Performance tests
"Not surprisingly, replacing DRAM with uncached Optane DC reduces performance by 20.1 and 23 per cent for memcached and Redis, respectively, whereas enabling the DRAM cache means performance drops between 8.6 and 19.2 per cent.
Reducing performance may sound undesirable, but it enables applications to work with a much larger in-memory dataset, as the test server could accommodate 1.5TB of Optane DC memory per socket, compared with 192 GB of DRAM."
Optane's challenge is cost/performance vs DRAM or NVMe. A quick comparison shows NVMe is around £400/TB vs Optane at £1000/TB and DRAM at £16,000 @64GB
Depending on what you are doing (i.e. is the server a VM or physical, how important is ECC/redundancy, what is the cost of managing servers vs the business benefit of the additional performance), I'm not sure the market it will be addressing will be that large for the applications mentioned (databases/caching).
i.e. would you take a server with 192GB RAM and mirrored 1TB Optanes or a server with 256GB RAM and mirrored 1TB NVMe drives? Particularly if the next step up maxs out Optanes at 1.5TB/slot versus 2TB/4TB NVMe if running large queries. YMMV but in 90% of the cases I would probably take the additional RAM, knowing that it will give me the most performance/flexibility for the majority of the workloads.