Fab 5

Like it or not fab 5 is coming.

As Eli put it at Investor Day describing the need for new production capacity in the industry:

“You cannot grow from 10,000 petabytes [today’s industry capacity] to 60, or 70 or even 40,000 PB without new fabs. The question is not if, the question is when.”

For SanDisk and Toshiba this will mean a new fab 5.

Existing production space in fab 4 will be pretty much maxed out late 2010. Today SanDisk and Toshiba’s fab 3 + fab 4 deliver 35% of the world’s NAND output. To hold their own, fab 5 is a necessity for both SanDisk and Toshiba.

As is usually the case, the devil- and the promise- is in the details. This post is a first pass at what fab 5 means.

Mega-Fab

Fab 5 will be a mega-fab. It will be a beast. The slide below is from this year’s Investor Day.

Fab 5 will be expensive at $8 billion ± . For perspective that’s about $3 billion more than aircraft carrier.

At full capacity fab 5 will be able to produce 200,000 300 mm wafers/month.

It will need to support EUV and it will need to be able to adapt to post-NAND technologies.

From SanDisk’s perspective the post-NAND technology will be (hopefully) SanDisk’s 3D R/W which was discussed in my last post.

SanDisk expects that this transition to post-NAND will be more complex than the transition from SLC to MLC which took 3 years- 2002- 2005. Fab 5 will be expected to accommodate both NAND and post-NAND designs- at the same time- for an extended period.

Why a mega-fab and not something smaller?

According to Eli, its all about economies of scale. From Investors Day:

“Eli: You have to build a mega-fab because otherwise you are not going to make much difference. [laughs] First of all, but more importantly, you are not going to be competitive.

Small fabs don’t cut it. OK, its really a mega-fab or stay at home.

That’s the magnitude. But that by itself has an implication on when, because if you could do it in tiny little increments, or something [you can’t do it in small pieces]. Now I would say that those mega-fabs don’t necessarily ramp overnight. This is not like a 6 month exercise.

A mega-fab could easily take two to three years to ramp- and the investments would be spread accordingly.”

Industry Supply

The slide below from this year’s Investor Day is not only provocative, but instructive.

Heretofore, on seeing this slide, I have been focussed on the markets for flash and their hockey-stick ramp- year by year. This time around- the left hand scale: millions of gigabytes (GB)- jumped out at me.

1,000,000 GBs= 1 petabyte (PB) = 1000 terabytes (TB).

PB is often used in describing annual capacity of the flash industry. Today’s supply and demand is typically pegged at about 10 to 11,000 PB. Looking at the slide, sure enough- 10 to 11,000± PB for early 2010.

The graph, based on data from Gartner, shows supply and demand growing to 75,000± PB by 2013.

Wishful thinking.

To my mind, supply is going to come up far short.

Eli ran through the basic math in response to a question at Investor Day (included below). Today’s 10 to 11,000± PB will likely be able to grow to 30,000 PB in existing fabs primarily on the strength of shrinks to 1x.

It will take an additional 7+ new mega-fabs- fully ramped and running at full capacity to add another 45,000 PB by 2013.

My guess is that the equivalent of only three new mega-fabs will be up and running by that date. Fab 5 (meg-fab) for Sandisk/Toshiba. A mega-fab for Samsung. A regular fab (Singapore) for Intel/Micron. And an equivalent regular fab for Hynix.

I just don’t see how these four players will be able to fund, build, and fully ramp more by 2013.

Not in this uncertain economic environment. Not with the looming technological uncertainties. Not with both of the big guys- Samsung and Toshiba/SanDisk currently preaching caution.

And not with the cost, complexity and time required for each new fab.

Just for kicks I went back and looked at the timeline for Fab 4 (a mega-fab).  Fab 4 was announced mid 2006. Construction started that same year. The building was finished a year later, in early fall 2007. Mass production began at the end of 2007. By the end of 2008 production had ramped to about half of full capacity. By the end of 2010+, Fab 4 will be at full capacity.

