Micron Buys Out Intel's Stake in Optane-Fabbing Partnership

The IM Flash Technologies (IMFT) joint venture (formerly called Intel-Micron Flash Technologies) is now officially dead, with Micron exercising its right to buy Intel out of the venture. 

Although Micron announced intentions to buy out Intel’s share of the IM Flash Technologies joint venture last October, Intel was quick to tell us that Micron wouldn’t be able to do that until January 1, 2019, as per the two companies’ contract. Now that January 1 has passed, Micron didn't waste any time in exercising its right to buy Intel out of the venture. Micron will have to pay $1.5 billion for Intel’s non-controlling (49 percent) share, and the company said it will eliminate the $1 billion member debt from its balance sheet.

Micron Technology President and CEO Sanjay Mehrotra seems to think that its development and commercialization of 3D XPoint, a non-volatile memory tech used in Intel Optane, will no longer be slowed by Intel: “The IM Flash acquisition will enable Micron to accelerate our R&D and optimize our manufacturing plan for 3D XPoint. The Utah-based facility provides us with the manufacturing flexibility and highly skilled talent to drive 3D XPoint development and innovation and to deliver on our emerging technology roadmap.”

Optane Sales Lower Than Expected

In 2015, Intel and Micron announced the 3D XPoint technology, claiming the memory would be up to 1,000 times faster than flash.

The two companies created the IMFT joint venture, which has been developing this technology back since 2006. The joint venture has also been responsible for the two companies’ shared production of NAND SSDs.

Even though Micron announced its own retail 3D XPoint product back in 2006, QuantX, the company hasn’t started to sell it. Instead, it kept selling all the 3D XPoint chips it was making to Intel. However, Intel’s Optane SSDs haven’t caught on as much as the company had hoped, in part because competing NAND SSDs have become almost as good for a much lower price.

Furthermore, the Optane DIMMs, which are arguably the more exciting application of 3D XPoint because they turn 3D XPoint into a sort of (slower, but still useful) DRAM alternative, have only recently come to market, due to the lack of support within Intel’s CPUs. The delayed support has hurt Intel’s ability to sell Optane DIMMs to the data center market.

Now, Micron seems to think it'll be better off on its own. It plans to release its own 3D XPoint SSDs later this year, once the second generation technology is ready.

Lucian Armasu
Lucian Armasu is a Contributing Writer for Tom's Hardware US. He covers software news and the issues surrounding privacy and security.
  • JamesSneed
    This might be good for both companies. Intel can free up some fab space and get a little money. Micron can start selling Optane how they want which hopefully is decent sized and affordably priced M.2 drives that at least compete with the "pro" SSD drives on price.

    Optane has been a weird you can use it as a cache or you can use it in a DIMM slot kind of tech that really isn't catching on.
    Reply
  • jimmysmitty
    21682069 said:
    This might be good for both companies. Intel can free up some fab space and get a little money. Micron can start selling Optane how they want which hopefully is decent sized and affordably priced M.2 drives that at least compete with the "pro" SSD drives on price.

    Optane has been a weird you can use it as a cache or you can use it in a DIMM slot kind of tech that really isn't catching on.

    The DIMM slot is only in servers currently. It wouldn't have been viable cost wise for consumers for a few more years. However I think thats the ultimate route to go. PCIe SSDs are good but they still are not as fast as memory. Having more direct connection to the CPU and lower latency would kill off one of the largest bottlenecks in the PC market.
    Reply
  • JamesSneed
    21682078 said:
    21682069 said:
    This might be good for both companies. Intel can free up some fab space and get a little money. Micron can start selling Optane how they want which hopefully is decent sized and affordably priced M.2 drives that at least compete with the "pro" SSD drives on price.

    Optane has been a weird you can use it as a cache or you can use it in a DIMM slot kind of tech that really isn't catching on.

    The DIMM slot is only in servers currently. It wouldn't have been viable cost wise for consumers for a few more years. However I think thats the ultimate route to go. PCIe SSDs are good but they still are not as fast as memory. Having more direct connection to the CPU and lower latency would kill off one of the largest bottlenecks in the PC market.

    I was aware :) I like to build PC's but am also in IT so follow servers and gaming.

    Yeah it's an odd time as servers can have rather large memory footprints which mitigates a lot of the needs for Optane especially since it has endurance wear as DIMM's effectively don't. To me Optane makes since as a SSD for pro/enterprise users or as cache inside of an SSD drive much like SLC was used as cache on cheaper MLC drives to help with the endurance but in the case of Optane it could have real performance impacts if implemented correctly..
    Reply
  • USAFRet
    Key words:
    " if implemented correctly"
    Reply
  • AgentLozen
    Optane memory is objectively better than flash and a worthy successor. Everything about this technology is desirable except for its price.

    Imagine if some car company built the first practical flying car. This flying car promises to:
    -Be lots of fun to pilot
    -Be safer
    -End traffic forever
    -Get you to your destination faster

    Get yours today for the low price of $1,999,999!!!

    Despite being the holy grail of driving, it would never catch on at that price point. I think technologies like Optane and OLED are in that spot. Once the price drops to a reasonable level then there will be no reason to keep the predecessor technologies around.
    Reply
  • Co BIY
    Ending the partnership with Intel doesn't seem like the best way to encourage them to add support for the tech in server CPUs which the article says is the key to selling in the data-center market.
    Reply
  • JamesSneed
    21683845 said:
    Ending the partnership with Intel doesn't seem like the best way to encourage them to add support for the tech in server CPUs which the article says is the key to selling in the data-center market.

    Maybe or maybe not. It may just be to free up fabs on Intel's side more than anything.
    Reply
  • mlee 2500
    I love my Optane 905p. The random and mixed performance is great, sequential very nearly the same as my Samsung M.2's, and I don't have to worry about endurance in a drive that I intend to keep and heavily use for several years.
    Reply
  • USAFRet
    21684721 said:
    I love my Optane 905p. The random and mixed performance is great, sequential very nearly the same as my Samsung M.2's, and I don't have to worry about endurance in a drive that I intend to keep and heavily use for several years.

    I'd love a 905p.
    But at $1200 per 1TB, that's a bit much.
    Reply
  • mlee 2500
    21685124 said:
    21684721 said:
    I love my Optane 905p. The random and mixed performance is great, sequential very nearly the same as my Samsung M.2's, and I don't have to worry about endurance in a drive that I intend to keep and heavily use for several years.

    I'd love a 905p.
    But at $1200 per 1TB, that's a bit much.

    Oh I agree, and that's why I use Samsung EVO's for my 1TB needs (Steam library etc.).

    However I was able to pickup a 480GB Optane disk (U.2 with M.2 adapter) for only about $500 which I use for my lower capacity but more transactional I/O. Since I'm not space constrained at all in that regard, spending an extra $200 for a faster drive which will last me several years or more wasn't such a bad value proposition.
    Reply