A couple of years ago, I remember hearing about the possibility of micron based Octa Level Cell (OLC, 8 bits per cell) SSD technology (therefore even higher capacities per storage cell), but subsequently, these rumours came to nothing.
Below is an article about Penta cell (5 bits per cell) SSD flash memory and the reasons why we might not see it for many years (likely 2026) here:
https://hexus.net/tech/news/storage/147964-dont-expect-ssds-using-plc-2026-says-wd-president/
TL;DR
Shifting from 4 bits per cell (known as QVO) to 5 would result in a large increase in errors, and therefore, more logic dedicated to error correction (and additional logic required for redundancy, in case NAND cells fail). It would also only improve capacity over QVO NAND by 25%, so a somewhat marginal gain.
Additionally, "a PLC controller must be able to read/write 32 voltage states per cell. This is a much more complex / exacting task for controllers to do". It looks like there is a linear increase (+100%) in the complexity of memory controllers , for each increment in bits per cell.
That would mean a 400% increase in memory controller complexity (vs existing QVO SSDs), in order to create a hypothetical Octa Level (OLC) based SSD, with likely less than double the storage per cell of a QVO SSD (presumably, this could improve by adding additional layers to the fabrication processes used to create NAND cells).
The industry as a whole seems currently more interested in adding more layers (as this seems be be a much more cost effective way to improve SSD density) to TLC (3 bits per cell) and QVO, with a couple of manufacturers already developing 176 layer NAND fabrication processes. Link here:
https://www.anandtech.com/show/16299/sk-hynix-announces-176layer-3d-nand
According to Samsung "if we apply a double-stack technology, a 256-layer stacking is mathematically possible". Link here:
https://en.yna.co.kr/view/AEN20201201006900320
The current gen. Samsung 870 EVO (128 layer, TLC based) EVO costs ~£198 for the 2tb model, with 176 layer (cheaper??) Samsung SSDs coming later this year.
Sorry if this is old news to some.
Below is an article about Penta cell (5 bits per cell) SSD flash memory and the reasons why we might not see it for many years (likely 2026) here:
https://hexus.net/tech/news/storage/147964-dont-expect-ssds-using-plc-2026-says-wd-president/
TL;DR
Shifting from 4 bits per cell (known as QVO) to 5 would result in a large increase in errors, and therefore, more logic dedicated to error correction (and additional logic required for redundancy, in case NAND cells fail). It would also only improve capacity over QVO NAND by 25%, so a somewhat marginal gain.
Additionally, "a PLC controller must be able to read/write 32 voltage states per cell. This is a much more complex / exacting task for controllers to do". It looks like there is a linear increase (+100%) in the complexity of memory controllers , for each increment in bits per cell.
That would mean a 400% increase in memory controller complexity (vs existing QVO SSDs), in order to create a hypothetical Octa Level (OLC) based SSD, with likely less than double the storage per cell of a QVO SSD (presumably, this could improve by adding additional layers to the fabrication processes used to create NAND cells).
The industry as a whole seems currently more interested in adding more layers (as this seems be be a much more cost effective way to improve SSD density) to TLC (3 bits per cell) and QVO, with a couple of manufacturers already developing 176 layer NAND fabrication processes. Link here:
https://www.anandtech.com/show/16299/sk-hynix-announces-176layer-3d-nand
According to Samsung "if we apply a double-stack technology, a 256-layer stacking is mathematically possible". Link here:
https://en.yna.co.kr/view/AEN20201201006900320
The current gen. Samsung 870 EVO (128 layer, TLC based) EVO costs ~£198 for the 2tb model, with 176 layer (cheaper??) Samsung SSDs coming later this year.
Sorry if this is old news to some.
Last edited: