Intel and Micron announce 3D Xpoint, a new memory technology that's 1,000 times faster than NAND

Shawn Knight

Posts: 13,022   +130
Staff member

Intel and Micron held a joint press conference earlier today to announce a breakthrough memory technology called 3D Xpoint (pronounced 3D crosspoint) that’s up to 1,000 times faster (read speeds) than today’s NAND, features 1,000 times better endurance and is 10 times denser than conventional memory.

3D Xpoint is described as whole new class of memory – not just a new type of NAND or DRAM – that operates fundamentally different than what’s available today. As the name suggests, it’s built on a three-dimensional architecture that uses a new type of switch instead of traditional transistors.

3D Xpoint is also non-volatile meaning it can be used for storage purposes as well as function as traditional memory although the two companies repeatedly implied that it will could also lead to new types of computing experiences not yet conceived.

The duo also pointed out that this is real technology, not just a PowerPoint presentation. To help drive home that idea, one of the first wafers from their joint production facility was shown off on stage.

It certainly sounds impressive, but what does it mean for the average consumer? Probably not a whole lot right now.

Intel and Micron said they will be sending out samples later this year before products using 3D Xpoint arrive in 2016. If I had to guess, it’ll be huge companies with data-driven needs like Amazon, Facebook or Microsoft that will initially put the technology to use in the real world.

As for consumer-based products, those will probably come further down the road. On the PC side, a new storage-minded interface capable of handling the kind of speed that 3D Xpoint can deliver will be mandatory as today’s SATA III interface is already a bottleneck with existing SSDs.

The technology could also find a home in mobile devices although if it’s more power-hungry than NAND, that could be a hindrance. It’s entirely possible that a new type of battery technology could neutralize that shortcoming but I wouldn't hold my breath on that happening anytime soon.

Permalink to story.



Posts: 292   +253
I hope this really brings about the death of the hard drive. And it can't come soon enough.

Gimme a 1 TB Xpoint SSD that doesn't break the bank and I'm sold.


^ probably 3-5 years from now on will get to consumer market, being affordable after 1-2 years from launch.
Newspapers and magazines chose to not even try to fathom the implications of internet connectivity until, for many of them, it was too late to catch up. The music industry, television broadcasting, phone industry, all the way down to cab services, same story. So many sleeping. So many now hurting. The same hurt may be on the nearing horizon for the tech industry as well. The source of so many marvels is quickly becoming it's own bottleneck. Technologies like 3D Xpoint memory, WiFi that is a hundred times faster than what is now available, this and that, blah, blah, blah, won't be available for YEARS! because of . . . well, think of any excuse you can imagine. Here's a good one. Who needs more than 640K of memory? An oldie but goodie, but it serves as a typical example of the lingering grip that clinging to set ways is just fine. "If it ain't broke . . .", and so on and so forth. The widely revered and innovative tech industry is settling into the lumbering and slumbering that others have found so inviting, so soothing, so self satisfying. Pinocchio is having a great time at the carnival.


I think we need DividedbyZero...where arrrrrrre you??
I must know more! :D


Posts: 553   +193
Imagine a memristor array connected in a spatial matrix, in between DRAM and flash. it helps?


Posts: 311   +50
1,000x faster.

Latency wise.

Pretty huge difference from actually being 1,000x faster, bandwidth wise, which is what pretty much everyone reading the article title is expecting.


Posts: 46   +79
Why don't you say what you really mean? the only reason this tech. doesn't become mainstream for end users in '16, is because of money. companies have a roadmap, and they don't want to DEVALUE what they have going on now. even though they could give it to consumers now, they won't, because of the money they would lose on current tech sales. they want to bleed every penny out of consumers with OLD tech, before starting their profits on new tech. it sickens me how much technology has suffered because of the greed of giant corporations. imagine where we would be if research was truly allowed to flow at its real cutting edge rate, instead of tech. being bought up by bigger corp.'s and shelved, to inhibit it from ruining what they are currently making money on. I can name at least twenty mainstream industries where this goes on every single year.