Where is Micron’s QuantX?

Micron Quantx LogoFor more than a year The Memory Guy has been fielding questions about Micron’s QuantX products.

First announced at the 2016 Flash Memory Summit, this brand name has been assigned to Micron SSDs and DIMMs that use the Intel/Micron 3D XPoint Memory.  Originally QuantX products were scheduled to ship in 2017, but Micron is currently projecting availability in 2019.  My clients wonder why there have been these delays, and why Micron is not more actively marketing this product.

The simple answer is that it doesn’t make financial sense for Micron to ship these products at this time.

Within two weeks of the first announcement of 3D XPoint Memory, at the 2015 Flash Memory Summit, I knew and explained that the technology would take two years or more to reach manufacturing cost parity with DRAM, even though Intel and Micron loudly proclaimed that it was ten times denser than DRAM.  This density advantage should eventually allow XPoint manufacturing costs to drop below DRAM costs, but any new technology, and even old technologies that are in low-volume production, suffer a decided scale disadvantage against DRAM, which sells close to ten billion units per year.

If the sales price of a chip is lower than its cost to manufacture the manufacturer loses money.

Intel is losing money on 3D XPoint Memory by selling it at a price that is below its production cost (as I recently pointed out in this post.)  Since Micron owns the XPoint manufacturing plant (IMFT) Micron deeply understands this, and has a very solid knowledge of where production costs are today, and where they are likely to be over the next several quarters.

Why is XPoint losing money?  This is relatively simple to explain.  3D XPoint Memory won’t fit into the memory/storage hierarchy unless it is sold for a lower price than DRAM.  This is something that I regularly explain by using this chart, which shows how the Memory/Storage Hierarchy works.

The Memory/Storage Hierarchy

The short explanation of this diagram is that each level of the hierarchy, from the three cache levels on the upper right to tape on the bottom left fits within the hierarchy as long as it’s cheaper than the next faster level, and faster than the next cheaper level.  If it doesn’t fit, then computers will use a cheaper alternative that provides equal or better performance.  3D XPoint Memory can fit between DRAM and SSDs as long as it’s faster than NAND flash (which it is by its very nature) and it costs less than DRAM, which today is an issue.

3D XPoint must sell for a lower price than DRAM or it will not be adopted.

Since XPoint uses a new technology that is not yet in high-volume production it will continue to cost more to produce than DRAM until its production volume approaches that of DRAM.  Production costs will drop below DRAM costs as volume increases, but volumes will only increase if it’s priced below DRAM.  This is a Chicken-and-Egg problem.  Since mid-2015 Objective Analysis has held the position that for the next 2+ years Intel will need to sell XPoint at a loss to cultivate a market.  It is now 3½ years later, and this product still seems to be having trouble reaching cost parity with DRAM.

So it costs more to produce than the price it must sell for in the market.  Anyone who sells it will lose money.

Why, then, is Intel producing a money-losing product?  3D XPoint is a necessary part of Intel’s platform strategy.  Without the support of Optane DIMMs Intel’s most costly new Xeon server CPUs won’t perform any better than the company’s cheaper/older CPUs – systems without Optane will bog down the processor’s performance

Intel has been addressing the issue of platform performance since the early 1990s.  A part of this has been the company’s orchestration of DRAM interface migration: FPM, EDO, SDRAM, DDR, DDR2, DDR3, & DDR4.  The company discovered that poorly designed 486-based systems could impede their high-end CPU sales, so they have worked tirelessly ever since then to assure that the platform does nothing to slow the performance of Intel’s highest-performance processors.

Still, how does it make sense for Intel to lose money on Optane Memory?  Here’s how I usually explain it:

Intel may lose $10/system for the Optane it sells, but that will enable it to sell a $50 more expensive CPU (these are completely fictitious numbers).  While this is good business, it means that only Intel can make money by selling 3D XPoint at a loss.  Micron will need to wait until XPoint’s cost drops below DRAM costs to make money selling Quantx.

As a result, readers shouldn’t expect to see Micron promote QuantX until 3D XPoint’s cost falls below that of DRAM.  And with a DRAM price collapse already started, this could be a very long time.

There’s no wonder that Micron has been so quiet about QuantX.  Its manufacturing cost isn’t yet low enough to make sense.

It also makes sense that Samsung’s response to Optane is to make a fast NAND flash – Z-NAND.  The company can do this by customizing the circuitry inside of standard NAND flash chips, and these changes don’t require any new technology.  Z-NAND will be built using Samsung’s existing NAND process but with a different chip layout (i.e. lots of planes & sense amplifiers).  They are betting that they don’t need to debug a new memory technology to participate in the new memory layer.

Other companies hope to compete against 3D XPoint with various alternative technologies (i.e. ReRAM, MRAM, etc.)  Any company that embarks on this path will need to simply eat the loss for 2+ years without any hope of ever recovering it.  It’s hard to imagine that anyone will really do this.

A more detailed understanding of 3D XPoint Memory can be gained by reading the report: A Close Look at the Micron/Intel 3D XPoint Memory, which can be purchased for immediate download on the Objective Analysis website.

21 thoughts on “Where is Micron’s QuantX?”

  1. intel just killed the cpu+optane bundles, at least for consumer side, that says a lot.
    marketing 16gb optane + 4gb dram as 20gb ram didnt help either..

    1. Yes. It’s hard to see why a consumer would be attracted to one of the Optane bundles. The few who understand that they want more DRAM are likely to buy more DRAM.

  2. I want to see this tech or tech like it put out for mass storage to replace the deadend nand flash with it’s limited re-write lifespan. I don’t care about using it for system ram, because it also has a limited re-write life, just a lot longer the flash. Flash is fin for storing stuff like settings, but not for main storage. I want stuff I buy to last.

