Steam Machine also doesn't use an APU, so it needs both RAM and VRAM. With the RAM and NAND shortage, this significantly increased its starting costs almost immediately.
PS6 will use unified GDDR7 RAM + a small amount of DDR5 for CPU cores for system UI response, but it'll still be over $1000. AMD has to fit in fab time at TSMC for this semi-custom APU at a time when its datacenter CPUs and GPUs have demand at all-time highs, and getting anything more than 1TB of NAND flash at decent prices (with volume discounts) is going to be a near impossible task, even for Sony. The console chips could actually be fabbed at Samsung Foundry, as there are deep rumors of AMD holding serious talks with Samsung for a wafer supply contract. It wouldn't surprise me if consumer-grade chips were manufactured there, down to SF2 and maybe SF3, while older IODs remain on TSMC N4/5 and N6/7 (PRO and EPYC parts are supported for 10 years). New IODs may move to N3 coupled to N2 CCDs, as well as any Samsung Foundry options.
Agree with consensus that Sony should delay PS6 until the shortage eases, as the one game that we all know will push PS5 to its limit is GTA6. Just one game this entire generation? Nothing drives novel solutions in gaming like a lack of new hardware. Let them cook.
PS5 Pro versions of GT7 and FF7 Rebirth are definite improvements over base PS5. I'm getting really tired of terrible upscalers in console games that give nasty vaseline screen effect (checkerboard and spatial upscaling). So, if devs release more PS5 Pro patches with PSSR 2.0 and improve RT effects (and/or performance) by using Pro's RDNA4-based RT hardware, I think that'll help. PSSR 1.x is too artifact prone (though it has improved quite a bit since initial release) and I don't think 3rd party devs are too happy with it.
If PSSR 2.0 is a big improvement, maybe we'll see more PS5 Pro game patches if dev adoption improves and studios allow it. GTA6 absolutely needs to use it. Base PS5 will probably struggle with both framerate (unstable 30fps) and image quality, so it'll be a game that can push PS5 Pro's hardware improvements to a degree where the difference will be more substantial. Sony creating tools to find artifacts in a temporal multi-frame upscaler like PSSR is a good step forward too, and all of this directly benefits eventual PS6.
Generally agree about the poor state of PS5 Pro support. I suppose the larger issue stems from how long games are in development now (3-5+ years). Games don't need to get larger and larger (+ longer and longer). Honestly, as long as there's an interesting story and game mechanics, and maybe some kickass graphics effects (not solely RT) and high-res textures (tired of that muddy look), I think many would be completely fine with that.
PSSR 2.0 (FSR4-based) is probably the one good thing that'll be coming to Pro in 2026, but I hope the OS can override PSSR 1.x games with the new PSSR 2.0 API, otherwise it'll be a waiting game for new patches in supported games.
Getting good RT support in Pro does require developer time, so unless RT effects are completely in your face (eww), the difference between subtle RT effects in base PS5 and Pro won't be readily apparent. No one is going to be pixel peeping while gaming.
Also, I think Fidelity mode on all PS5s should target 40-48fps to get the most out of VRR TVs, not 30fps. With dynamic resolution scaling, this shouldn't be too difficult to achieve, and some distant rendering can be scaled back to regain fps within the same resolution scale as previous 30fps mode. I know some games use 40fps now, but I think it should be standard across the board.
Comments 3
Re: Opinion: The PS6 Cannot Come Out in This Era of $1,000+ Consoles
Steam Machine also doesn't use an APU, so it needs both RAM and VRAM. With the RAM and NAND shortage, this significantly increased its starting costs almost immediately.
PS6 will use unified GDDR7 RAM + a small amount of DDR5 for CPU cores for system UI response, but it'll still be over $1000. AMD has to fit in fab time at TSMC for this semi-custom APU at a time when its datacenter CPUs and GPUs have demand at all-time highs, and getting anything more than 1TB of NAND flash at decent prices (with volume discounts) is going to be a near impossible task, even for Sony. The console chips could actually be fabbed at Samsung Foundry, as there are deep rumors of AMD holding serious talks with Samsung for a wafer supply contract. It wouldn't surprise me if consumer-grade chips were manufactured there, down to SF2 and maybe SF3, while older IODs remain on TSMC N4/5 and N6/7 (PRO and EPYC parts are supported for 10 years). New IODs may move to N3 coupled to N2 CCDs, as well as any Samsung Foundry options.
Agree with consensus that Sony should delay PS6 until the shortage eases, as the one game that we all know will push PS5 to its limit is GTA6. Just one game this entire generation? Nothing drives novel solutions in gaming like a lack of new hardware. Let them cook.
Re: 'It's a Huge Upgrade': PS5 Pro Poised to Improve Massively in 2026
PS5 Pro versions of GT7 and FF7 Rebirth are definite improvements over base PS5. I'm getting really tired of terrible upscalers in console games that give nasty vaseline screen effect (checkerboard and spatial upscaling). So, if devs release more PS5 Pro patches with PSSR 2.0 and improve RT effects (and/or performance) by using Pro's RDNA4-based RT hardware, I think that'll help. PSSR 1.x is too artifact prone (though it has improved quite a bit since initial release) and I don't think 3rd party devs are too happy with it.
If PSSR 2.0 is a big improvement, maybe we'll see more PS5 Pro game patches if dev adoption improves and studios allow it. GTA6 absolutely needs to use it. Base PS5 will probably struggle with both framerate (unstable 30fps) and image quality, so it'll be a game that can push PS5 Pro's hardware improvements to a degree where the difference will be more substantial. Sony creating tools to find artifacts in a temporal multi-frame upscaler like PSSR is a good step forward too, and all of this directly benefits eventual PS6.
Re: PS5 Pro Users Starting to Feel a Little Shortchanged After Poor or Absent Support
Generally agree about the poor state of PS5 Pro support. I suppose the larger issue stems from how long games are in development now (3-5+ years). Games don't need to get larger and larger (+ longer and longer). Honestly, as long as there's an interesting story and game mechanics, and maybe some kickass graphics effects (not solely RT) and high-res textures (tired of that muddy look), I think many would be completely fine with that.
PSSR 2.0 (FSR4-based) is probably the one good thing that'll be coming to Pro in 2026, but I hope the OS can override PSSR 1.x games with the new PSSR 2.0 API, otherwise it'll be a waiting game for new patches in supported games.
Getting good RT support in Pro does require developer time, so unless RT effects are completely in your face (eww), the difference between subtle RT effects in base PS5 and Pro won't be readily apparent. No one is going to be pixel peeping while gaming.
Also, I think Fidelity mode on all PS5s should target 40-48fps to get the most out of VRR TVs, not 30fps. With dynamic resolution scaling, this shouldn't be too difficult to achieve, and some distant rendering can be scaled back to regain fps within the same resolution scale as previous 30fps mode. I know some games use 40fps now, but I think it should be standard across the board.