The closing panel discussion at Denali's MemCon Silicon Valley last Thursday (July 19), "Flash: Let a Hundred Flowers Bloom", was both aptly named

[More:]

In any market that is price-reducing 50% per year and growing volumes 150% to 200% per year, there's bound to be a lot of unanswered and unanswerable questions at any point in time as to where the whole juggernaut is heading, and what it means to all who inhabit places as far as the furthest reaches of the NAND food chain.

Such is NAND today.

Following the opening keynote, this year's MemCon event opened with a panel discussion whose charter was more focused on the future of NAND flash in the compute segment, namely Solid State Drives (SSDs) and Hybrid Hard Drives (HHDs)...namely, what is the outlook for NAND to replace HDD in laptops and when and how?

Then, after three days and 36 speeches and panels (over 20 of which dealt with NAND, flash memory, the NAND-enabling ecosystem and NAND applications), the closing Panel aimed at only a slightly larger topical NAND target: "Which applications were going to drive the NAND market in coming years, and what will that market evolution look like?"

The Cast of Panelists was a nice cross section of the analyst community, with end systems, focused on Consumer Products, represented by IdaRose Sylvester from IDC and Chris Crotty from iSuppli; mainstream memory market analysts Jim Handy of Objective Analysis; Alan Niebel from his own Webfeet Research, plus Lane Mason of Denali Software, doubling as Memory Market commentator and for Denali's own inside business planning support, plus Memory Market Consultant Jim Cantore from JLC Research.

Major issues:

Growth forward: The era of +200% GB growth for NAND flash is over...probably. So says Alan Niebel, arguably the most seasoned and detailed NAND flash market watcher. After a 6-year run up averaging more than 200% GB growth year after year since the late 1990s, Webfeet's GB growth forecast is looking at +130% for 2007 v. 2006, +100% for 2008 v. 2007, and steadily slowing to a steady state +80%, year over year "in the outer years"...this GB growth expectation, cast consistently in accordance with a like slowing of the price/GB and cost/GB reduction. Contrast this with Mr. Mason's assumption (position) that sustainable growth in the historic levels of +180 to 200% year after year ad infinitum, were probably unsustainable due to NAND flash's high capital costs and unlikelihood of sustaining the cost (price reduction) of –50% y/y very much longer. NAND CapExp is expected to make up 20-25% of the total semiconductor industry's CapExp budget for 2007, despite NAND's paltry 5-6% of total semiconductor IC sales...Clearly the money is NOT coming from NAND profits, which were nil at the end of the 1Q07 'price to the death' rush. The $10B CapExp for 2007 is a one-time investment bubble that is not sustainable, though it MIGHT have to be, to sustain the historical GB growth...something does not fit together neatly.

However unsustainable the historic rates might be, one cannot help but see a conflict of "realism/planner's droop"...which slows the 'unsustainable growth' in the outer years of the common 5-year forecast...in conflict with the unbridled NAND enthusiasm that says, "Why do we expect the market to stop growing fast NOW? We have lots of demand drivers, known and unknown, quantified exactly and merely guessed at, just waiting to be revealed and bloom forth like iPod did and the alphabet soups of card formats and DSC's did before them."

All of the existing NAND flash markets of today have proven to be remarkably price elastic, running roughshod over the question of who really needs to store 2000 songs and 10,000 photos, but settling on price points for purchase of $20-30 as 'consumer purchase tipping points", and so long as NAND prices decline, volume will build proportionally in existing applications AND next, more price sensitive applications will reveal themselves, as submerged rocks do when the shoreline tide goes out.

This issue is probably one where we all agree to disagree...for the time being...and we need to wait to see the results come, in year after year, to see if the bulls or bears are more correct in their assessments. As one questioner said, "The only thing we can be sure of is that all these forecasts are wrong"...except Jim Handy's, added parenthetically by Mr. Handy himself

iPod or iPhone: The eternal 'which platform will host which features" debate has risen to a higher level with the release of Apple's iPhone (Editor's note: This was before the tepid actual iPhone interconnect sales numbers became public earlier this week.). Is it an iPod with communications? Or an iPhone with lots of song-able memory? So far, the phone has tried to put everything on it's own platform, and succeeded to a degree. Camera? No sweat! Internet and Email access? Add more NAND flash and higher speed circuitry for downloads. GPS? That's a no-brainer...but not all consumers are following suit, and lively markets still exist for standalone single-function systems...(Now, here's a market for the flexible, fully user-customizable utility 'system,' with remote reconfiguration capability and and and...)

