It appears to me that the initial "cached" Optanes were a loss-leader, to help Intel test the latest Optanes which combine 3D XPoint with QLC Nand. Nevertheless, without having their expense sheets in front of me, I suspect Intel had to spend much more money and far more time finishing the R&D, than originally projected. And, when that R&D was finally finished, Intel could have launched with drop-in replacements for 2.5" NVMe SSDs, and followed immediately with drop-in replacements for M.2 NVMe SSDs with adequate storage capacity. Those two form factors were already in widespread use worldwide. The 16GB and 32GB M.2s were B-A-D out of the gate, in part because they only used x2 PCIe 3.0 lanes = another BIG mistake. Meanwhile, Samsung stayed way out front, with superior price/performance proven in both form factors. Optane's faster latency did NOT justify the higher prices, which Intel demanded evidently to speed recovery of their massive R&D investment to reach market. Just my 2 cents, FWIW. Intel might increase demand by 40% if it reduced prices by 20%, but they won't know how "elastic" Optane prices will be, unless they try.