So far, so good. SanDisk’s Q1 was a substantial improvement over Q4. In late Q4, NAND pricing started to improve and then improved some more. Call it benign pricing, stable pricing, favorable pricing or whatever, it’s a welcome relief.
On top of benign pricing, demand for SanDisk products in Q1 was stronger than expected. Remarkable actually, given the current sorry state of the consumer market.
Where we go from here is anybody’s guess. The next real test will be the second half when idled capacity comes back on line.
Bulls predict demand will soak up this additional supply. Bears prognosticate otherwise.
Personally, I side with the optimists. Demand is the key. To my mind it all comes back to Mobile where NAND rules. In this high growth arena, there is no substitute for NAND.
And the brakes are on for all the NAND suppliers. The credit squeeze coupled with the worldwide economic meltdown had to be a rude awakening.
The days of easy CapEx money appear over. The nonsense days of a billion dollars invested for a billion dollars of revenue are a thing of the past. NAND manufacturers are going to have to show they can make money to get money- SanDisk included.
The last three years have been most painful for the NAND manufacturers. Billions of dollars invested and pretty much nothing to show for it- come 2009. While there is plenty of blame to go around, one company, and one man, in particular seems responsible for this sorry state.
The company, of course, is Samsung. The man is Hwang Chang-gyu, the former head of Samsung’s chip unit.
Hwang’s Train Wreck
At this last January’s Industry Strategy Symposium held in Half Moon Bay California, G. Dan Hutcheson, CEO of VLSI, nailed it with the train wreck metaphor:
“Flash has been the fastest growth market in the history of semiconductors, ”but it’s been squandered,” he [Dan Hutcheson] said. ”What about Hwang’s train wreck?”
He was referring to Hwang Chang-gyu, formerly the head of Samsung’s chip unit, who devised something called ”Hwang’s Law.” The axiom implied that NAND transistor count would double every year in leading-edge devices, thereby driving down prices by some 40 percent a year as means as to enable new applications.
The idea worked–and backfired. Prices for NAND have fallen like a rock, enabling new applications like MP3 players, solid-state drives and others. But too many vendors followed the same path, creating excess supply and falling average selling prices (ASPs).
Now, NAND vendors are losing money. Even Samsung is reeling due to the glut and Hwang has apparently paid the price. In recent times, Hwang was demoted to chief technology officer and president of the corporate technology unit. Last week, he apparently left Samsung, according to reports.”
Hwang is gone, and NAND pricing has improved significantly. Call it a coincidence if you like, personally I don’t.
I suppose it could be that Hwang was just the fall guy and responsibility lies elsewhere organizationally. It really doesn’t make any difference, if, as it appears, Samsung has changed its tune regarding pricing.
Of course pricing doesn’t happen in a vacuum. Supply and demand need to be in better balance.
In 2008 industry-wide NAND supply (bit output) grew a stunning 130% year-over-year. And this was down from 2007/2006’s 160% growth. In each of the previous three years bit output grew over 200% year-over-year.
This year the brakes are finally on. Market research firms are forecasting only 50% to 60% supply growth- and this is with 100% capacity utilization in the second half of 2009.
In Q1, industry-wide NAND production cutbacks got serious. Desperate measures for desperate times. Supply came into better balance with demand, and bingo- NAND component pricing stabilized.
We shall see whether this a lesson learned. 2009 at least looks safe. As Eli said in the Q1 conference call:
“So I think it’s very encouraging and I believe that really if somebody decided today, if the biggest supplier in this market [Samsung] decided today that they are going to pull all the stops…, it’s still not going to make much of a difference in 2009 and I think we are really talking about, I think a period of return to much better health.”
Along with the rest of the gang, SanDisk has cut back. The SanDisk/Toshiba NAND fabs have dialed back 30%± for Q1 and Q2 and put additional wafer starts on hold for 2009.
Of the major NAND players, SanDisk is in a most enviable position, thanks, in curious twist of fate, to Samsung’s unconsummated hostile takeover bid of last fall.
In a gesture of support for SanDisk, Toshiba agreed to restructure the two company’s joint ventures. Toshiba took over a bit more than 20% of SanDisk’s captive supply responsibilities, allowing SanDisk to more effectively balance profitability and volume.
In another curious twist, SanDisk’s excessive inventory is now generating money. How quickly the table turns. December’s curse becomes May’s promise.
SanDisk took a $184M inventory hit in Q4 2008. Expectations were that prices would continue to decline for the next couple of quarters. Inventory which cost $784M to produce was only deemed to be worth $598M in those dark days of December.
Thanks to the magic of strengthening pricing, inventory is becoming more valuable as time is passing in 2009. In Q1, inventory netted SanDisk a $34M profit. So far, Q2 looks every bit as good, maybe better.
Q2 Gross Margins appear they’re going to get a booster shot- thanks to inventory. Such income drops straight to the bottom line.
Accounting magic aside, inventory is an important part of SanDisk’s supply strategy for 2009. SanDisk expects bit growth in sales for 2009 to be quite a bit ahead of its supply growth in bits, which is projected as less than 50%. The difference will be made up from inventory.
SanDisk is counting on leveraging this inventory help its business get healthier again.
The big, pleasant surprise of Q1 was sequential quarterly bit growth. SanDisk’s Q4 is the strongest of the year, whereas Q1 is the weakest. As one would expect gigabytes-sold pretty much invariably drops from Q4 to Q1. This Q1 was different.
This first quarter gigabytes sold actually grew 9% sequentially. Quite stunning given the economic meltdown.
Retail unit sales from Q4 to Q1 declined 30%, similar to the last two years. SanDisk had expected seasonal retail unit decline to be far more pronounced because of deteriorating economic conditions, but this wasn’t the big surprise. That was OEM.
OEM drove the upside. OEM unit sales grew 18% sequentially. Most of this growth came from microSD cards sold in the mobile market, in particular to mobile network operators.
Who would have thought that the lowly TransFlash card would have come this far?
Personally, I didn’t. I was far more enamored with msystems’ MegaSim, now morphing into a TransFlash cousin, the SanDisk Service Delivery Card.
TranFlash became microSD in 2005. Today microSD has become a mobile handset standard. It enjoys widespread industry support. It is the smallest memory card available commercially, with the lowest price per capacity and the highest capacity.
SanDisk microSD sales are even stronger in Q2 than Q1. From 05.04.09 LCM Semiconductor Report- Dan Amir:
“SanDisk microSD cards output is very strong. Based on our checks, SNDK’s daily shipments of microSD cards are up to 1.8M vs. 1.3M/day last quarter [Q1]. We believe most of the demand is still coming out of the China handset market and the increasing amount of smartphones entering the market. We view this data point as positive for SanDisk and it confirms our thoughts that business continues to strengthen at SanDisk.”
Almost 40% quarter-over-quarter microSD growth is rather remarkable in these dark days. I’m still having trouble getting my mind around SanDisk currently shipping 1.8M microSD cards a day. That’s a lot of cards.
MicroSD appears to be an important SanDisk story to watch. One could make the argument that it’s becoming the SSD of smart phones, as these handsets morph into powerful mobile computers.
All this said, SanDisk isn’t out of the woods yet. Demand can be fickle and benign NAND pricing is far from a given. The more time that goes by though, under these favorable pricing conditions, the healthier SanDisk’s business becomes.
Cost reduction is another very important factor. SanDisk expects its overall cost reduction on a cost per bit basis to be in the range of 40% to 50% for 2009.
Thanks to cost reduction, NAND pricing doesn’t need to go up for the magic of benign pricing to work. Holding steady-ish for a while, works wonders as well.