Even then, they still need a strategy for storage If the biggest problem is the distance between storage/memory and processing (bandwidth, latency, energy), then reducing and ultimately eliminating that distance seems like a reasonable strategy From a business perspective, the value is not in harvesting or storing data, it's in processing the data and extracting some gold out of it. Long term, don't even think that discrete memory and storage have a future, at all. Maybe cold storage could benefit from very low cost and high capacity dumb storage but all else got to be integrated, one way or another.
I can't even find funding data on "crunchbase"! My guess: WD owns Esperanto outright, as a means to growth like Yahoo with Alibaba and Yahoo Japan! Those investments carried Yahoo stock! The other designer in Esperanto sits on board of Mellanox a great outlet for Risc-V product! It's not about synergy. It's about GROWTH!
Seems that in the short term (whatever short term means), they are thinking of it as memory and aiming for a data-centric architecture with future systems. However, that's not easy, takes time, the economics must work and the interests of key players must align. Then there is the threat of embedded 3D memory chipping away at their SAM. So do they have that time and even if they do, is it the right path?
That's why I was wondering if maybe they are exploring 3D ReRAM as a synaptic device for neuromorphic computing. The long run is not about data and accelerators, it's about AI, that's where the value is. In data centers, CE, healthcare (enhanced humans and beyond), robots and more. BTW that goat in this month's cartoon is a cloud connected enhanced intelligence domestic animal - why enhance humans and machines but not other animals, that makes no sense. Anyway, integrating processing and memory, one way or another, got to be the better path. I suppose, an offload engine could be a first step towards that goal.