Total time for Fab 4 from announcement to full ramp= 4.5 years+. Unless SanDisk/Toshiba want to be working on two fabs at once, I don’t see how we’ll see more than Fab 5 by 2013. Heck and it hasn’t even been announced yet.

Other players are in the same boat.

If only the equivalent of 3 new mega-fabs are built, and the shrinks to 1x go as planned, it seems to me that by 2013, industry capacity will be far below the 75,000± PB shown in the Gartner slide above. 50,000± PB (max) looks more like it.

What this will mean is anyone’s guess. Likely some markets won’t experience the growth they otherwise would be able to.

In any case, those controlling supply would seem to be in the driver’s seat- demand willing.

Here is Eli’s math from Investor Day:

“Q: OK very good. And how many new fabs would need to be built then to reach your estimated demand in the out years [2013]? At one point a few years back you were talking about 6 new fabs.

Eli: OK, and don’t hold me to it. To do that math.

Lets say that 1x generation product is between lets say 2.5 and 3 TB per wafer, approximately that- lets say 3 TB per wafer. You have to be really really good to get 3 TB per wafer, and probably use 2 to 3 bits per cell.

That means 30,000 PB in 2013, an incremental 30,000 PB would require another 10 million wafers which is 5 mega-fabs, assuming 1x @ pretty good yields @ 3 bits per cell.

This is the point I was making earlier- that the projected growth in demand requires very substantial new capacity.

The existing capacity is already generating 10,000 PBs- 11,000. So lets say it can go to 30,000 PB. If you want to build another 30,000 you need to have another 5 or so fabs- mega-fabs.

Q: On an average of $8B per fab?

Eli: Roughly.

Q: Roughly, plus maintenance capex. That’s a big investment- a big risk.

Eli: Yes. Somebody told me that a nuclear aircraft carrier cost about $5B. That gives you an idea how big $8 billion is, but as I said, it’s spread over a three year period- investment wise. In any kind of new capacity it would be very likely with Toshiba so our number would be less than that number.

Q: I hope so.

Eli: Me too. [laughter]”

Previous Fab 5 Investment Model

Fab 5 has been a glimmer in SanDisk/ Toshiba’s eye for years now- at least since 2006. In early 2008, financing details were discussed at SanDisk’s Investor Day.

At least back then, SanDisk looked to have a lot of flexibility. The slide below is from Investor Day 2008.

As far as capital investment, Toshiba was willing to assume 75% of the deal. SanDisk would only have to cough up 25%. Sweetening the deal, SanDisk had the option to convert to 50% / 50%.

Capacity was to have been divided 50%/ 50% whereby SanDisk would receive 25% of captive capacity and 25% cost-plus committed foundry. SanDisk had the option to convert committed foundry to captive or to non-committed foundry.

Where things stand now is anyone’s guess. No one is talking. Last month Toshiba even released a press release to the effect that no decisions have been made for Fab 5. Eli has been doing his part in playing down the possibility of Fab 5- which is what the analysts have wanted to hear.

All that said, fab 5 is not a matter of if, but when. And when is getting closer.

My guess is that fab 5 will be announced this year. The cost will be closer to $9 billion once the dust clears. The location will be in Iwate in western Japan. The ramp will be similar to fab 4.

The building itself will be dedicated in 2011 and sometime in 2012 capacity will have ramped to about half of full capacity. EUV equipment will continue to be an issue. 3D R/W will be up and running by 2013, but only in small volumes.

My guess is that the fab 5 investment model from 2008 is still relevant.

SanDisk will at least hold its current market share, and total industry capacity will be closer to 40,000 PB than Gartner’s 75,000 PB by 2013 with demand exceeding supply.

Lots of guesses. In any case, the next few years should be interesting with plenty of twists and turns.

SanDisk and Toshiba have a lot riding on fab 5 and 3D R/W.

