The two analysts have worked together to produce a report entitled: ‘Emerging memory takes off,’ an update to the 2020 memory report by Objective Analysis and Coughlin Associates.
The emerging memory market comprises novel technologies, usually for non-volatile data retention, based on a wide variety of technologies. These include ferroelectric RAM, magnetic RAM, phase-change memory (PCM), resisitive RAM based on filament growth and destruction, and so-called correlated electron RAM.
Emerging memories have begun to replace established memory technologies in SoCs. Meanwhile, stand-alone emerging memory chips, such as MRAM, FRAM, and ReRAM, are winning designs in new applications.
Embedded non-volatile memories will grow by initially by displacing NOR flash memory and volatile memories DRAM and SRAM. Although a trend towards chiplet packaging may impact the need to embedded large memory arrays monolithically.
“The Internet of Things will be revolutionized as new embedded memory types slash power consumption. Larger systems are already changing their architectures to adopt persistent memories to improve latency and data integrity,” said Jim Handy general director of Objective Analysis, in a statement.
The report estimates that 3D XPoint memory could reach volumes of more than $20 billion per year by 2031. However, this represents a considerable reduction in the growth curve for the PCM-based technology. A year before Coughlin and Handy had estimated that the 3D XPoint market would be worth $25 billion in 2030.
It is notable that Micron has stopped making 3D-XPoint non-volatile memory and is selling the wafer fab at Lehi, Utah where production has been focused (see Micron turns its back on 3D-XPoint, puts fab up for sale).
Intel has sold its NAND flash and solid-state disk business to SK Hynix but has retained its Optane business base on 3D-XPoint (see Intel sells NAND memory business to SK Hynix and IEDM: SK Hynix makes PCM-based 3D crosspoint memory).
Over the same period standalone MRAM and STT-MRAM revenues are expected to grow to about $1.7 billion, or over forty-two times 2020’s standalone MRAM revenues. And embedded ReRAM and MRAM will compete to replace the bulk of embedded NOR and SRAM in SoCs, fueling even greater revenue growth.
Related links and articles: