Support NeoGAF

[FireFly]

We know even 9070 non xt, with its 220W tdp is at 225% of performance vs base ps5 gpu, and that card launched march 2025 on 5nm.
Gotta use logic here, for 2028 tech on 3 or 2nm process node we will get vastly more performance in ps6, compared to current desktop gpu's think something around 4090, including ai upscaling and rt performance too.

Ofc unpredictable stuff could happen like ww3 or dunno alien invasion, but if things go smoothly ps6 gonna be at least 3x stronger from base ps5 on top of few times better ai upscaling and rt capabilities(we can safely assume at least some form of basic raytracing will be default/baseline for next gen games).

The issue is that the cost per transistor is not going down significantly with newer processes. If you want more CUs on 3nm, you're going to have to pay for them. The 9070 is 357 mm^2 on 4nm, compared with 279 mm^2 for the PS5 Pro on 5nm/4nm. It has Infinity Cache which could be removed, but doesn't have a CPU. So if we take the Pro as the base, we're looking at adding more memory, still increasing the size of the chip even to hit 9070 level performance, and then reducing the price to $599? If the PS5 Pro breaks even at $550 say, then Sony would only have $50 to play with before they start taking losses.

(3nm looks like it would enable a ~25% boost to clock speed, and 2nm looks to be too expensive, hence Kepler implying it's not being used).

Edit: Another thing to point out is that a 256-bit bus on GDDR7 is only expected to give around 1 TB/s of bandwidth. That's what the 4090 has, but the 4090 has a huge increase in L2 cache to match the Infinity Cache on Radeon cards. If you're stripping out the IC to save die space, you won't get that benefit.

 

[PeteBull]

The issue is that the cost per transistor is not going down significantly with newer processes. If you want more CUs on 3nm, you're going to have to pay for them. The 9070 is 357 mm^2 on 4nm, compared with 279 mm^2 for the PS5 Pro on 5nm/4nm. It has Infinity Cache which could be removed, but doesn't have a CPU. So if we take the Pro as the base, we're looking at adding more memory, still increasing the size of the chip even to hit 9070 level performance, and then reducing the price to $599? If the PS5 Pro breaks even at $550 say, then Sony would only have $50 to play with before they start taking losses.

(3nm looks like it would enable a ~25% boost to clock speed, and 2nm looks to be too expensive, hence Kepler implying it's not being used).

Thats why i think 3x powerjump is doable but we can forget about 400 or even 600$ price, its gonna be 999$, including discdrive, so maybe bestcase scenario 800$ discless but who really knows maybe its 999$ discless
Inflation is real thing after all and that 300$ that seemed like a bargain back in 1995( first playstation launchprice) is equivalent to 632$ currently, and lets remember we got almost 3 and a half years of inflation left still

 

[PeteBull]

To be fair PS4 Pro did not deliver a 2x performance jump but a 2x TFLOPS one. We are mixing a few things here IMHO.

Some games it was at 2x, some only +78% aka 1080p to 1440p jump (+78% pixels on screen, like in case of uncharted4), but important thing is, it was achieved at same price of 400$/€ and took only 3 years, aka 2013 to 2016.

 

[Panajev2001a]

Some games it was at 2x, some only +78% aka 1080p to 1440p jump (+78% pixels on screen, like in case of uncharted4), but important thing is, it was achieved at same price of 400$/€ and took only 3 years, aka 2013 to 2016.

To be fair the PS5 Pro is likely less subsidised now (likely maybe even turning a little profit) than it was in the PS4 Pro case, so that is something objectively worse but then again it was kind of expected too. Both consoles were quite a bit memory bandwidth limited.

Otherwise, with silicon manufacturing process scaling being what it is (it takes longer and longer to make meaningful performance jumps and it is costlier and costlier to design and manufacturing chips each time you jump to a major new node… each performance jump also needs to be greater and greater to make a meaningful impact) it was expected for PS5 Pro to have a tough time bringing a massive impact to games especially for devs that did not spend a lot of time making Pro optimised versions and even more so for those that did not push hard on PSSR and RT adapting their RT code to take advantage of the changes in PS5 Pro's GPU (HW stack management was one of them… not all improvements did avoid the need for that to take full advantage of). With that said with FSR4 like PSSR we may see PS5 Pro pull ahead more.