Who drives platform features? A corollary to this 'appropriate platform' question is the one, also avidly discussed on the panel, as to what was driving the addition of new features, Demand pulled? or Supplier Driven? Who needs 2GB for songs and a 4MPxl camera, when most camera-on-phone users do not use the camera? Just testing the waters and waiting and watching for market take-up? Likely...

Video as NAND market driver; high-frame-sequence DSC or "Digital Camcorder"? Though all current dissections of NAND end demand show video applications having only a small slice of the current pie, about 5-10% of GB shipments, there was near unanimity that video...in some form, in some application...constituted a potential growth upside for NAND. From the 'moving edge' of today's Digital Still Camera, which can shoot fast frame sequences, to bona fide Digital Solid State Camcorders, akin to but challenging tape- and HDD-based systems, comes a fast-moving market segment with an explosive growth potential. Anything that eats GB of storage in such potentially vast quantities is interesting to NAND flash makers. Salivation now, but when will it be fed? When will we witness the key "Apple iPod" event for video flash camcorders?

Security on NAND-based systems: Always a concern. And the security capabilities are available, no matter HOW much and what kind of security you need. Only problem: it seems is that few people use what is available. Another example of 'feature and utility overload'...until it is too late, and your X squirrels all the money from your bank account while you are otherwise not paying attention.

Market Characteristics, Land of a Thousand Niches (na, nananana, nanana, nanana nanana) Two relevant comments were added into this important issue from IdaRose Sylvester and Alan Niebel, on the valid observation that the NAND demand side is broad, multifaceted and rapidly evolving. First, when the issue of market cyclicality of the NAND forecasts and market issues came up (up-up-up-down with a crash), Alan pointed out that we may be analogizing too much with the DRAM business, with its 70%+ driver by PCs and associated computer equipment, and such a one-horse show is more susceptible to wild swings, when the PC market slips, weakens, or otherwise fails to live up to expectations (e.g., Vista ramp up is 'slow'), whereas the NAND flash market is firmly resting on many significant demand contributors—a variety of special-format cards, cell-phones, MP3 and PMP players, digital cameras, and smaller NAND installations in a variety of computer and consumer products for frame store and booting functions. Most of these markets for NAND today, also have a significant NAND growth upside, as phones become more high-end and add NAND-intensive features, as DSC's morph or are supplemented by video cameras (That's YOUTUBE talking), and PCs add SSD. The demand side is more diversified, and driven by many threads, some of which have yet to respond to today's prices and NAND capabilities, but have clear upside growth paths for increasing NAND demand far beyond what is shipped and used today. PCs, on the other hand, in addition to being a more monolithic mainstay of DRAM demand, have far more predictable MB/system growth paths...1.2GB grows to 1.3GB/system...grows to 2GB per system with the full Vista DRAM complement out in 2008 or 2009.

In their fastest growth phase, MP3 players, flash cards and DSC flash demand grew far faster than the PC-driven DRAM market, which by today's NAND growth standards looks positively pedestrian in its heyday 'explosive growth' phase.

Summary: OK, so not everything was resolved in the 1:20 of this MemCon closing panel session...in fact, nothing was TOTALLY resolved...but the arguments were moved forward and bets (opinions) were placed, and the discussion was moved to the next level. "Major uncertainties exist in the NAND market evolution through 2010 and 2012", is something everyone can seem to agree on.

We hope the discussion does not stop here, and we urge readers of this article to take a look at the webcast on the Denali website, and wade in with comment, criticism, opinion, assertion, your personal prejudice, hard indisputable facts. Add fuel to the NAND fire.