Ummm. I think you are confusing the roles of embedded NVRam and...

Ummm. I think you are confusing the roles of embedded NVRam and SSD NVRam. The latter is intended for data centres and computer enclosures i.e. highly environmentally controlled and shielded environments it is not intended for sticking in a wearable or the car where the temperatures can scale. It is intended for massive Terabyte storage sites where power costs, read write speed and unit cost are the critical considerations as these issues can mean millions of dollars difference in set up and operating costs. In these locations the slower write speed, higher power consumption and high per unit die cost are nonstarters.

The versatility characteristic doesn't come into it. It is simply not a relevant factor for the targeted use case of the 4DS reRam. It may or it may not be as versatile - but nobody buying it cares. They can't use auto-temperature insensitivity in a data centre, so they aren't going to pay for a feature they aren't going to use. They need lots of NVRam, they need it cheap, they need it to work very fast, they need it to use little power, they need it to last for ages, and they need it to withstand an obscene number of read-write cycles without cell failure - because they are using it in insane quantities for high volume access with a ludicrous density of access cycles. This is the space where the 4DS reRam has been designed to excel. The polar opposite of what embedded reRam needs to be able to achieve.