SLC Price Premium and Profit Potential Persists

SLC NAND is More Profitable than MLC, Though Market is Quite Limited:

Digitimes reported this week that NAND contract prices were again rising, a little or a lot, in response to the supply takedowns of Oct-Dec, the idling of 200mm capacity, and a general, industry-wide production slowdown over the year-end holidays. Demand also appears to have been getting stronger, after the fall Shock Wave impact has worn off (we hope). Given the economic climate and lack of good forward visibility, the longer-term persistence of none of these trends is guaranteed, however.

In the price uplift, however, SLC NAND, though a small fraction of total NAND consumption, continues to show better pricing and better gross margins, as shown in the Digitimes chart below, assuming about a 2x cost-of-manufacture differential between SLC and MLC NAND:


Source: Digitimes, 16 January 2009


One might wonder why this premium has not been eroded in the fiercely competitive NAND market of 2008, in which any semblance of profitability has been sniffed out, and snuffed out by competitors seeking to right their own financial ships. But in the case of SLC, it has not, and it has been present for more than a year and a half, to varying degrees.

In an earlier BLOG, we discussed the long period from about late 2005 into 2H07, during which SLC was a worse financial performer than MLC...way worse. Gradually, and consistently since then, SLC has been considerably more profitable than MLC, which costs about half as much per GB to build as SLC NAND.

Certain applications need SLC NAND, and its longer endurance and generally superior R-W performance, and those applications are willing to pay for it. As successful as MLC practitioners have been at overcoming certain SLC advantages for some applications, not all users are willing to settle for the MLC solution; SLC persists in the market today, and we expect it to be here indefinitely.

The next wave of ‘price and performance’ differentiation is now coming to market with x3 and x4 NAND versions, with their own specifications for read and write times and endurance. We know for sure that those parts will be lower cost per GB to build and test. However, their performance specifications are still a moving target, and vary from vendor to vendor; some early x3 NAND product have little loss in performance compared to 2MLC. For sure, these new versions of NAND will find their own market niches, as flash chipmakers and their customers find ways to get the performance they need at the lowest possible price.

It is hard to envision the NAND market of 2-3 years down the road, and NOT see it fragmenting into many different price and performance domains, including all of the four basic ones listed above (SLC, 2MLC, x3 and x4), but also in the direction of the Micron-Sun Extended Endurance Roadmap, in which they have increased endurance by 10x for SLC and MLC, plus HS NAND options (ONFi or Toggle), and whatever other variations show up in the meantime. NAND Flash makers, controller makers and system architects can all work to close performance and specification gaps between SLC and MLC (or x3 and x4), but not all will be able to do it totally, or in the same timeframe. Different NAND makers and system products will specialize in different market segments (which, apparently, is now happening in SLC NAND)

From a profit perspective, with today’s NAND low prices, it is also easy to see that one 16M SLC NAND part contains more profit potential than an infinite amount of MLC, whose prices are probably below costs of manufacture for most vendors. This begs the question as to how long this price and profit disparity between SLC and MLC NAND can persist (or, how it has persisted so long already). Though SLC NAND is only about 8-10% of total NAND GB shipments, it is of sufficient size to support more competition and increased NAND makers' interest in these dire times.