The speed of SSDs is vastly overstated, as they are generally rated at the speed of the fastest component and the real world slows them down. For a comparison of SSDs in actual devices check this out:
That's a comparison between macOS and Windows, not the SSDs themselves. I would not have believed how terrible Windows is at small file I/O unless I had seen it myself dozens of times over the years. If they had put Linux on one of those laptops for comparison, I would consider a chart like that fair.
If you actually read technical deep dives like AnandTech posts on SSDs and MacBooks alike, you would see that they really do perform identically to what Apple ships... because Apple is shipping industry-standard SSDs. Apple chooses expensive SSD chips, sure, but it's the same NAND everyone else has. Only the most recent generation of MacBooks (since the T2) has actually integrated their own SSD controller, inside the T2, but it's still the same physical industry-standard NAND chips underneath.
And that's exactly my point - the SSDs spec sheet numbers are based on the NAND chips, but the controllers slow them down in real world performance. I've still yet to see a benchmark that shows Macbook Air/Pro/iMac Pro level SSD permormance from any other SSD equipped computer.
You can buy this SSD today. There are others like it, and some, like Intel's Optane SSDs actually substantially outperform anything Apple has put in their computers in terms of IOPS, even though the Optane line hasn't focused on raw sequential throughput yet. Off the shelf SSDs get the much vaunted performance of "Apple's" SSDs (they're just normal SSDs...) being discussed in this thread.
In fact, the SSDs in those windows computers are as fast as the one I've linked to. The benchmark you were linking to was benchmarking NTFS vs APFS under a small file I/O workload, which NTFS sucks at. It was not benchmarking the SSDs in any effective manner. I am certain that I pointed this out in my comment above! If those laptops were copying a few large files, the performance would have been identical. If those laptops were running Linux, the performance would have been identical.
https://www.laptopmag.com/articles/2018-macbook-pro-benchmar...