MU - Micron Technology Inc.

[K2SO]

I am looking at Micron, and it looks like a good little performer trading at a mid 6 PE. 

Beware investing in cyclicals like this on the basis of P/E. Counterintuitively, the time to buy these is usually when the P/E is high. You really need to understand the industry supply/demand dynamics.

[DTEJD1997]

I am looking at Micron, and it looks like a good little performer trading at a mid 6 PE. 

Beware investing in cyclicals like this on the basis of P/E. Counterintuitively, the time to buy these is usually when the P/E is high. You really need to understand the industry supply/demand dynamics.

There is an argument to be made that, "This time it is different" with Micron.

20+ years ago, there were maybe 20 manufacturers of DRAM.  Now I believe there are 3.

Perhaps even more importantly than that, there appears to be an informal agreement among the manufacturers to not act irrationally on  memory pricing.

An example of this behavior is the fact that DRAM pricing for desktops is up about 300% in the last 16 months.  At the low point, I was able to buy a 16gb memory kit from Micron for about $50-$60.  Today, that price is about $175 or so.  Over time, the price of memory usually goes down as efficiency kicks in. 

In the near term, my contacts claim that the price of memory will be flat OR will maybe come down by $10 or so.

In the past, the price of memory almost always went down.  Every once in a blue moon, price might spike up on demand or a factory/supply problem.  In 20+ years of buying memory, I've never seen pricing like we have had in the past 1.5 years.

So maybe, "it is different this time"?

[walkie518]

it's different until a player increases production for purpose of simply generating more income

today there is an increasing number of database transaction activities done in memory rather than on disk, ssd, and hybrids ... while the business isn't going anywhere for the foreseeable future, there will likely be swings in value as a result of tweaking the supply chain

I owned Micron and sold in the $30s...I am waiting for the next leg of the cycle to buy more.

[walkie518]

it's different until a player increases production for purpose of simply generating more income

today there is an increasing number of database transaction activities done in memory rather than on disk, ssd, and hybrids ... while the business isn't going anywhere for the foreseeable future, there will likely be swings in value as a result of tweaking the supply chain

I owned Micron and sold in the $30s...I am waiting for the next leg of the cycle to buy more.

[DTEJD1997]

it's different until a player increases production for purpose of simply generating more income

today there is an increasing number of database transaction activities done in memory rather than on disk, ssd, and hybrids ... while the business isn't going anywhere for the foreseeable future, there will likely be swings in value as a result of tweaking the supply chain

I owned Micron and sold in the $30s...I am waiting for the next leg of the cycle to buy more.

I think there is a strong case to be made that "this time it is different" with MU.

I didn't think that way in the past, but I'm convinced now for TWO reasons.

#1

I've been working with/building/selling computers for over 20 years now.  Sometimes it is a hobby, sometimes it is my main source of income.  In all the time I've been working with them, I've NEVER seen memory prices do what they have done over the past 18 months.  NOT EVEN CLOSE.  It is simply unprecedented that DRAM prices would TRIPLE and stay that way for a good period of time.

#2

As time progresses, the number of memory manufacturers continues to get smaller & smaller.  At this point, there are only 3.  It is so much easier for 3 manufacturers to act "rationally" than it is for 7,10,15, or even 20.

Sure, the Chinese might get into this area, but the cost of setting up a foundry is hundreds of millions of $, perhaps even a billion.  That is going to keep a lot of competitors out.  If they do come in, they will likely act rationally.

I don't think we will ever get back to having 10 or a dozen different manufacturers.

Finally, if MU is making money at the clip they have been recently...that is going to change things real fast.  A stock simply won't trade for a P/E of 4 in a normalized market.

We will see, but real money is to be made when the normal way of doing business has radically changed.

[walkie518]

it's different until a player increases production for purpose of simply generating more income

today there is an increasing number of database transaction activities done in memory rather than on disk, ssd, and hybrids ... while the business isn't going anywhere for the foreseeable future, there will likely be swings in value as a result of tweaking the supply chain

I owned Micron and sold in the $30s...I am waiting for the next leg of the cycle to buy more.

I think there is a strong case to be made that "this time it is different" with MU.

I didn't think that way in the past, but I'm convinced now for TWO reasons.