I recall posting several times well ahead of launch that I was not super excited about a Pro release because I did not think releasing more HW more often is the answer, it is less and less so as time moves on… especially when it is not a new console generation allowed to have exclusive games designed around its features at all and where effortlessly compatibility is the name of the game.

 

[PeteBull]

To be fair the PS5 Pro is likely less subsidised now (likely maybe even turning a little profit) than it was in the PS4 Pro case, so that is something objectively worse but then again it was kind of expected too. Both consoles were quite a bit memory bandwidth limited.

With all the loss on sony gaas initiative its very likely but us/end customers dont give a fuck that sony's concord blew up so much cash they gotta make it up on higher ps5pr0 price.
Lets even assume sony makes 50$/€ on ps5pr0, its price is horrendus for the low specs it has, here in europe even now over 6 months after launch u gotta pay 850- 900€+ including discdrive
U can tell its selling terribly coz its getting those promos already:

https://www.coolblue.de/en/product/955502/playstation-5-pro-disc-drive.html

First few months it was even over 1k euros including discdrive, now both ps5pr0 and discdrive getting cheaper and cheaper (still way above US price tho)xD

 

[nkill20]

With all the loss on sony gaas initiative its very likely but us/end customers dont give a fuck that sony's concord blew up so much cash they gotta make it up on higher ps5pr0 price.
Lets even assume sony makes 50$/€ on ps5pr0, its price is horrendus for the low specs it has, here in europe even now over 6 months after launch u gotta pay 850- 900€+ including discdrive
U can tell its selling terribly coz its getting those promos already:

https://www.coolblue.de/en/product/955502/playstation-5-pro-disc-drive.html

First few months it was even over 1k euros including discdrive, now both ps5pr0 and discdrive getting cheaper and cheaper (still way above US price tho)xD

pro bom is 550 max, sony is upselling it hard.

 

[Panajev2001a]

pro bom is 550 max, sony is upselling it hard.

They are upselling it but not that much, it is a premium model meant to generate revenue (again, most people will buy it without a disc drive). Same reason they increased Switch 2's price to the degree they did.

It is expensive but not a waste of money I think. People need to have reasonable expectations too, releasing HW more frequently is not the solution but it makes the problem worse (including cross gen titles window).

 

[Proelite]

The CPU and GPU does for Magnus can be split to use in discrete solutions. They're put together using advanced packaging.

AMD GPUs starting with RDNA 5 all have Xbox BC.

Logically working backwards what does that mean?

 

[Proelite]

My guess for the tier of Xboxes / OEM devices powered by Magnus.

Zen6 CPU
16GB of DDR5 on 128bit bus. ~100GB/s
18GB of GDDR7 on 192bit bus. 768-896GB/s.
68CUs up to 3ghz. up 52 teraflops fp32.
2TB SSD.

Legacy Windows games / apps run perfectly without any modifications due to split system + gpu ram. No latency problems going to GDDR7.

New / bespoke games can treat use the DDR5 pool as they see fit. Series X BC games can use the DDR5 pool as an extra cache the same way that BC games on X already does.

OS is obviously all in DDR5 pool.

 

[vaibhavpisal]

My guess for the tier of Xboxes / OEM devices powered by Magnus.

Zen6 CPU
16GB of DDR5 on 128bit bus. ~100GB/s
18GB of GDDR7 on 192bit bus. 768-896GB/s.
68CUs up to 3ghz. up 52 teraflops fp32.
2TB SSD.

Legacy Windows games / apps run perfectly without any modifications due to split system + gpu ram. No latency problems going to GDDR7.

New / bespoke games can treat use the DDR5 pool as they see fit. Series X BC games can use the DDR5 pool as an extra cache the same way that BC games on X already does.

OS is obviously all in DDR5 pool.

Is there going to be a Series X and S?

 

[Proelite]

Is there going to be a Series X and S?

