3D XPoint is finally history for Micron by now at the latest: The only factory for production is sold. This ends the humble chapter for Micron, because there was almost nothing left of the cooperation with Intel for the group. Intel now has to handle the production for itself in its own factories.
A loss for Micron
Established in 2006 by Intel and Micron as a joint venture IMFlash, the factory in Lehi, Utah, was most recently completely subordinate to Micron. Micron bought out Intel two years ago when it was foreseeable that 3D XPoint would not work as hoped. For this, Micron paid Intel $ 1.5 billion.
Now the entire factory will completely change hands again for 1.5 billion US dollars. US $ 900 million went to the factory and thus the purchase price to Texas Instruments (TI); Micron sells the equipment for US $ 600 million. TI could continue to use most of the tools, equipment and tools, although the new owner plans to upgrade the 300mm wafer factory and build 65nm and 45nm chips there. Most of the employees in the factory should continue to be employed there.
Micron's problem child is history
The chapter 3D XPoint for Micron now ends ingloriously with many millions US dollars in depreciation and loss. Micron has been on the losing side from the start when it comes to the whole 3D XPoint issue. Extremely hyped at the start, Micron was ultimately just the manufacturer for a single customer: Intel.
The future of 3D Xpoint: uncertain even with Intel
With a few exceptions, the market hardly wanted and still wants the memory today, further development is lagging behind, the chips are still there of the first generation, although somewhat improved, in use. There are other generations on the roadmaps, but this technology is still not under a bright star. Fab11X will build the chips and advance the technology at Intel in the future, while Pmem 300 is still on the roadmaps for Sapphire Rapids.
In the end, however, it wouldn't be surprising if the Technology will be discontinued sooner or later – just as Intel recently separated from the NAND division. The new CXL interconnect, for example, should produce suitable storage solutions. But even these developments do not happen overnight, research has been carried out on 3D XPoint for decades and so far the right one for a broad market has not been found.