Predict: Next gen console tech (10th generation edition) [2028+]
- Thread starter RDGoodla
- Start date
Flappy Pannus
Veteran
Likely, no. Radeon R9 Fury aside, HBM has never been cost effective except for data center applications.
Tkumpathenurpahl
Veteran
Is it the packaging that makes HBM more expensive? And, if so, might a chiplet based console find it more cost competitive than they presently do?
Yes and the more HBM stacks you have the worse it gets since all of them have to be perfectly stacked on the interposer. Long term, there are companies and start ups looking at using organic interposers and simpler stacking to reduce the cost. I think that's still in the R&D phase and likely won't be commercially viable for a while.
HBM is really the right solution for bandwidth and energy requirements in a console, it's just too expensive still.
Frenetic Pony
Veteran
And supposedly in 2026, MS rushing out a new console gen to replace the failed Series.
Dear MS:
Mobile: Zen 6c, XDNA3, RDNA5+ on TSMC N3P. On package GDDR7, 18gb ram is enough (you only need PS5 + ai model), large MALL cache needed for maximum perf/watt. I'm talking like 96mb stacked SRAM here, you need to share it between CPU/GPU/NPU and ideally it'll hold an entire 1080p render target set (48mb just there). You need as short a trace as possible, "LLC include tag" as a compiler hint would be really useful here.
Dedicated: RDNA5 is chiplet so this can be biiiig, 100(50 WGP) is entirely plausible, go for a straight 4090 performance. Don't need a big CPU, Zen 6 full 8 core with large LLC is enough here, you'll be more limited by single core stuff as always. Only need around 128mb cache here, if it goes out to a 256bit bus who cares. The big question here is streaming. Private streaming to anything nearby that actually works well is a good selling point for high end, talking high quality 120hz 4k HDR encoding @100mbps, wifi-direct, mobile optimized 10mbps 1080p60 streaming, stuff like that. If your rich audience can play their high end gaming console on any of their TVs, or their Ipad, or their laptop, at the flick of a switch they'll be happy.
Controller: Do you have someone that actually, you know, likes video games? Not some accountant that thinks Series S and Gamepass are innovative? Good, put them in charge, especially of the controller. This thing could be a dual-sense, plus the Switch 2's haptic joysticks, plus it could break into two (hey it's needed for mobile anyway), you put accurate motion sensing in them and you have your possible VR controller/Wii like controller already just in case you somehow need that. Wif-Direct so you can carry it to the screen you're streaming to and get low latency controls back to the console. Speaking of, wifi direct audio, USB-C audio, 3.5mm headphone jack if needed, audio needs to be low latency when streaming but don't do Sony's droolingly stupid move of making it a proprietary codec that doesn't even exist only when there's plenty of better options.
Flappy Pannus
Veteran
Wccftech quoting Redgamingtech for specs of a console years out from launch, like what are we doing here?
I recall the RGTech dude saying the Series X would have the GPU on a daughter card for upgradability, and this was relatively close to launch. These are not people you need to pay attention to.
I recall the RGTech dude saying the Series X would have the GPU on a daughter card for upgradability, and this was relatively close to launch. These are not people you need to pay attention to.
I would add to that stop with all the security chip nonsense people should be free to use whatever controller they want
Had some spare time a while back and got to thinking about chiplets, got a bit carried away with some mockups/concepts for the entire [PS6] generation. Kind of out there with stacked HBM and all kinds of craziness.. I know there's some unlikely stuff here but I thought it'd at least serve to get those brain-juices flowing about different approaches. One can dream!
Also, I know there's likely a strong aversion to the concept of a low-powered, budget system at launch due to the XSS, but I think that's just a matter of poor execution (~40% bandwidth rather than 65-75%, low absolute memory, 1/3 gpu power rather than 1/2-to-2/3 etc.) so I baked the idea of a better-executed lower power system into this, along with "midgen" consoles and a "twilight" console for the transition to the gen afterwards.
In a spoiler-tag to avoid image spam:
Posted elsewhere but thought I'd share them here too.
Also, I know there's likely a strong aversion to the concept of a low-powered, budget system at launch due to the XSS, but I think that's just a matter of poor execution (~40% bandwidth rather than 65-75%, low absolute memory, 1/3 gpu power rather than 1/2-to-2/3 etc.) so I baked the idea of a better-executed lower power system into this, along with "midgen" consoles and a "twilight" console for the transition to the gen afterwards.
In a spoiler-tag to avoid image spam:
Posted elsewhere but thought I'd share them here too.
Johnny Awesome
Veteran
I don't agree about the S. I think it would have been nice for it to have an extra 2GB, but one of the advantages of its current power level is that it might be shrinkable into a handheld by 2026 that can launch with a massive GP game library. Forza Horizon 5 on a handheld? Sign me up!
I'll say again that I'm fine with MS launching a new system in 2026 along with a handheld. Especially now that they aren't as concerned about winning hardware wars.
Any speculation on how good a 2026 Xbox could be? I'm betting we're only heading to 20-24 GB of memory, better RT and ML upscaling.
I'll say again that I'm fine with MS launching a new system in 2026 along with a handheld. Especially now that they aren't as concerned about winning hardware wars.
Any speculation on how good a 2026 Xbox could be? I'm betting we're only heading to 20-24 GB of memory, better RT and ML upscaling.
Last edited:
The next-gen Xbox console may not be designed by the team behind the Series X, with one industry insider claiming that the project has been reassigned to Microsoft's Surface division.
cheapchips
Veteran
Who knows with the rumor mill stuff, but assuming it's true, what benefit does it offer to switch teams? Surface devices are not exactly cheap mass market devices like consoles.
Last gen and this, Xbox haven't exactly done an amazing job vs Sony with regards to cost reduction. Maybe volume provides Sony with this luxury, but their cost reduction isn't just when they get to switch process nodes. Xbox seem to tied to those bigger steps.
It's not going to be a surface device per se. This suggests a different future strategy and product redesign if true. MS is probably not planning to follow the traditional console path this time around. Just like Nintendo is also doing things differently and succeeded. Whereas MS just can't maintain momentum for the last two generations, being like Playstation. So they are reconsidering their product offering.
Flappy Pannus
Veteran
Bear in mind Surface got a big shake-up recently with Panoy's ouster. Frankly I was never impressed with Surface as a product line, far too expensive and slow to update.
Well I could imagine that they now go for an open PC platform with min requirements for the hardware and everyone (certified) can bring out such a console with the Xbox os.
With the current generation (and the one before) it is clearly visible that not many developers optimize for the hardware even if it only because dev cycles are so incredibly long, that the developers even don't know for which generation a game gets released and optimization is costly (time & money). So most of the time games seem just to get a different ini-file for the settings and get ported to the API of the target.
At least for most games this seems to be the case. Everything else is more or less PC hardware. Even the decompression hardware lost it's magic with the PC ports (as they just save a few more CPU cycles).
Also the demand for switch and steam deck shows, graphics quality is not that important as many in forums always say. So even a mobile variante is thinkable. Maybe even sooner with a mobile series s.
With the current generation (and the one before) it is clearly visible that not many developers optimize for the hardware even if it only because dev cycles are so incredibly long, that the developers even don't know for which generation a game gets released and optimization is costly (time & money). So most of the time games seem just to get a different ini-file for the settings and get ported to the API of the target.
At least for most games this seems to be the case. Everything else is more or less PC hardware. Even the decompression hardware lost it's magic with the PC ports (as they just save a few more CPU cycles).
Also the demand for switch and steam deck shows, graphics quality is not that important as many in forums always say. So even a mobile variante is thinkable. Maybe even sooner with a mobile series s.
Tkumpathenurpahl
Veteran
I think it bodes well if the Surface team are responsible for the next XBox hardware.
A tablet/laptop kind of form factor may be achievable by the time of 3nm. Going by MS's own figures, the Series S tends to hover in the 72W range (although loading screens and in-engine cutscenes are ~150W for some reason) which would place a 3nm Series S equivalent at (72*.7*.7=) 35.28W for typical usage: too high for a Switch/Steamdeck style portable, but feasible for the aforementioned form factor.
A tablet/laptop kind of form factor may be achievable by the time of 3nm. Going by MS's own figures, the Series S tends to hover in the 72W range (although loading screens and in-engine cutscenes are ~150W for some reason) which would place a 3nm Series S equivalent at (72*.7*.7=) 35.28W for typical usage: too high for a Switch/Steamdeck style portable, but feasible for the aforementioned form factor.
Similar threads