There are going to be way more than 2 SKUs in the mid to high tier, depending on what OEMs and Xbox want to mix and match CPU And GPU wise.

That being said the low / mobile / handheld S tier will be monolithic APU / the successor to Z2/Z3 extreme.

 

[vaibhavpisal]

There are going to be way more than 2 SKUs, depending on what OEM wants to max and match CPU And GPU wise.

How many gpu configs are expected?

I imagine there will be only 1 cpu.

 

[Proelite]

How many gpu configs are expected?

I imagine there will be only 1 cpu.

At least 2 for the GPU? AT0 and AT2. Basically the same die as in the discrete off the shelf GPUs. CPU idk, maybe they have a Zen 6 with v-cache too. Depends on how many consumer Zen6 dies without integrated GPUs and with the bridge latch are made.

 

[MrAlwaysRight]

Why are we not seeing AMD's 3D V-Cache tech anywhere else besides PC?

Heat mainly, silicon budget and manufacturing yields are the other reasons in the main.

 

[Kazekage1981]

Will the new xbox sku's:

1) mobile ones have 5G connection along with Wifi7?
2) PCIE Gen 5 based?
3) Have USB 4.0 ports?
4) Mobile ones have some sort of graphene type battery that last long and does fast charging?
5) Latest display port and HDMI specs?

 

[twilo99]

Heat mainly, silicon budget and manufacturing yields are the other reasons in the main.

Shame. Makes a huge difference in gaming

 

[MrAlwaysRight]

Will the new xbox sku's:

4) Mobile ones have some sort of graphene type battery that last long and does fast charging?

no

 

[memoryman3]

My guess for the tier of Xboxes / OEM devices powered by Magnus.

Zen6 CPU
16GB of DDR5 on 128bit bus. ~100GB/s
18GB of GDDR7 on 192bit bus. 768-896GB/s.
68CUs up to 3ghz. up 52 teraflops fp32.
2TB SSD.

Legacy Windows games / apps run perfectly without any modifications due to split system + gpu ram. No latency problems going to GDDR7.

New / bespoke games can treat use the DDR5 pool as they see fit. Series X BC games can use the DDR5 pool as an extra cache the same way that BC games on X already does.

OS is obviously all in DDR5 pool.

Only 16GB of DDR5? 32 would be more futureproof no?

 

[Proelite]

Only 16GB of DDR5? 32 would be more futureproof no?

Total ram amount is 34GB. Considering PS6 will have 24 or 32GB GDDR7 + 4GB LPDDR5, it should be fine.

 

[memoryman3]

Total ram amount is 34GB. Considering PS6 will have 24 or 32GB GDDR7 + 4GB LPDDR5, it should be fine.

Wouldn't a 32GB + 4GB PS6 pose potential problems for this configuration? I also expect most gaming PC's to also come with 32GB of system memory + 12 or 16GB of video memory by 2028, and possibly 24-32GB of VRAM in 2032. A developer could just decide to use 20GB or so for video RAM, or use a lot of system memory for simulations. There's also the issue of games without DirectStorage needing to copy textures/meshes from system RAM (good luck with only 16GB!) so where does that leave this build?

Stalker 2 has low 1% lows on 16GB RAM vs 32GB…

 

[Proelite]

Wouldn't a 32GB + 4GB PS6 pose potential problems for this configuration? I also expect most gaming PC's to also come with 32GB of system memory + 12 or 16GB of video memory by 2028, and possibly 24-32GB of VRAM in 2032. A developer could just decide to use 20GB or so for video RAM, or use a lot of system memory for simulations. There's also the issue of games without DirectStorage needing to copy textures/meshes from system RAM (good luck with only 16GB!) so where does that leave this build?

Stalker 2 has low 1% lows on 16GB RAM vs 32GB…

I think the Magnus system should be able to hit 4k60 using FSR4 on all optimized games released up by 2026, with all discernable graphical features activated.

Future proofing for PC games imo won't be necessary as refreshes with Zen7 and/or RDNA6 will come in 2-3 years. Also OEMs will address any performance gaps with more/faster ram, more CUs enabled and higher core clocks.

