DRAM always provides us with something to talk about, and as we head into the second half of 2012, there's been quite a bit of speculation lately surrounding the memory market's next price points.
After a relatively serene first half of the year during which DRAM average selling prices (ASP) managed a relatively small yet steady 20 percent overall increase, volatility returned. Since early July, computing DRAM prices decreased approximately 10 percent, dragging consumer DRAM prices down too and leaving little doubt the trend will continue for the short term. The question to be answered is when will it end and where will pricing be when it's time to buy memory for Christmas and year-end builds?
My current answer is: about right here, right where July's market currently stands. Many say the current, sudden ASP downtrend was caused by Elpida Memory Inc. , who began aggressive PC DRAM sales to quickly bring in cash. Competitors followed suit, just to sell something. That, however, might have simply been the last gram of saltpeter tipping the scale in the wrong direction -- sooner than expected, but not unexpected.
For the last year plus, DRAM manufacturers attempted to keep ASPs up (some think artificially so). While production capacities grew due to smaller wafers and new fabs, manufacturers did not loosen the reins of that full global capacity. When PC DRAM demand grew less than expected, they moved capacity to other, less volatile and higher margin niche products. With the excess inventory manufacturers accumulated, they competed in the enterprise/server market, fighting for shares in the clouds of data storage. Finally, some DRAM makers simply refused to lower ASPs below profit-making levels.
But a flood that kept PC makers from attaining first half 2012 sales projections, a worsening global economy, and more wafer shrinkage, meant mounting inventory. Though ASPs increased, it was a matter of time before DRAM makers let off some pressure and sold some of that excess inventory. When inventory levels become acceptable again, off-selling will stop and prices will bounce back. That should come soon, before the end of the third quarter, when consumer electronic and PC manufacturers need to buy for Christmas production.
So, now we're wondering how high ASPs will bounce back. Will the introduction of Windows 8, increased tablet sales, or those Christmas requirements return us to pre-July watermarks?
I'm reserved about the answers I hear in the marketplace: "Strong DRAM Sales in the second half," "inventory indexes strengthening," and "With Elpida gone, others should benefit."
Intel Corp. (Nasdaq: INTC) has already announced they anticipate their PC market sales to be lower in the second half, with hope of tablet demand offsetting some loss -- despite the introduction of Windows 8 to lure buyers back to stores. Inventory indices remain higher now than in the first quarter of 2011. While they may trend lower over the next few quarters, they will remain at precarious levels.
Meanwhile, much of Elpida's lost capacity will be filled by Micron's increased capacity, if that prospective merger bears fruit. And the global economy has a ways to go before a mass of consumers from any country earn enough disposable income to generate increased PC sales above current projections.
The good news is DRAM manufacturers have a better handle on inventory management now than ever before. Tablet wars, cloud servers, and increasingly complex consumer electronics eat away at the elephant capacity of PC DRAM and will help the memory market overall balance average pricing to acceptable levels for manufacturers.
Now, if Taiwan DRAM makers -- those connected to Elpida and those who are not -- rupture at all, that'll require an update. Until then, I would expect ASPs to bottom out in the next month, followed by another slow but steady climb through the middle of the fourth quarter. Prices won't top the end of the second quarter but will be close to it. So, "strong" DRAM prices may be too strong a phrase for what lies ahead.