Micron Technology (MU)

Question 1

9.16.2015 - Tech Conference Webcast

Answer 1

• How does end market demand look?
o 2H 2014 they saw nice uptick in PC demand driven by corporate refresh
o This year, half over half looks better but compared to last year it looks a little weaker mainly related to PC
o Enterprise and cloud server segments look consistent
o Supply side doesn’t look much different, most of the disconnect in the market is demand related
o Think inventory on DRAM side has gone lower so that’s starting to help as well
• Typical linearity in your business
o 60% of demand is 2H 2015 – q3 AND q4 IS the strongest demand
o Back to school build late in the summer
o Christmas build in the fall time frame
o Calendar Q4 to see a pretty good increase in overall demand sequentially, anything with a consumer end market
• Growth drivers that get you excited
o DRAM - PC should be better, but don’t expect anything spectacular – LT view is pretty flat unit growth
 Builds of PCs has been off substantially more than that
 Inventory replenishment from PC build side of things
 Directionally looks better
 Mobile – pretty consistent, bit demand is 50% in CY2015, DRAM per phone growing dramatically, lots of phones doubling
 Server – growing 40% in aggregate, consistent demand
o NAND
 Key markets are still SSD and Mobile (includes tablets)
 Both are growing in 40-50% range YoY, above average
 Legacy markets, card and storage to grow significantly below the average
o Bit growth and cost down
 Above market bit growth in DRAM, not dramatically, but a lot more than this year – bit growth is where you get cost leverage
 Inotera agreement shifts next year - Market minus structure to margin sharing next year even with fairly conservative assumptions on pricing
• DRAM – overarching theme is how can 2016 be better than 2015
o Benefits of consolidating market is supposed to be better pricing yet DRAM prices have gone down 40-50% this year
o Industry is structurally different now
o DRAM ASP reduction has been focused on one area, not broad based
o Bad market in PC environment, one of the worst year in recent memory
o Prior DRAM cycle created cash burns and negative gross margins
• Signs of stabilization
o Positive turn in the spot market and in the channel with real volume too, buyers paying more today than they were 2-3 weeks ago
o Contract price is basically flat month over month and that’s the first time that’s happened all year
• Bit growth should be above market growth – can you talk about the roadmap
o 20 nanometer ramp going ahead of schedule in terms of yields
o Two FAP – mobile DRAM and server DRAM moving aggressively towards 20 nanometer
o Big thing is that they have opportunity
o At a point where your investing in the technology after all these acquisitions
o 1x process node
• Industry supply outlook for next year
o Samsung and Hynix
o Think industry supply growth is at a lower rate than this year, you’re getting past some of the big conversion cycles for Samsung on 20 nanometer
o If you add it all together, low to mid 20s supply growth in 2016 – don’t expect supply growth to accelerate
 Assume they add wafers in those fabs
• When you look out to 2020, how much 3D NAND will be greenfield vs. conversions
o Majority of bits to be converted to 3D NAND by end of CY2016
o Focused on getting converted to 3D now
o First year will mostly be conversions, 2 fabs in Singapore which are commencing conversion now – next year to year and half is primarily conversions but CY2017 you get increase in greenfield capacity
o 2/3 conversion 1/3 incremental
• How much bit growth will come from conversion of 3D?
o The initial bit conversion of 3D, it might be a downside to supply over the next few quarters, you don’t get immediate benefit of conversions, wafers declining, a couple quarters until you get benefit of bit growth
o Mid 30s plus or minus vs. this year which is a little higher
o Don’t see an acceleration in the industry supply – what will all the players do in terms of capacity in the next 3 years, that’s the big question
o NAND market is more elastic
o Next few years, think NAND industry might be supply constrained – how much TLC vs. MLC is converted to 3D
• NAND margins stabilized, but much lower than your competitors - what other factors should we think about in terms of driving margins higher for this business?
 Better end market mix, doing that every quarter, reducing sales into the channel, down 50% from a couple quarters ago
 Less volatility overall
 Lack of TLC in the mix, not lack of technology, its lack of a product portfolio
• 3D Crosspoint
o How do you position this? What applications will be addressed first?
 In terms of bits, it could be half of DRAM in a few years
 Think the first couple of years of adoption will really be an incremental market – markets that don’t exist for DRAM or NAND today
• Capex and FCF conversion
o Long run capex and FCF – accurately MU is now in a phase now where gross capex is trending up, but net capex is lower
 Includes one time investments, shell investment of $1bn
 $4bn capex is more normalized number
 Will generate significant returns on that investment going forward
• Customer perspective from 3D crosspoint
o Customers will be able to dual source
• Crosspoint included in the non-volatile memory capex this year, it will be a small piece of that