But if that baseline needed for the 1st part is 32GB of DDR5, then $50 added extra to MSRP makes sense.

 

[Kazekage1981]

Lets not forget the most important thing about the new consoles:

The SSDs!!!!!!!!!! XD

Will the SSD speeds be higher based on PCIE Gen 5? I am hoping 10GB/S-20GB/S

 

[StreetsofBeige]

I havent kept up with this thread, but if the next Xbox console is around $1000 US, forget it. That would be $1,300-1,400 CDN. I like console gaming, but I dont like it that much. I'll just game on my laptop.

PS5 Pro is already around $950 here and even that is getting absurd. But PS5/X going for about $600 at launch I can stomach. Checking them now, disc versions now go for about $700-ish after price hikes. That's stretching it already.

 

[Panajev2001a]

Lets not forget the most important thing about the new consoles:

The SSDs!!!!!!!!!! XD

Will the SSD speeds be higher based on PCIE Gen 5? I am hoping 10GB/S-20GB/S

They pretty much have to if they want to keep up with the GDDR7 main pool bandwidth increase (proportionally). Likely 10+ GB/s RAW bandwidth as well as Kraken decompression rate improvements (there is a limit on how much silicon they will want to spend on increasing the I/O block, but it is fixed function optimised silicon still)… as well as some maybe AI related tricks like neural texture compression would do the trick.

They will not want to go for the craziest SSD specs as it would make storage extension crazy expensive for customers and kind of defeat the purpose of cost efficiency of PS6 (sure some of that is for better profit margins).

I do not think Sony is going for the Apple model of no expandable storage and lots of SKUs with varying storage (2 TB, 4 TB, and 8 TB at crazy prices albeit it would allow them to get the kind of bandwidth they need cheaper as the SSD would be completely custom) though… so you want customers to expand SSD storage and do it with the same performance of onboard one… so even in 2 years you want fast but affordable 1 TB prices.

 

[Physiognomonics]

Lets not forget the most important thing about the new consoles:

The SSDs!!!!!!!!!! XD

Will the SSD speeds be higher based on PCIE Gen 5? I am hoping 10GB/S-20GB/S

This is already PS5 speeds with Kraken compression. Average is 10GB/s with Oodle textures. Up to 20GB/s with some type of data like geometry. I even expect them to keep the same PS5 I/O as few developers are actually properly using it.

Total ram amount is 34GB. Considering PS6 will have 24 or 32GB GDDR7 + 4GB LPDDR5, it should be fine.

32GB? Isn't that expensive and overkill? They could do 192bit bus (to reduce APU size) with 2GB Chips used in clamshell configuration and replace that with 4GB chips as soon as possible to reduce costs. total size: 24GB. Bandwidth probably going to be on the low side as usual for Cerny. Mayby only 768GB/s with 32Gbps chips. They could use 36gbps chips to get to 864 GB/s which sounds right for PS6.

 

[Panajev2001a]

This is already PS5 speeds with Kraken compression. Average is 10GB/s with Oodle textures. Up to 20GB/s with some type of data like geometry. I even expect them to keep the same PS5 I/O as few developers are actually properly using it.

The problem will be keeping up with bandwidth improvements with main RAM otherwise you will need to go back and buffer a lot more data compared to the PS5 dreams of more on demand paging. Still I do see RAM increasing and it should be possible to have in 2027 cost effective 10+ GB/s RAW bandwidth SSDs that with Kraken can get in the 20-30 GB/s range once you factor compression (Kraken decompressor would go from a max of 21 GB/s or so to 30-ish GB/s but it should be possible to hit 20 GB/s more often although lowering latency is probably even more important than increasing raw bandwidth).

32GB? Isn't that expensive and overkill? They could do 192bit bus (to reduce APU size) with 2GB Chips used in clamshell configuration and replace that with 4GB chips as soon as possible to reduce costs. total size: 24GB. Bandwidth probably going to be on the low side as usual for Cerny. Mayby only 768GB/s with 32Gbps chips. They could use 36gbps chips to get to 864 GB/s which sounds right for PS6.

