The oddest thing about last week's announcement was that it called phase-change 'a relatively new memory technology' when it has been under development for 40 years.
Phase change has been dismissed as a Techno-Ponzi technology - good for raising a development budget but for little else.
The densest commercial phase change chip is 128Mbit. Obviously two-bit-per-cell gets you to 256Mbit but, when flash is at 64Gbit, IBM's new advance doesn't go far towards solving the density limitations of phase change as a mainstream memory.
Indeed, IBM's announcement last week didn't focus on the technology's density potential, instead it went on about reliability.
All that IBM would claim for last week's advance was that it was 'a big step towards enabling practical memory devices based on multi-bit PCM'.
Apart from the signal achievement of inventing DRAM, IBM left it to others to invent the mainstream memory types: SRAM (Transitron), EPROM (Intel), EEPROM (Sperry), Flash (Toshiba).
IBM says it doesn't want to make phase change memories - merely find licensees - and commercialisation is seen as five years' away - meaningless in the semiconductor industry.
So phase-change lurks in the sub-fusc murk of the Techno-Ponzi underworld.