Advertisements

20 Responses to Fab 5

  1. joshua maddux says:

    Savo- Just a ‘thank you’ for your analysis and insights on SNDK and running a great site. I think your clarity and objectivity are what i appreciate most. I have been an investor/trader in SNDK for 7-8 years. The structure of your site/blog is great because it reveals a timeline of SNDK’s breakthroughs, success/failures and market dis-equilibrium. Because there are looming game changing technologies such as 3D/ Fab 5 etc… it is great to have a deeper understanding of today and tomorrows potential.

    Best Regards,
    Josh

  2. MorganBucks says:

    Right on the money ….

    UPDATE 1-Toshiba to build new NAND memory factory
    Mon Mar 22, 2010 9:57pm EDTStocks
    TOSHIBA CORPORATION
    6502.T
    ¥466
    +16.00+3.56%
    1:00am EDT
    Samsung Electronics Co Ltd
    005930.KS
    ₩809,000.00
    +8,000.00+1.00%
    1:00am EDT
    SanDisk Corporation
    SNDK.O
    $33.82
    +1.57+4.85%
    12:00am EDT
    *Says to resume a NAND plant construction project in July

    *Sees seeing growing demand for flash memory for smartphones (Recasts, adds background)

    TOKYO, March 23 (Reuters) – Toshiba Corp (6502.T) said on Tuesday it will resume plans to build a new NAND flash memory factory, a further sign of growing optimism in the sector about growth in chip demand.

    The world’s second-largest NAND flash memory maker behind Samsung Electronics (005930.KS) said the popularity of smartphones and other gadgets was behind rising demand for NAND memory.

    Rivals like Samsung and SanDisk Corp (SNDK.O) have also given bullish outlooks for the chip market, which had been battered amid the global economic downturn. [ID:nTOE62H04T] [ID:nN26219612]

    Toshiba said it would start the construction of its fifth plant in Mie, central Japan, where all of its NAND memory production is located but added that it has not decided on the scale of investment or the plant’s output capacity.

    The company had originally planned to start the construction of the plant in spring 2009 and for it to be completed this year but the project had been put on hold following the industry slump.

    • savolainen says:

      Hey MorganBucks,

      Must say I’m pleasantly surprised that the market seems to actually like this fab 5 announcement. I’ve been thinking that this would probably be a difficult day.

      I suppose Eli and Toshiba deserve a bit of credit for the orchestration.

      Quite clever to have Toshiba announce just the shell construction- with fit-out scale and tool-install schedule to be determined later- based on industry demand.

      Hard to argue with that.

      Especially when SanDisk doesn’t look obliged to ante up the building, which means no fab 5 dedicated SNDK capex for 2010.

      My guess is that just like the fab 4 shell, the fab 5 building itself was never going to be dependent on SanDisk $$ anyway.

      Nice spin.

      Personally I’m very happy that construction is starting sooner rather than later. This schedule will allow max. flexibility down the road. This announcement also telegraphs confidence with minimum risk.

      Now all SanDisk has to do is run at the max in 2010- as it looks like they will- and then when analysts ask where future capacity will come from, point to fab 5.

      All the while saying, they’re being cautious, while making big $$$$. Not a bad scenario.

      Regards,
      Savo

  3. savolainen says:

    from LCM SNDK report 25 March 2010:

    SanDisk (SNDK-BUY, $41 PT)

    Investment thesis. We continue to like SanDisk for the following reasons: 1) the company is well positioned in terms of cost-downs in 2010 with the transition to 32nm and 3bit and will likely see cost-downs of 30%-40%, better than ASP declines; 2) the stable NAND environment in 2010 is positive for SNDK’s vertical model; 3) we believe the company continues to gain share with its product variety and pricing flexibility; 4) the mix change toward OEM should provide longer-term stability and opportunity for new revenue streams. Our price target of $41 is based on 15x our 2010E EPS.

    Risks and valuation. The biggest risk is that NAND overcapacity will return. Our new $41 price target is based on 15x our 2010 EPS estimate.

    NAND

    OEM NAND demand remains very strong. Our channel checks suggest that demand for MLC NAND by OEMs remains robust; in some cases, demand is 50% more than available supply. Our checks suggest that May and June are completely sold out. Demand is especially strong from smart phones and from iPad look-alike products. We believe that Apple ordering in the spring for its expected iPhone refresh is to further tighten the market. We believe that demand is tight for both the card and the embedded OEM NAND market currently. Retail NAND demand is soft but pricing not falling; SNDK benefitting from private label. Overall, demand for cards remains fairly soft, especially for the low-density 1-2GB. As all 3-bit capacity goes to retail, this is somewhat negatively impacting pricing in the market. However, the size of the impact is fairly small due to the limited 3-bit mix. With regards to SNDK, we hear that many card players are purchasing from SNDK due to better pricing. This is a significant positive for the SNDK private label OEM business, which is ~10% of sales.

    Toshiba announces Fab 5. This week, Toshiba announced that it is planning to begin construction of Fab 5, which was put on hold due to the recession. We expect Toshiba to complete the Fab by end of 2010 and have tools installed by the 2H11 for production. While Toshiba has yet to lay out a plan on the production output, we believe that it most likely will be initially ~50K wafers. We are not concerned about oversupply in 2010; however, increasing plans for production increases in the industry next year lead us to believe that 2011 could be an oversupplied year.

    The tablet market — the next biggest demand driver in 2010. We could see a significant portion of demand coming out of the tablets that are slated to be released this year from both Apple and HP/Microsoft. While some of our checks suggest that some of these devices will have card slots, we anticipate that embedded memory will be a big driver for the industry. All eyes continue to be on Apple and whether or not it will need to pull more NAND from the industry. We believe SNDK will benefit from any upside at Apple on the iPhone or iPad side, as this would further tighten the market for NAND and benefit the SNDK vertical business model. Raising 2Q estimates on SNDK. With recent data points suggesting a better NAND market for 2Q than we previously expected, we are raising our estimates for 2Q and 2010. Our new estimates are $1.04B/$0.61 and $4.48B/$2.71, compared to $1.01B/$0.58 and $4.39B/$2.57 previously.

  4. bob77977 says:

    hi savo,
    “today’s 10 to 11,000 PB will likely be able to grow to 30,000 PB in existing fabs primarily on the strength of shrinks to 1x”.
    keep in mind that if 3D succeeds, production will be much higher – assuming 1x production to be 50% X2 and 50% X3 then sandisk/toshiba’s part of 10.500 PB will grow to 33,600 PB in 2013 and the total production reaches 53,000 PB (8/2.5 multiplied by 35% of 30,000 PB.only 28,000 PB if all production is X3).
    true, 3D’s production is not going to prevail yet but it has to be taken into consideration lest oversupply occurres shortly after.
    I think this should confront sumsung with a big dilema. they could invest in a mega fub which could be worthless in 2015 or before. that grants a lot of power to sandisk over them.
    the question is will sandisk license 3D to sumsung ?
    what is your opinion ?

    • savolainen says:

      Hi bob779977,

      Yes, I am guessing that SanDisk will license 3D to Samsung- but it won’t come cheap.

      Samsung will have to pay more than they want to.

      It is in SanDisk’s interests to have the industry follow its chosen path and collect the big licensing/ royalty $$$$.

      That said, SanDisk still needs to prove that 3D can be manufactured and manufactured efficiently and cost effectively. Often these things take longer than one might think.

      EUV is no walk in the park.

      Therefore I think Samsung will go ahead with a mega-fab- but only one- until the dust clears. Basically I agree with you that the uncertainty over the end of NAND as we know it will slow NAND capex down- industrywide.

      Also, I based my comments on bit growth potential in existing fabs on Eli’s comments.

      Eli said he had already accounted for the potential of X3. He was making big simple calculations to make the point that the projected growth in demand requires very substantial new capacity.

      I am going to be away this weekend so my next post will postponed at least a week. My tentative plan is to make another pass at SSDs.