16 to 24 GB would be a very small jump and this is total system memory not just VRAM that needs to last you till 2034 as base specs (Pro console might only add auxiliary OS memory to free a bit more resources for games). I think 32 GB is a sensible generation upgrade.

 

[Physiognomonics]

The problem will be keeping up with bandwidth improvements with main RAM otherwise you will need to go back and buffer a lot more data compared to the PS5 dreams of more on demand paging. Still I do see RAM increasing and it should be possible to have in 2027 cost effective 10+ GB/s RAW bandwidth SSDs that with Kraken can get in the 20-30 GB/s range once you factor compression (Kraken decompressor would go from a max of 21 GB/s or so to 30-ish GB/s but it should be possible to hit 20 GB/s more often although lowering latency is probably even more important than increasing raw bandwidth).


16 to 24 GB would be a very small jump and this is total system memory not just VRAM that needs to last you till 2034 as base specs (Pro console might only add auxiliary OS memory to free a bit more resources for games). I think 32 GB is a sensible generation upgrade.

In 2027 you can't have a 32GB console as 4GB GDDR7 chips won't be available in mass production. They are not even on any roadmap right now. And this would mean use a more expensive 256bit bus, increasing size of APU forever. This is why I think they could use some clamshell configuration with a cheaper 192bit bus and save on cost as soon as 4GB chips are available, which is what they did with PS4.

 

[HJuggernaut]

Wouldn't 384bits be a first for console?

no xbox one x had that

 

[Panajev2001a]

In 2027 you can't have a 32GB console as 4GB GDDR7 chips won't be available in mass production. They are not even on any roadmap right now. And this would mean use a more expensive 256bit bus, increasing size of APU forever. This is why I think they could use some clamshell configuration with a cheaper 192bit bus and save on cost as soon as 4GB chips are available, which is what they did with PS4.

That is a possibility which I think they will attempt. Next-generation will be RAM hungry, especially if they bet on PT and AI.

 

[Proelite]

32GB? Isn't that expensive and overkill? They could do 192bit bus (to reduce APU size) with 2GB Chips used in clamshell configuration and replace that with 4GB chips as soon as possible to reduce costs. total size: 24GB. Bandwidth probably going to be on the low side as usual for Cerny. Mayby only 768GB/s with 32Gbps chips. They could use 36gbps chips to get to 864 GB/s which sounds right for PS6.

After the recent MLiD's leak, I agree with you. I need to be more conservative with my PS6 speculations.

24GB of 36gbps chips to get to 864GB/s should be plenty.

 

[DrKeo]

This SOC design screams 3 things IMO:
1) It's an Xbox.
2) It's powerful. This SOC is freaking huge.
2) It's expensive. Either they will sell it for a lot to make a profit, or they will take a big hit on each console to make it competitive in the console space.
3) They are probably aiming for a 2026 release, considering the AMD timeline and the fact that they are using a 3nm node.

 

[viveks86]

Ha... DF is already speculating that Magnus would likely not be the only SoC and other OEMs could make their own variants of the GPU portion of the die, with potential higher or lower end versions. I thought this was unlikely to happen in the near future. Are we there already?

 

[HeisenbergFX4]

Ha... DF is already speculating that Magnus would likely not be the only SoC and other OEMs could make their own variants of the GPU portion of the die, with potential higher or lower end versions. I thought this was unlikely to happen in the near future. Are we there already?

I think its possible

 

[viveks86]

I think its possible

The split design makes a whole lot of sense now. Damnnn...

 

[Proelite]

Ha... DF is already speculating that Magnus would likely not be the only SoC and other OEMs could make their own variants of the GPU portion of the die, with potential higher or lower end versions. I thought this was unlikely to happen in the near future. Are we there already?

Rumor  Post in thread '(Grain of Salt) Moore's Law is Dead: PS6 SoC codenamed 'Orion', PS6 Portable SoC codenamed 'Canis''

Jul 30, 2025

"This week, Xbox announced it is actively building its next-generation lineup across console, handheld, PC, cloud, and accessories. As part of this, Xbox unveiled that it has entered into a strategic, multi-year partnership with AMD to co-engineer silicon across a portfolio of devices, including future first-party consoles and cloud."

