Optane is faster than any PCIe 4.0 SSDs, there is no point in having SSDs that can transfer 1000000TB/s of data except maybe people that work with VERY large files. The important thing is latency and Optane SSDs destroy all the competition.
Optane's latency and IOPS are so much better than anything a consumer workload needs, though.
I'm not pretending that Optane isn't the highest-IOPS, lowest-latency product you can buy; I'm saying that regular NAND SSDs already have enough IOPS and low enough latency that consumer software rarely or never even
approaches their limits.
Meanwhile with NAND SSDs hitting 7GB/s transfers at 1/10th the price of Optane which caps out at 2.5GB/s, Optane is massively, obviously deficient in a very visible, measurable consumer metric, whilst costing 10x too much. If you're looking at squential performance/$, something like
the Samsung 980 Pro is 25x better than Optane. That's a performance/$ ratio that is too hard for almost anyone to ignore.
I think Optane is a great product, but like Intel, I don't agree it's a good fit for
most consumers. The remaining, tiny handful of consumers who would be interested in Optane's very specific latency advantage and who can also can swallow the cost and limited transfer rate disadvantages is such a tiny sliver of a market that it's not worth Intel catering to them.
With the best NAND SSDs available today, Optane latency is still 60% lower, but it's sequential throughput is also 60% lower. That means that for it to be a better choice than a PCIe SSD you need your consumer workload to exceed 20K IOPS and not involve any significant sequential components. Honestly, it's damn near impossible to find anything that needs even half of that without admitting that it's no longer a
consumer workload at that point.
If you need enterprise-class IOPS then Optane isn't going away, they're just not selling it to consumers anymore, and I think that is the right decision.
