I posted some information in reply to this on my subreddit, but here are some additional details:
This is Micron's second replacement gate NAND generation, the 128L does exist but you don't see it very often and it's not nearly as much an improvement over the floating gate 96L. Replacement gate is similar to what Samsung uses, based on TCAT, which originally offered advantages like faster erases, faster operations (read/write), better data retention, and wider voltage threshold windows. In any case, Micron also adds CuA to the mix, which we will also see from Samsung (finally) and also with BiCS6 for that matter. Floating gate is still used for Intel's QLC (e.g. 144L) and may be the technology of choice for split-gate/split-cell due to the advantages of the semicircular cell shape with it.
The Phison E18 is a penta-core design - triple Cortex-R5 cores, with dual-core CoX - that in the 12nm process node can be clocked higher to handle the higher MT/s requirement. B27B can't quite achieve what Phison wanted with this controller. Speaking of which, Micron also has B48R with B57R in the works. There seems to be some NAND temperature throttling possible here, though, which is interesting. It's noteworthy because common knowledge is that "heat is good" for flash which is not really true. I'd go more deeply into that, but I'd like to write an entire article on the subject to fully explain why that's not an accurate view. However, for succinctness sake, you don't want cTLC exceeding 100C, but one reviewer had issues with the E18 throttling too early based on NAND temp.
Not much else to say technically, I guess the SLC caching result - direct-to-TLC and folding speeds specifically - speaks for itself. Given its a 32-CE design we can speculate as to per-die performance, although I think we're seeing a 96-LUN (plane) limit here. Micron released information on the flash late last year and Phison will have more on these in a week. I guess we'll see how new SKUs will be marketed!