"Lisa Su, Chair and Chief Executive Officer of AMD, shares how Xbox and AMD are building on two decades of partnership, innovation, and trust. AMD will extend its console work to design full roadmap of gaming-optimized chips combining the power...

• Proelite

Why wouldn't it be likely? The more unlikely is this bridging of discrete CPU and GPUs being a one off for Xbox.

Mixing and matching discrete GPUs and CPUs chiplets and bridging them via TSMC info lsi has been the goal of AMD since years ago.

If you move to this model, it removes a ton of r&d cost from companies that want a high performance APU. They just pick the CPU and GPU needed without worry about upfront cost of designing and APU nor economic scaling.

Sure you lose a bit of performance for your die area, and cooling is a bit more expensive, but those scale way better than monolithic die sizes.

 

[Lagspike_exe]

I think its possible

They could mandate specific CPU since the game logic relies on it as well as the GPU uarch. Then, allow OEMs to scale the CU/clock to hit different market segment.

They'd lose on economies of scale though but who knows what contracts TSMC is offering right now and how big of a hit that would be.

 

[viveks86]

Why wouldn't it be likely? The more unlikely is this bridging of discrete CPU and GPUs being a one off for Xbox.

Mixing and matching discrete GPUs and CPUs chiplets and bridging them via TSMC info lsi has been the goal of AMD since years ago.

If you move to this model, it removes a ton of r&d cost from companies that want a high performance APU. They just pick the CPU and GPU needed without worry about upfront cost nor scaling.

Makes sense. It just didn't hit me that it is already being designed with that in mind. Had this feeling that there was one more generation where OEMs don't get involved. Seems like that's not the case. Makes it even more idiotic to compare Magnus and Orion now.

 

[DrKeo]

If the next Xbox is an actual console, it makes sense that all Xbox devices will have the same CPU and only the GPU chiplet will be different. Maybe it will also apply to the mobile model, maybe that's why it has performance cores.

 

[Proelite]

Makes sense. It just didn't hit me that it is already being designed with that in mind. Had this feeling that there was one more generation where OEMs don't get involved. Seems like that's not the case. Makes it even more idiotic to compare Magnus and Orion now.

Rumor  Post in thread '(Grain of Salt) Moore's Law is Dead: PS6 SoC codenamed 'Orion', PS6 Portable SoC codenamed 'Canis''

Jul 30, 2025

30% more powerful coming out before ps6. MS is crazy this is nothing.

Comparing PS6 and Xbox will be meaningless.

• Proelite

• 
• 

Read my posts around this post. I already hinted at what's happening.

 

[Proelite]

If the next Xbox is an actual console, it makes sense that all Xbox devices will have the same CPU and only the GPU chiplet will be different. Maybe it will also apply to the mobile model, maybe that's why it has performance cores.

Why does the CPU have to be the same? As long as there is a CPU that supports bridging it'll end up in an Xbox OEM or 1st party. Maybe the ones with 3d v cache might be too complicated to cool along with a flatter GPU.

 

[Darsxx82]

Wccftech on what DF say about Magnus

The Magnus APU rumored to be powering the Xbox next-generation system will provide, with its unique design, a lot of flexibility that will allow Microsoft to scale an entire stack of systems that are more likely to compete with pre-built PCs rather than the PlayStation 6 and other traditional consoles.

In the latest episode of their weekly podcast, the tech experts at Digital Foundry went over the recently leaked Magnus APU said to power the next-generation Xbox, highlighting how different it is from the APUs powering the past and current generation PlayStation and Xbox consoles, not only in terms of size, but also in terms of design. Not only does this design, featuring separate CPU and GPU dies, steps away from the traditional monolithic design that balances power with costs, it will also be more powerful than anything we have seen so far, and supports a potential iterative approach to console design, giving Microsoft the ability to mix dies for more frequent upgrades than in traditional console cycles. This design could also be carried across multiple generations, falling in line with a "phases" approach, as opposed to the current "generations" approach, that Microsoft seems to be aiming for, considering recent developments.
Click to expand...

