Posted by on
Categories: 3d Xpoint apple Micron NVDIMM

Analysts are beginning to forecast near-term #NVDIMM sales in the billions. @Apple once again is projected to be an early NVDIMM adopter. Is @Micron ‘s “New Memory” a potential player? In my last article we explored the notion that the #DRAM producers are likely to be deterred from being aggressive building additional wafer capacity in the face of the prospect of the (relatively) near-term arrival of new, non-volatile memories that will initially augment, and over time, replace DRAM. We cited IMFT’s announcement that they completed an expansion of the Lehi 3D XPoint fab in November of last year as a clear signal of the confidence of Intel (INTC) and Micron (MU) that the new technology was on track for commercial success. We then went on to speculate about the potential for Apple (AAPL) to be an early adopter of this technology. As it happens, other folks are thinking about that possibility as well. Recently, Alan Neibel of Webfeet Research presented just this scenario at SNIA’s recent “Persistent Memory Summit.” It seems that Mr. Neibel has a theory about the future of the memory industry, and it’s one that has some tantalizing elements in it for Apple and IM that are worth teasing out. Let’s consider what Mr. Neibel has to say. The Neibel gambit, if we may call it that, begins with a supposition about a revolutionary new memory, called “Non-Volatile RAM” (NVRAM for short) that’s close at hand. He outlines its characteristics and capabilities in this slide:  This slide is notable for a number of reasons, but probably most of all because it defines, in the combination of its features, a nearly perfect memory. Its power consumption should be excellent because it’s persistent, and that means that, unlike DRAM, it does not need to be refreshed. It has decent endurance (100’s of billions of cycles) and DRAM-like latency. It’s also DRAM-like in that it is, unlike many NVMs, DDR4 compatible. So what is this new memory? Mr. Neibel is coy on this point. Let’s compare it against the memory taxonomy generously provided in the Handy/Coughlin presentation that preceded Neibel’s talk. Take a look at the chart below. Do you see a fit?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.