    1. Peter, Not at all plagiarism. The Seeking Alpha writer who goes by the name “Gold Panda” reported on an article I wrote and a presentation I made. I, in turn, decided to turn that presentation into a blog post.

      Both cite the same source, but that happens all the time in articles.

      Also consider the fact that the author and I came to completely different conclusions.

      Thanks for pointing out the article, though. I hadn’t seen it before.

      Jim

    2. That article literally cites Objective Analysis. So he did not plagiarize himself.

  4. As opposed to what is written here, 3D-Xpoint price per GB is cheaper than DRAM by a factor of 2 or more (Amazon).

    1. Dror,

      Thanks for your comment. I didn’t realize that the post was a little confusing since it discusses both cost (the manufacturer’s cost to produce the product) and price (what that product can sell for in the marketplace.)

      I have made some small corrections to the post that will hopefully clarify the point.

      Jim

  5. The cheaper/faster hierarchy depicted in your graph makes good sense for equivalent types of memory, but that would seem to break down when you compare non-volatile memory to typically faster volatile memory products like DRAM.
    With the coming boom of the various, portable IOT products, QuantX may well find a nice market niche where memory capacity and speed are valued in a low power consumption, frequently switched-off product.

    1. Good point, Michelangelo. What matters most is the price of the total solution. In certain cases XPoint will offer that.

      But if Intel sells Optane below cost and Micron tries to make a profit on QuantX SSDs and DIMMs, guess who the customers will buy it from?

      On the other hand, if we assume that Micron sells it in chip form (which the company has never said they would do) there might be an opportunity in applications like the ones you suggest. It might be a better, less expensive option than NOR plus SRAM. Everspin is certainly trying to do that.

      I’m not holding my breath.

      Jim

  6. What Intel needs is a “killer app” to spur adoption of Optane DIMMs as an intrinsic component of big data applications.
    I have created the world’s fastest C++ hash table for non-volatile memory. A version can be downloaded for free at http://ThreeMisses.com
    It’s part of a C++ library that will make it very easy for C++ programmers to access terabytes of data by key with microsecond-scale latency.

    1. Steve,
      I usually don’t allow people to pitch their products in the comments for this blog, but this time I will make an exception. Your Three Misses program seems like a good match for the ginormous memories that are possible with Optane DIMMs.
      Good luck with it.
      Jim

      1. Certainly! And I hope I am not too touchy about this.

        You wouldn’t BELIEVE how many people want to post ads for Nike shoes etc. on this site!

        Jim

  7. “3D XPoint must sell for a lower price than DRAM or it will not be adopted” you say. But WHY?? It is two completely different products as one hold the data when powered off while the other does not!!! Just like some wanted to pay extra fro a SSD over a HDD there will be people who want to pay more for the 3D XPoint than they the cost of the SSD in order to get that absolutely insane performance upgrade!! 3D XPoint is to be compared to SSD / NAND and NOT DRAM!!

    1. Niels,

      Thanks for the comment.

      The way that 3D XPoint fits into the memory hierarchy it must be cheaper than DRAM (although it can be more expensive than an SSD) and faster than an SSD (although it can be slower than DRAM). This is how every layer fits into the memory/storage hierarchy.

      This is also why XPoint needs to be compared to both DRAM and an SSD.

      https://TheSSDguy.com/3d-xpoint-memory-at-the-storage-developers-conference/

      Intel agrees, as you can see from this post:

      https://TheMemoryGuy.com/intels-optane-dimm-price-model/

      Jim

  8. When should Micron release consumer 3D XPoint drives or Intel releases their second generation of 3D XPoint?
    My 58GB Optane is really small for a boot drive and I constantly need more space; but Optane is so expensive!

    1. Nuno,

      I am sorry to say that the outlook for profits has gotten worse thanks to this year’s DRAM price collapse. As a result, Micron may not ship QuantX products for a long time!

      I don’t know when the next generation of 3D XPoint will be released, but you can be certain that it will still have the same pricing: Half as much as DRAM. Optane will always be expensive when you compare it with a NAND SSD.

      Thanks for the question!

      Jim

  9. Why wouldn’t Micron seek out AMD to enable AMD to make the same loss in memory, gain on CPU type sales? Would that not enable Micron to begin building a larger 3D Xpoint market?

    1. Jac,

      That would be a smart thing for Micron to do, but I believe that a lot of this technology (if not ALL of it) belongs to Intel, so Micron would probably be disallowed from selling it to support AMD processors.

      It will be interesting to see what AMD does in response to Optane. They have already had four years to determine a strategy, so I would guess that their response will become clear to the computing community in the next year or two.

      Jim

  10. Maybe you could discuss historically how companies were able to insert technologies into the chain. For example, which companies inserted HDD into the chain? I suppose it was IBM. Did they sell at a price below cost for a period? Did HDD cannibalize their punched card business? Why did punched cards leave the chain? Was the chain shorter in the past? Will the chain always continue to get longer as links i.e. technologies are inserted?

    1. LLoyd,

      You’re not the first to ask for something like this. I made copies of the hierarchy for the 1960s and 1970s that I will share in a future post.

      I don’t think I can find the information I need to analyze the losses for each of these technologies. HDDs were developed in the 1950s, and memory chips started to replace core memory in the late 1960s. That’s a long time ago, and many pioneers in those fields are no longer in existence. The ones who still are don’t hang onto their records that long. I believe that only 7 years is legally required.

      I would argue that Punched Paper Tape and Hollerith Cards are input devices rather than storage. That moves them off the chart.

      Thanks for taking the time to share your thoughts.

      Jim

Comments are closed.