2x faster SSD may be more important than a 20% difference in TFlops. The latter is something you can adjust through resolution scaling, but you can't magically make up for a shortcoming in asset streaming/load times.
They're not comparable metrics, which is why it would make the "next gen power victor" crown completely muddy if that were to occur.
Based on credible rumors, the PS5 is more powerful (see: Reiner, Greg, Benji, Klee, Pachter, etc). They are all confirming the same thing. From what we can gather, they are talking strictly about TFlops, but we also have rumors that suggest a strong SSD advantage for Sony.
While the system architecture is the same (Navi, Zen, etc), there are components that could easily differ, and given the same price targets for both consoles one platform holder may prioritize different elements. I don't think we are going to see a universal advantage in every single elements from whoever is the victor unless the price targets are different.
No, that's literally not how technology works, man. Firstly you have to ask yourself what are the realistic speeds for an SSD drive at a price economical enough to include in a games console. These systems will be using QLC NAND drives, which is generally slower than TLC, MLC, or SLC. However, they will also likely be using 3D NAND in those drives, which is generally faster than 2D planar NAND.
After that though you have to figure drive size (very likely will be 1TB) and form-factor (possibly M.2 drives). Some of the SSDs packing 6-8 GB/s are pretty hefty-sized; granted some of that includes cooling and also the controller and DRAM cache (all three of which could be removed depending on just how custom the drives are, but in that case they would be completely custom and likely not using existing Samsung/Gigabyte etc. drives; those guys would probably not be interested in stripping down a current product line just for a custom part in a games console as they'd have to make due with tailoring manufacturing and assembly costs/pipelines for that too), but it's still something to consider.
Whatever differences there are with the SSDs, are not going to lead to massively varying levels of performance. We aren't talking CDs vs. cartridges here like in the PS1/N64 days. Fundamentally these SSDs will be the same medium and same technology; speed differential will come down to allotted PCIe 4.0 lane interconnects provided to the drives. Which we aren't even sure the final number will be, for either system. However I doubt we will see any drives approaching 6 GB/s or 8 GB/s in either system (I had a thorough discussion with another member in this thread (or the Phison SSD memory controller thread, can't recall which) and we basically broke down the practicality (in terms of costs, board real estate, size, TDP etc.) of 6-8 GB/s drives in either system. OTOH, the 40x number quoted for XSX's drive, no one has confirmation on what state system dev was in when the statement was made, or if it was even referring to an internal custom drive or simply the speed connection for a user-optional secondary "cold storage" drive.
For the TFLOPs thing, look I don't want to come off like I don't put any faith into the insiders on this account, but as far as substantiated leaks are concerned the only one that was completely deleted was the Github document with Oberon testing of clocks suggesting PS4 and PS4 Pro back-compat. That document was dated as June of 2019...which isn't THAT long ago, and Komachi has found other documents showing further Oberon and Ariel testing, all of them basically going with the same 36/40CU spec. TF range has ranged from as low as 8 to as high as 10.5 or so on those, all of which are possible depending on how low or high the clocks are set to on the GPU.
Given how quickly that particular Github leak was deleted, plus other database testing showing the same Oberon and Ariel chips, that to me seems like the most credible leak regarding PS5 at the moment. I'm also combining this with certain events and rumors that transpired regarding Sony and PS in 2019 before any leaks were happening, such as sudden SIE and corporate departures, PS headquarters relocation, rumors of a 2019 launch being delayed (supposedly due to PS4 sales; tbf PS4 sales were very very strong in 2019), and more along that ilk.
I'll put it this way: where are the testing database figures for this other possible PS5 that is 12-13TF? Why haven't these leaked yet? Why hasn't there been a single database test for such a chip? At least one such leak would have occurred by now. And keep in mind that the Github leak that was deleted also had Arden (XSX) results in there, which seem to be more or less accurate and relevant to the final design. Planning a system, even completely revising major parts of its design, takes LOTS of time and PLENTY of money. There are people outright thinking the Ariel and Oberon chip tests have no relation to PS5 despite the fact there is evidence of testing modes for clock speeds related exactly to PS4 and PS4 Pro. Unless MS plans to add native emulation for Sony consoles now, which I figure would be illegal.
While I always consider what those who claim to be insiders have to say on the matter, I need hard facts and data to work with here and in regards to PS5 the only consistent data that's popped up for it this late in the game are related to those Ariel and Oberon chips, and we know the performance levels for those through their phases. I take a lot of time to study things in terms of the tech world, and look into real-world factors that can affect the viability of certain things. It's why I was on #Team ReRAM before doing some more research and hopping right off that team when I saw where that tech is *actually* at in terms of potentially being mass-market ready. That's one such example but you see the point.
There's been other things in terms of rumors from other insiders too that just don't add up to me, and I guess I blame this on knowing a bit about how some of this tech stuff works. But, ultimately that isn't particularly important to mention here nor is it anything that sees these guys as any less potentially credible. However I also know that you can't control the authenticity or relevance of your sources, and I do think there's some rumors about with both systems that are either wrong (due to being impractical if you research into them) or may've been correct at some point but don't have a lot of context for them or have since changed and they aren't in the loop with said changes. Happens all the time.