You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Does anyone have an opinion on the potential of SMIC becoming a serious player in the memory space with Crossbar memory. Will they be a threat to 3dXpoint from Intel and Micron? Any chance of dominating SSDs? Does SIMIC have the technical expertise to make this a dominant technology? Are there any variants of this technology on the horizon? Any thoughts when SIMIC will be delivering this to customers? Can Crossbar become a dominant technology against the competition? Anyone know of other companies licensing Crossbar? All thoughts and comments appreciated.
Upon further reading I have found the view that Crossbar will be combined on other chips that currently use a separate memory. Will this lead to a fundamentally different way into to how memory is incorporated into designs of systems? Thoughts and comments on this also appreciated.
tty2, About four years ago, they were talking about 1 terabyte memory and range of uses. Do you think this is just vaporware? I originally heard reports that TSM was developing a production process three years ago, but have heard nothing since. If there was potential, I think TSM would be using it. Any information welcome on this one, especially if anyone else might be working with Crossbar.
Intel was founded on a single produce, the DRAM, however as competition increased the Japanese vendors became dominant in DRAM and Intel decided to cut their losses and exit the DRAM business entirely, instead to concentrate on their new product, the Microprocessor.
Intel then eventually partnered with Micron on the 3Dxpoint memory technology, and it takes a huge financial commitment to lead in this memory market and stay in business through all of the wild boom and bust cycles that memory makers have endured for decades now.
SMIC may gain some initial market share with ReRAM, but let's see what Micron and Intel actually start selling in 2017.
Technology announcements on new products like 3Dxpoint need to be taken with a large grain of salt. Realty is when you are shipping product and the systems companies cannot order fast enough.
Crossbar reliability data would be nice to benchmark but probably just as hidden as 3D XPoint. Right now their main claim to fame or distinction is for the application and threshold-switching cell architecture similar to 3D XPoint. So I would say they are more appropriate to compete in that arena. However, being fabless, they would suffer a disadvantage compared to Intel/Micron, as logic-based foundry fabs such as TSMC, SMIC, etc. do not invest in process modules more suitable for making standalone high-density memory.
For embedded macro, their main challenge is they are quite behind other ReRAM companies in product development, such as Panasonic, Adesto, etc.
Dan, one factor that we can count on is that what is here and now will advance at a blistering rate. It will be interesting what platform has the best migration path to superior performance. I consider both these technologies just starting points and maybe there will be room for both, maybe not, only time will tell.