      Best,
      Savo

  5. bob77977 says:

    thanks savo,
    sandisk will make the $$$ anyway. either by collecting heavy license fee or through fat margins.
    the second way allows sandisk a better control over samsung. they will beware of overproduction and try not to make sandisk angry. I must admit I like it better.

  6. bob77977 says:

    actually, things depend on price elasticity at the future point of time when 3D will be manufactured. higher price elasticity will justify licensing whereas low price elasticity will not justify it.

  7. bob77977 says:

    hi savo,
    appearance of memristore definitely changes the picture.
    if 3D R/W turns out to be more efficient than HP’s memristore then licensing will be the way to go. otherwise
    HP will license sumsung and the rest.
    what are sandisk’s chances to win this race against HP in your opinion?
    somehow 3D looks to me more reliable then NAND, can you please relate to that. is the usage of X4 will be easier?
    making it 10-11 instead of 8?

  8. jexzz says:

    Hey, Savo– I’m sure you’ve read Dan’s post on the potential INTC takeover of our favorite NAND player. I think the most significant fact that speaks to this rumor is Toshiba’s silence on SNDK’s investment in FAB 5 or, rather, Toshiba’s declaration that SNDK won’t participate. Why would they if they were about to be bought? Anyhoo, I was wondering if you had any thoughts on this? Dan was the only one who named tha “big new product” before the announcement (32gb). Smart guy, as you know. Your thoughts?

    Hope all is well, and thanks in advance.

    jexzz

    • savolainen says:

      Hey jexzz,

      Personally I don’t read too much into the silence on SNDK’s investment in fab 5. Seems pretty much business as usual. Toshiba typically is responsible for the fab shell with SanDisk’s share applied to tools.

      I interpret no announcement of SNDK’s involvement as simply good expectations management- SNDK doesn’t have to say anything and therefore they hold off as long as possible.

      In this scenario SNDK gets to play the prudent one- meanwhile racking up the $$$ this year- so that when the fab 5 announcement is made, everyone understands the need and SNDK has that much more cash on the books.

      Also, if INTC is waiting in the wings, I would think they would want in on fab 5 and Toshiba as a partner- and therefore would be most supportive of fab 5. Cutting edge capacity could be on the scarce side in the next few years and Toshiba would make a much better partner for INTC that MU.

      In the really big picture, I am dubious about INTC. Not because the match wouldn’t be right, but because INTC seems somewhat retarded when it comes to memory. They hung onto NOR far too long, and then teamed up with MU.

      Maybe INTC has changed its stripes, but I’m not counting on it.

      IMFT hasn’t been either a happy or profitable experience. My guess is INTC wants to cut its losses on the memory front, but hasn’t been able to execute on an exit strategy.

      That said, if INTC is interested in SNDK at a big premium, it would be because of 3D R/W.

      Maybe INTC knows far more than we do, or more than SNDK is letting on, but I doubt it.

      Best,
      Savo

  9. jexzz says:

    Savo – thanks. Hope you have a great weekend. Yout comments always appreciated.

  10. Sandiskrocks says:

    Savo,

    Good day. I saw your response to Jexzz. You say that Intel wants to exit the memory business, yet they say that SSD is a strategic growth engine for their future. They had also commented last Tuesday that they were accumulating cash for a large strategic investment. Both of these things were said, and therefore, how would they not stay in the memory business to grow SSD’s? Nand will be suitable for true SSD growth, or in taking over hard disk, only 3D. Intel knows this, and do you really think they’re going to sit back and pay SanDisk all of the royalties that will come of this? In my opinion, I think not. They don’t even have a current licensing agreement with SanDisk as it expired this year already. If you don’t think that they will buy SanDisk, how do you think they’ll be a top player in the SSD market then? As you said, and as we all knew, they are exiting their partnership with Micron, so they must be looking at something else. The bottom line is they said they’re going to buy somebody, and it’s not going to be STEC by the way, that’s not a “large strategic investment”. Your comments appreciated.

  11. Sandiskrocks says:

    Sorry, I meant Nand “won’t be” suitable for true SSD growth in taking over hard disk drives because of cost, and process limitations.

