 1X, we’ve given a bit crossover at the end of the calendar year. We are also actively developing subsequent nodes. And if you look at where our competitors are, they remain ahead but that gap has closed significantly. We still have a ways to go, but we have travelled a fairly good distance there and we have every intention of continuing that progress.
On the NAND front, the transition to 3D NAND was a really defining moment for the company. We were able to successfully produce economically at 32-layer that was not the case for broadest array of the industry. But certainly, everyone is norming at around 64-layer output. We have the smallest die size as a result of transistor into array [ph]. And that enables us to fight little bit above our weight in terms of the cost structure.
We believe that, that fundamental architecture is going to be very viable for the next several generations, certainly in third generation, fourth generation, but even well beyond that time we see no reason why that architecture is not sustainable. And we are also deploying CapEx to NAND to get that leading edge technology deployed into production, which is in turn helping us drive that level of bit growth that you saw last year, which was more than 60%. What we've articulated this year, which is above an industry that is approaching 50 and the corresponding cost reduction.
So we feel very, very good about our progress in both on a relative basis, I’d say we are a little closer to the competition in NAND than we might be in DRAM, but certainly closing the gap on both. And if you look at the margin performance of key industry participants, you can see that Micron's relative margin performance has improved. I would be remiss if I also didn't say that to complement this low cost output is our strategy around selling more high value bits.
So it's great to have cheap bits or inexpensive bits or industry leading cost bits, it's even better when you have those and you're putting them into products that customers value and are willing to pay a good margin for it like SSDs, like these embedded solutions in automotive. So both things are necessary for the company to fully realize its potential and to continue to drive improvements in our margin profile. And I think as we've demonstrated through the commentary we've made on the calls, we have very active programs around the execution of actually both of these paths.
Great. Well we're just about out of time, Ernie. Thank you for joining us today.
Thank you. My pleasure.
Thank you for your insights.
Yes.