Flash memory maker SanDisk (UNKNOWN:SNDK.DL) may have caught a break, as Samsung (NASDAQOTH:SSNLF) has supposedly pushed out its 3D NAND ramp after reports that it cut specialty equipment orders. SanDisk's timeline doesn't include 3D NAND technology until the second half of 2015, whereas Samsung began mass production last August, and Micron Technology (NASDAQ:MU) intends to start sampling its 3D NAND products in the first half of the year.
Meanwhile, SanDisk remains a leader in planar, or 2D, NAND solutions, providing the lowest cost. The company will ramp its 1Y (19nm x 19.5nm) technology this year, which made up just 15% of production in the fourth quarter of 2013, and transition to 1Z next year. Meanwhile, Micron's 16nm technology has room for improvement as well as its 20.5nm in the bit line direction.
Both companies will likely squeeze as much as they can out of planar technology, and stand to gain from any delays in Samsung's ramp.
There's a physical limit to developing flash memory cells, and the cost reduction from approaching that limit is not as great as earlier leaps in technology. As returns on investment diminish for planar memory chips, chipmakers turned to stacking NAND strings vertically on top of each other. In this way, they can cram more transistors onto one die, thus reducing the cost per bit.
But 3D isn't more cost effective yet. Samsung's first iteration that it started producing in August of last year featured 24 layers of stacked cells. The process uses just a 40nm node (Micron and SanDisk are using 19nm), so there's room for improvement using a smaller node. But, the next generation of cell stacking is where significant cost reduction will come in, as 24 layers grow to 32 layers or more.
As a result, analysts and insiders don't expect for cost crossover to occur until the second half of 2015. At an analyst conference, Micron's VP of R&D said, "Now, when we look at the projection, we think it's more like a second half of 2015 story when [3D NAND] actually starts becoming something significant in the market place. And a lot of that is, again, driven by the fact that planar technology is cost competitive."
SanDisk is focused on cost first, which means squeezing its node size down to the smallest possible where it still gets cost improvements. That will likely end with its 1Z technology, which is expected to be 16nm, before the company transitions to 3D manufacturing.
3D is more expensive to produce because it's more difficult to check a cell's operations when it's buried under a bunch of other cells. The process also requires additional specialty equipment to fabricate chips.
As a result, SanDisk believes it can get a better return on investment through 2D NAND over an early transition to 3D NAND. On the company's fourth-quarter earnings call CEO Sanjay Mehrotra said, "We really believe that as long as you can extend the 2D NAND life, it is in the best interest in terms of ROI and in terms of really applying that technology across a broad range of industry applications."
The company's timeline for introducing 3D NAND remains in the second half of 2015, the same time cost-crossover is expected, and in the meantime leads the industry with the lowest-cost planar solution.
A little breathing room
Samsung's 3D NAND technology might be first to market, but it might not be seeing the demand Samsung initially anticipated. It's not a big part of its flash business, let alone entire business, but it was a threat to flash memory makers that are slower to transition like SanDisk.
SanDisk's core business looks to have a bit breathing room on pricing with Samsung unable to make as big of a dent in the market with 3D NAND. The company will continue to face pressure on its pricing, however, as 3D technology costs move closer to planar costs.
The real test will be in a year and a half when SanDisk needs to manage its own transition to 3D NAND.