  12. sam says:

    Savo, INTC is retarded on memory, although they did (finally) get out of DRAM, and they did (finally) get out of NOR (sort of). But that is a reason for buying Sandisk–give Eli a seat on the board and put him in charge of the Sandisk division, giving them someone who understands the sector. Nothing really changes except that Sandisk stockholders become Intel stockholders (I am assuming that this would be a stock deal, at least mostly so, as Intel will need their cash for new fabs). And Eli gets to be an important leader of a major international company, even if it isn’t exactly his own company. If Intel is serious about staying in flash, they would be foolish not to buy Sandisk, assuming that Eli and crew are willing sellers and would stick around for awhile.

    But when the Numonyx deal closes, Intel will be a big stockholder of MU–I hope that they sell those shares before any deal with Sandisk.

    • savolainen says:

      Good Morning Sam and Sandiskrocks,

      Don’t get me wrong. I think it makes a lot of sense for INTC to buy SNDK. For all the reasons you mention.

      I just am not so sure INTC sees it that way. They seem to have made a lot of dubious decisions over the years- particularly when it comes to the memory business.

      INTC has not been interested in SNDK for a long time- when they should have been.

      INTC could have scooped SNDK long ago when SNDK was cheap. Say in 1998 when SNDK’s market cap was only $400M or 2002 when SNDK was valued at around $1 billion.

      Even last year when Samsung made its move. INTC could probably have had SNDK for about where SNDK sits today.

      Maybe INTC has just been waiting until SNDK gets back to all time highs before putting an offer on the table. Seems kind of backwards to me, but stranger things happened.

      ****

      The cross license agreement situation is interesting. It never occurred to me that there could be an upside to the expiration of some of SanDisk’s early (and fundamental) patents, but there appears to be.

      As SNDKs early patents expire so does the INTC/SNDK cross-license agreement.

      The INTC/SNDK cross-license agreement was signed in October 1995 and extended for the life of the patents. As the patents expire, so does the license to use them.

      Clearly INTC won’t need to license the expired SNDK patents, but at the same time INTC would seem to have lost any implied access to related subsequent patents- and therein may lie the problem for INTC.

      Samsung recently faced pretty much the same situation. As I recall Samsung toyed with the idea, that they didn’t need to license again, because certain SNDK patents were expiring. On further reflection Samsung apparently decided all those other SNDK patents were worth a lot after all.

      it will be interesting to see what INTC decides.

      Regards,
      Savo

  13. Sandiskrocks says:

    Savo, thanks for your reply. In my mind at least, it really comes down to what Intel said in their conference call, which was they were going to make a large strategice investment and had to accumulate cash, which has to be an acquisition. They would surely announce a fab, or say they were positioning for that if it was the case. Secondly, as I mentioned, they also said SSD’s were a strategic future growth driver for them. Because of these two things, it seems clear to me, they must make an acquisition that makes sense. The only one is SanDisk, period. 3D is worth a lot more to Intel now, then MLC would have been way back when Savo, by far, and if they don’t get it now, there is no point to future SSD’s for them unless they want to license 3D from SanDisk at a mightly royalty rate because you know it’s going to be better than the old MLC royalty rates as well, or at least my guess

  14. Sandiskrocks says:

    bob77977, yes, 3D is the answer to replacing all of nand eventually, and yes, it will also be used in the same products nand is today. Let’s just say that 3D doesn’t go into production until 2014. At that point, it does not make any sense for Nand players to invest heavily into equipment for nand. It just doesn’t make any sense, because nand will be done by the late turn of this decade anyways. Limitations on process technology will become too difficult, and 3D will be that much cheaper of a solution.

    • bob77977 says:

      sandiskrocks,
      the point is not that 3D will be the most efficient nonvolatile memory in market. intel might be after it as there might be a strategic change in how computation works. no more dedicated CPU, software residung on the SSD will do the job.
      again, a big if.
      if this is the case sandisk worth is a few tens of $B.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: