Wikibon CTO urges caution over 3D XPoint memory tech


Intel Corp. and Micron Technology Inc., set the IT world abuzz this week when they unveiled their new 3D XPoint memory tech that offers up to 1,000 times the performance and endurance of regular NAND flash. The companies are touting 3D XPoint as a “breakthrough in technology”, but big question marks remain over its ability to surpass DRAM, NAND flash and processor chips in the enterprise, writes Wikibon CTO David Floyer in his latest piece, “Intel and Micron take the Plunge with Non-volatile 3D XP”.

3D XPoint is being lauded by its creators as a “new class of memory tech,” but Floyer urges caution, saying that Intel and Micron will need to generate significant volume demand for the product if it’s to succeed in dislodging conventional memory technologies.

In their announcement, Intel and Micron said they would initially target 3D XPoint at high-performance SSD storage in the data center. However, 3D XPoint is unlikely to become a replacement for NAND Flash for the forseeable future, unless the companies can produce it at a significantly cheaper cost. As such, Floyer doesnt expect the technology to surpass Flash in the enterprise IT market any time soon.

Instead, 3D Xpoint’s best prospects for success seem to be as a replacement DRAM, starting in consumer products, where its power efficiency makes it a far superior option, Floyer writes. According to Floyer, the most likely use case for 3D XPoint will be in mobile devices and wearable tech, where it could significantly improve the battery life of such gadgets. Even so, these gains would need to be substantial enough to justify rewriting operating systems to take advantage of 3D XPoint’s non-volatile qualities.

Floyer says that ultra-small devices and military applications the most likely volume market for 3D Xpoint, as everything can be manufactured in a single microchip. However, he warns that the time to volume in these markets would take at least five years, if not longer.

One key question that could affect all this is whether or not Intel will allow 3D Xpoint to be used with ARM processor chips. On the one hand, Intel could decide to keep the tech exclusively for its x86 chips and take one last crack at penetrating the mobile market. However, Floyer believes that such a strategy would almost certainly fail. As such, Intel may well decide to let Micron provide 3D XPoint on ARM processors and go after the mobile space alone.

What with so many uncertainties at the moment, Floyer recommends enterprise professionals take a “wait and see stance”. Everything depends on whether or not 3D XPoint can gain traction in the consumer and/or military markets first, the analyst argues. If and when that happens, enterprise adoption would likely follow 2-3 years later.

Floyer’s full report is available on the Wikibon Premium Website.