Investors welcomed news from Micron Technology (MU +10.95%) that the company had developed, along with Intel Corp. (INTC +12.93%), a next-gen memory technology that doesn't use the traditional transistor architecture. Micron popped 9% on the news, as the new memory claims to be 1,000 times faster than traditional NAND and 10 times denser than current DRAM. While that sounds impressive, a look into the details reveals that investors may be getting ahead of themselves, as the technology isn't likely to drive financial results in the near to medium term.
A product without a market?
The new XPoint technology marks an entire change in chip architecture, using not traditional transistors but instead a property change in the material itself. Memory manufacturers have been able to pack more transistors onto a single chip or wafer, but they've always known they would eventually reach a limit on how small these transistors can be, which is why they've been investing in 3D NAND and other technologies such as XPoint.
The technology has properties that place it between traditional NAND and DRAM, while it will probably be costlier than current NAND and slower than DRAM. Computers and mobile devices use both DRAM for "scratch-pad" memory and NAND for storing files and data, and since XPoint doesn't excel at either cost or speed, it isn't going to be replacing either one of these two main types of memory. That's something both companies admit to.
In other words, because current computers and mobile devices are built to use the best qualities of both DRAM and NAND, it seems XPoint may be a solution in search of a problem. Currently the only applications analysts see XPoint being used for are enterprise applications and in-memory databases where speed is essential and it's helpful to bring the data close to the processor and remove bottlenecks. Granted, over time XPoint's best properties may be harnessed for new applications and uses, or the cost will come down, but this process will take months or years.
Industry fundamentals aren't changing
I think investors got particularly excited about this announcement because it differentiates Micron's portfolio, potentially adding a product that nobody else on the market currently has. That gives investors hope that Micron may someday not be in a purely commodity business and can break free from the frustrating boom-bust cycle all memory manufacturers must deal with. Adding to the hope was Micron's statement that this technology would be ready in the next few quarters, showing up in products by 2016, as it doesn't require a huge amount of retooling or new aspects of manufacturing.
But even if this were true, it isn't clear there's a huge demand for this type of memory yet. Traditional transistor NAND and DRAM will remain the bread-and-butter for Micron and the industry for the next few years. That said, the industry is increasingly looking to be peaking, with indications that a supply glut is forming. The demand side isn't looking too hot in the near term, either. PC shipments declined 9.5% in the second quarter of 2015, according to Gartner. Most companies have already upgraded their computers after the end of XP support, and it isn't likely we'll be seeing an uptick from Windows 10 until late 2016. Also, according to Gartner, tablets are expected to continue to contract, as the market is completely saturated and people are holding on to their devices longer. The mobile-phone market is expected to grow only 3.3%, largely driven by a slowdown in China, which is also looking increasingly saturated.
Micron and Intel should be applauded for such a revolutionary breakthrough in memory technology, and we could see some interesting use cases for the new memory down the road. But while this development is good for competitive positioning in the long term, investors shouldn't get too excited in the short to medium term, as Micron's profits and stock will continue to be dominated by the industry supply and-demand cycle.