#1

I've been working with/building/selling computers for over 20 years now.  Sometimes it is a hobby, sometimes it is my main source of income.  In all the time I've been working with them, I've NEVER seen memory prices do what they have done over the past 18 months.  NOT EVEN CLOSE.  It is simply unprecedented that DRAM prices would TRIPLE and stay that way for a good period of time.

#2

As time progresses, the number of memory manufacturers continues to get smaller & smaller.  At this point, there are only 3.  It is so much easier for 3 manufacturers to act "rationally" than it is for 7,10,15, or even 20.

Sure, the Chinese might get into this area, but the cost of setting up a foundry is hundreds of millions of $, perhaps even a billion.  That is going to keep a lot of competitors out.  If they do come in, they will likely act rationally.

I don't think we will ever get back to having 10 or a dozen different manufacturers.

Finally, if MU is making money at the clip they have been recently...that is going to change things real fast.  A stock simply won't trade for a P/E of 4 in a normalized market.

We will see, but real money is to be made when the normal way of doing business has radically changed.

I build rigs too. 

Not sure if this is the time/place to bring up DDR4 as incremental improvement to DDR3, but next gen memory will not be engineered like DDR3 and DDR4 anymore--there will be a bigger step up as planar went 3d for ICs.

That said, memory is in need and database transactions are much, much faster in memory than in any form of storage.  Maybe a hybrid product bridges the gap, maybe mainboards got he way of the dodo to better SoC designs.  Getting memory and other peripherals closer to processing is and has been the goal for decades, I don't see this changing.

I agree that the competitive landscape has changed, but I also think it's unlikely to remain as it is.  There is simply too much need and demand for product for there not to be continued price swings among the suppliers.  Some or all parties will need to make more to accommodate need. 

[Spekulatius]

Why is spending a few hundred or even a billion dollar considered a hurdle for Chinese? It’s pocket change, if the government decides they want some of their champions to go down that road.

[walkie518]

Why is spending a few hundred or even a billion dollar considered a hurdle for Chinese? It’s pocket change, if the government decides they want some of their champions to go down that road.

The Chinese need the IP to match Samsung and Micron, but how long will it take and how much will they need to invest?

[ritrading]

Stayed away from semiconductors in all my years of investing, but start researching last week after seeing MU at about 5 P/E. With respect to capital, semiconductor companies look like oil companies. There's an enormous upfront capital investment, and then the marginal cost of manufacturing is relatively minuscule, resulting in 24/7 operations to maximize usage of the equipment. Then, dumping product into the market as a price taker, with little ability to store inventory due to the the trend of declining prices and risk of being obsolete. Difference I see with oil companies is that semiconductor companies are more likely to need to upgrade equipment frequently due to new innovation.

My hesitations with MU right now are the enormous fluctuations with revenues, net income, and free cash flow. Is the 5 P/E even cheap? It kinda looks like the revs, NI, FCF are at the top of a cycle, and MU will turn very expensive once these peaks move to troughs.

[DTEJD1997]

Stayed away from semiconductors in all my years of investing, but start researching last week after seeing MU at about 5 P/E. With respect to capital, semiconductor companies look like oil companies. There's an enormous upfront capital investment, and then the marginal cost of manufacturing is relatively minuscule, resulting in 24/7 operations to maximize usage of the equipment. Then, dumping product into the market as a price taker, with little ability to store inventory due to the the trend of declining prices and risk of being obsolete. Difference I see with oil companies is that semiconductor companies are more likely to need to upgrade equipment frequently due to new innovation.

My hesitations with MU right now are the enormous fluctuations with revenues, net income, and free cash flow. Is the 5 P/E even cheap? It kinda looks like the revs, NI, FCF are at the top of a cycle, and MU will turn very expensive once these peaks move to troughs.

There is one problem with your assumption set...

In the past 18 months or so...MU has been a price SETTER, not a taker!

The price of DDR4 ram has approximately tripled...I've been working with computers/IT equipment for 25+ years and have NEVER seen this happen.   Memory pricing has been a downward slope with the odd spike up.  I've never seen price be this strong, for this long.

There are now 3 (or is it 4) memory producers.  It is MUCH easier for them to act as a cartel than in years past when there were a dozen or more producers of memory.

I would think that MU SHOULD trade a discount, but maybe not as much as "the street" thinks, as it appears that pricing on memory is much more rational than it has ever been.

Interesting times ahead and we'll see!