Although the Xbox next-generation system's Magnus APU looks expensive, its flexibility will potentially also impact the gaming PC market, Digital Fundry speculates. From AMD's point of view, being able to ship an entire Xbox PC with this condensed, likely cheaper design could be huge, allowing them to gain a bigger share in the market and provide a cheaper alternative to pre-built PCs, which are far from being affordable. As such, there's a real possibility that the concept of the "console war" between Sony and Microsoft will change forever, as the Xbox next-generation systems will likely compete with pre-built gaming PCs, rather than the PlayStation 6.
Click to expand...

Xbox Next Magnus APU Will Provide Flexibility for Scaling an Entire Stack of Systems That Will Likely Compete More With Pre-Built PCs Than the PlayStation 6

The design of the Xbox Next Magnus APU will provide enough flexibility for scaling a stack of systems that will likely compete with pre-built PCs rather than traditional consoles

wccftech.com

 

[DrKeo]

Why does the CPU have to be the same? As long as there is a CPU that supports bridging it'll end up in an Xbox OEM or 1st party. Maybe the ones with 3d v cache might be too complicated to cool along with a flatter GPU.

It depends on what strategy MS is taking with the new Xbox. If this chip is just one machine that runs the new windows-gaming-lite and any OEM can make whatever they want using PC parts and it is still "an Xbox" then the CPU doesn't really matter. It's basically a PC.

But if the next Xbox is a fixed series of pre-made chips that have their own Xbox API and OEM makers buy them in order to make white-label dedicated Xbox machine? The CPU should probably be the same, just like the XSX and XSS have very different parts but the CPU is the same with a tiny clock difference.

 

[Loxus]

Wccftech on what DF say about Magnus

Xbox Next Magnus APU Will Provide Flexibility for Scaling an Entire Stack of Systems That Will Likely Compete More With Pre-Built PCs Than the PlayStation 6

The design of the Xbox Next Magnus APU will provide enough flexibility for scaling a stack of systems that will likely compete with pre-built PCs rather than traditional consoles

wccftech.com

It may turn out to be pre built PCs with Xbox logo on it and not a console.

 

[THE:MILKMAN]

It may turn out to be pre built PCs with Xbox logo on it and not a console.

I'd imagine it would have to be at least a M-ATX size PC case as a console size SFF system around 10-15L would cost a fortune.

HeisenbergFX4 have you heard anything about Magnus casing/form-factor?

 

[SupremeJ]

It may turn out to be pre built PCs with Xbox logo on it and not a console.

If it can be turned on with an Xbox joypad and feels like a console, then for the average user, these will be consoles.

 

[HeisenbergFX4]

I'd imagine it would have to be at least a M-ATX size PC case as a console size SFF system around 10-15L would cost a fortune.

HeisenbergFX4 have you heard anything about Magnus casing/form-factor?

No clue about any form factors

I would think the actual console will be a pretty standard size but the OEMs might go crazy

 

[THE:MILKMAN]

No clue about any form factors

I would think the actual console will be a pretty standard size but the OEMs might go crazy

Interesting....So Microsoft could do something similar to Series X form-factor then? I'd imagine it would have to increase in volume quite a bit from the <7 litres though.

 

[HeisenbergFX4]

Interesting....So Microsoft could do something similar to Series X form-factor then? I'd imagine it would have to increase in volume quite a bit from the <7 litres though.

No clue in all honesty

 

[DrKeo]

It may turn out to be pre built PCs with Xbox logo on it and not a console.

I don't think it's going to be that considering they said it will have native Xbox BC and I doubt PC is getting that anytime soon.

My guess is that MS is going to make a family of chips and sell them to OEM companies. Just like GPUs - you have the Nvidia founders edition and lots of OEMs that use the same chips but have their own crazy designs.

 

[Loxus]

If it can be turned on with an Xbox joypad and feels like a console, then for the average user, these will be consoles.

Not if it can do more than gaming.