Pretty cool to see Kepler pop up in these parts
Two SOCs for SIE next-gen sounds plausible. We've already heard about the potential PS portable; should be a safe guess one of those SOCs is in relation to that.
But now I've been wondering, what if the PS6 is more like a Gamecube -> Wii type of situation, at least in raw specs, and the
real gains are in custom accelerated silicon that's also shared in some way with the portable? Look at it this way: PS5 Pro is going to be $800 when you also factor in the stand and disc drive. SIE probably have good margins on it, but I doubt production prices become so low in the next 4 or so years that they can put out something like a PS5 Pro @ $499 while retaining the same margins, especially if they go with a more advanced node process (2nm is rumored).
At the same time, SIE know that a proper new next-gen PS6 at $699/$799 is going to lead to a much lower overall install base, and that could limit them on the one aspect of growth that makes some sense: new younger players, including new young adults. Is it worth cutting them off just to have the highest possible margins through a high price? And it's not like SIE won't have to greatly increase budget in some areas like RAM (they'll need at least 32 GB's worth, and at higher bandwidth).
The hardware at this point, at least looking at things like the PS5 Pro, aren't even the limiting factor in game performance anymore. Time, budget, and manpower are; putting out some monster PS6 costing $800 is just going to ultimately mean even fewer AAA games as they take longer and longer to make, costing more to make, meaning safer games. How can the AAA market survive that? PS6 needs to focus on technologies that help cut down costs, team sizes, and time of game development, both in the console hardware and in their development suites. In terms of base TF for example, does a PS6
really need to be anything more than maybe 25% over the PS5 Pro? Perhaps even less than that?
Meanwhile I could see them targeting the lower-end SOC to be something around PS5 performance, maybe as a portable that's like a PS4 Pro in performance undocked and PS5 level while docked. Have smart tech shared between both and make sure in the portable it can also just "automatically" configure PS6 games to run at the portable's specs; save as much of the profile building from the dev's hands as possible. But I think that means their next-gen tech will have to include things like AI-accelerated ways of smartly scaling mesh geometry, using framebuffers to determine it, and maybe even having an ability to generate predictive framebuffers on the GPU similar to how CPUs can do branch prediction, using results of previous framebuffers & metadata related to the player session to do so. A bit like frame generation in concept, but only partially (hopefully 10th gen systems also have some form of frame generation in the package though).
At least on the tech side of things, there are a lot of interesting possibilities for 10th gen SIE console systems.
