For a while now, it’s been looking like Intel has wanted out of the NAND fab business. The question wasn’t if, it was when, and how.
To date, Intel’s NAND JV with Micron hasn’t been either a happy or profitable experience. My guess is Intel wants to cut its losses on the memory front, but hasn’t been able to execute on an exit strategy.
After looking over Intel’s 2 November 2010 10Q, I’ve got a working hypothesis for what Intel’s 5 year exit strategy might be- hunker down and “Pass the Buck.”
I don’t think this was option #1. Option #1 was sell out yesterday. The insurmountable problem was that no one was interested in taking on the challenges facing a tier two NAND player.
So what to do?
Why not pressure- entice- lure Micron into taking on- what other’s won’t buy outright?
It’s not like this hasn’t happened before. Exhibit A- Numonyx.
Before getting to that- a graphic and some background are in order.
Intel’s slow motion train wreck of a NAND strategy seems to be moving ahead one frame at a time.
I know I’ve posted this stuff before, but like good wine (or good beer) it’s worth revisiting.
In February of 2009, Intel announced a two-year, $7 billion investment plan for its U.S.-based manufacturing facilities.
This would have been a great time for Intel to step up and show commitment to its NAND strategy.
The numbers told the story.
Not a dollar of those billions was going to NAND.
It seemed pretty obvious that Intel decided not to throw good money after bad.
In March 2008, Intel’s NAND problems became public when it disclosed that its money-losing NAND business would knock two points off its gross margins. Analysts were stunned given that Intel’s flash business generated such a small portion of Intel’s overall revenue.
Three classic articles from that time. The titles say it all:
10 March 2008 Why Intel Should Dump Its Flash-Memory Business
05 March 2008 Sell More NAND, Watch Margins Drop
Consensus- Intel would be far better off without this “rotten” business.
Intel’s CEO, Paul Otellini, vowed to “fix” its NAND business. He promised that NAND wouldn’t “be a drag on Intel Corp.”
As the articles above indicate, Intel has probably been shopping its NAND business for a while now, but has found no takers. So what’s option #2?
Hunker down and edge out.
2016 will be here before we know it.
Intel has JV NAND commitments with Micron which run through 2016.
Intel is focused on SSDs, but that doesn’t mean it necessarily has to produce its own NAND or next generation solid state storage. Intel has said as much.
A while back, Eli summarized the challenges that the Intel/ Micron NAND JV faces- “It is tough to be number 4 in a rough neighborhood.”
The Two Tiers of NAND
The big problem for Intel/Micron is that they are a distant #4 and a tier two player.
Tier one is Samsung and Toshiba/Sandisk. These three ship 80% of the world’s output of NAND.
Tier two is Intel/Micron and Hynix. These two divide up the remaining 20%.
The chances of Intel/Micron significantly closing the gap are slim to none.
The name of the game is the mega fab at $8 billion a pop. Intel/Micron don’t have any, or any on the drawing board.
At full capacity, the mega fab is able to produce 200,000± 300 mm wafers/month. They are highly productive and extremely efficient, but to build one takes serious capex.
The NAND capex spending by company graphic below is from the Bank of America/ Merrill Lynch (BofAML) Global DRAM/NAND report of 27 August 2010.
In 2010 Samsung’s capex is expected to be $3,326 million. Toshiba/SanDisk is expected to spend $2,691 million.
Micron/Intel is expected to invest $543 million. Hynix’s capex is expected to be $756 million.
Clearly tier two isn’t investing the bucks. They’ll reap what they sow.
The NAND supply model graphic below is also from the BofAML report of 27 August 2010. Lots of numbers and lots of interesting ways to slice the NAND pie.
Again no matter how you cut it, Intel/Micron and Hynix come up as tier two players.
The relationship between share by wafer capacity and share by bit shipments is both interesting and revealing.
Share by wafer capacity 4Q10E shows Samsung 40%; Toshiba/ SanDisk at 34%; Micron/ Intel 14%; and Hynix 11%.
Share by bit shipments 4Q10E shows Toshiba/ SanDisk at 42%; Samsung 32%; Micron/ Intel 14%; and Hynix 7%.
Suffice it to say that X3 is becoming increasingly important.
It’s rather interesting and encouraging that Toshiba/SanDisk is expected to ship close to 30% more bits than than Samsung this year.
Once upon a time, not that long ago- in early 2008, Intel was in the NOR flash business and wanted out. The solution was to spin off its NOR business and combine it with STM’s as Numonyx.
In 2008, Intel divested its NOR flash memory business in exchange for a 45.1% ownership interest in Numonyx.
At the end of 2009, Intel’s investment balance in Numonyx was $453 million.
The problem for both Numonyx (and Intel) was, according to EETimes, that Numonyx was “faced with a declining product line in NOR and a bloated overhead, Numonyx could not have survived on its own over time.”
In May of this year, Micron acquired Numonyx for $1.2 billion in stock.
Problem solved- at least for Intel.
In essence Intel had sold its ownership interest in Numonyx to Micron.
In exchange, Intel received 57.9 million shares of Micron common stock, with an additional 8.6 million shares held in escrow. At the end of September 2010, these shares were worth $479 million.
Intel also passed off its Numonyx credit obligations to Micron:
From Intel’s 10-Q 11/2/2010:
“In 2008, Numonyx entered into an unsecured, four-year senior credit facility of up to $550 million, consisting of a $450 million term loan and a $100 million revolving loan. Intel and STMicroelectronics had each provided the lenders with a guarantee of 50% of the payment obligations of Numonyx under the senior credit facility. The Numonyx senior credit facility that was supported by our guarantee was repaid in connection with the closing of Micron’s acquisition of Numonyx.”
So why would Micron be interested in Numonyx?
EETimes’ speculations from February of this year:
“Still, faced with a declining product line in NOR and a bloated overhead, Numonyx could not have survived on its own over time. So in other words, the Numonyx deal could backfire and break fragile Micron. Indeed, the U.S. memory maker has recently reported its first profit in three years and the company is still reeling from the nasty downturn.
The big question is why would Micron buy a company that primarily sells a declining and bland product like NOR? First, perhaps Micron wants to extend its reach in the handset with NOR. Second, maybe Micron wants to gain access to Numonyx’ next-generation PCM line, a technology that has generated interest from cell-phone giant Nokia Inc.
Or thirdly, perhaps Intel Corp. brokered this deal and Micron reluctantly agreed to appease its quiet partner. Intel is not only seeking an exit strategy with Numonyx, but it quietly wants to expand its leverage and influence inside Micron, observers speculated.”
I’m partial to #3:
Intel brokered the deal behind the scenes. It both took care of its Numonyx obligations and expanded its leverage and influence within Micron.
So what’s Intel going to do with this extra leverage and influence?
My guess is that Intel is going to use it to pass the buck to Micron once again.
While its hard to believe today, apparently in 2007, Intel and Micron were planning on opening a 300mm NAND wafer fab every year. Next up after Lehi was to be Singapore.
Today, Intel appears to be having second thoughts on Singapore.
IM Flash Singapore, LLP (IMFS) was formed in February 2007- the second JV between Intel and Micron. So far at least, this has not been a happy story for Intel, which early on dropped a few 100 million dollars into the building structure.
Ground was broken in Singapore in 2007 on schedule for an opening in 2008.
Today the fab facility structure is substantially complete, but only Micron has committed to the tooling and production ramp.
Intel is dragging its feet. Intel has said a decision will be announced before year end on its Singapore plans. The clock is ticking.
So what could all this mean for SanDisk shareholders?
First and foremost, it means that at a minimum deep-pocketed Intel isn’t going to be pushing the NAND supply/ demand balance out of whack any time soon.
This is not lost on Bank of America/ Merrill Lynch. From its 19 November 2010 report:
“Micron’s new Singapore fab (JV with Intel) will not feature NAND on a large scale (only 60,000wpm, even in 1Q12) due to Intel’s limited capital injection. The chances of NAND shortages in 2011-12 are likely to be high.”
From my perspective, regardless of whether Intel antes up for Singapore, Intel has the brakes on. This is good news.
Where this goes ultimately is anyone’s guess.
As noted in earlier posts, since Intel isn’t leaning into NAND fab production, there seems little to no chance of an Intel buyout of SanDisk.
If and when SanDisk winks out, my bet is on Toshiba.
I suppose there is the potential for SanDisk to be a NAND supplier to Intel, sometime down the road- but for the time being Intel seems to have it’s NAND supply needs covered in house.
There is also the IP angle to keep an eye on. Both for Micron and Intel.
Some time ago, Eli said that Intel has a cross-license with SanDisk, but that Micron doesn’t. It appears that Intel has somehow been protecting Micron.
Less Intel should translate into less cover for Micron within their JVs.
I’m not so sure about the extent of coverage provided by the Intel/SanDisk cross-license agreement either.
The INTC/SNDK cross-license agreement was signed in October 1995 and extended for the life of the patents- issued at that time.
As I understand it, the legal life of a patent is 17 years. 1995 was 15 years ago.
Sometime in 2012 it would appear that the INTC/SNDK cross-license agreement will expire along with the last of the patents.
It’s also not clear, at least to me what claim Intel might have on patents issued to SanDisk subsequent to the 1995 cross-license agreement.
Therein may lie a big problem for INTC/ MU for X3 or TLC.