Next Generation Hardware Speculation with a Technical Spin [2018]

[anexanhume]

You design the best console you can at the typical mainstream price. You then provide a better, higher-margin product to sell to a subset who are willing to pay more. You maximise your profits this way.

Think of it like a $500 8 GB phone and a $600 16 GB phone - the cost to add double the storage is nothing but some will pay more for it.

As Al* suggests, binning increases profitability. You can have a cost-effective solution with 4 disabled CUs, and then a premium for the lucky silicon that has no defects. If you don't have this higher tier, you can't charge extra for fully working chips and have to blow fuses to gimp them down to the baseline.

The trick will be getting the right top-end device, not stupidly expensive but different enough to be worth the added price. I have no idea what sort of premium gamers might pay for. If the baseline is $400, is $500 worth it for 10% more juice? 20% more juice? Would they go as high as $600 and what for?

Every other consumer product has different models at different prices for different consumers, from TVs to toasters. Why would the economics-of-consoles exist in its own space where consumers behave fundamentally differently?

The entire model of the console space is one box to define baseline specs that every game is designed to. Having ultimate SKUs with bigger or faster hard drives is nice, but they risk undermining their baseline performance if they want to gate some premium features behind a higher priced SKU. So, this is more than a phone that performs the same but has more storage.

People also don’t want to upgrade their TVs and toasters every 3 years. The console business model is fairly unique, and I think they risk the ire of consumers by bifurcation of their hardware too much. They already pissed off gamers last gen by forcing Kinect. I would think they approach this situation delicately.

 

[Shifty Geezer]

The historical model of the console space was limited to tech and economic viability of hardware that could not maintain BC with tools and titles. Every generation was a clean slate and had to be, in order to provide a power advance. As such, it had to be maintained a long time to recoup investment, and it wasn't at all easy to create a more powerful model. It was not possible to create a PS2++ at launch.

This also isn't about shorter generations and faster upgrade cycles. People don't need to upgrade their console every 3 years with this model. Six years is still the console life-cycle; it just would provide an optional higher tier at its start. There's also nothing forced about it, unlike Kinect. It's a choice, how much is gaming worth to you? Exactly like the mid-gen console choice that hasn't resulted in gamers crying over having an inferior experience with their standard console. Reason being, they're ordinary consumers. They know they can spend more to get a better experience, and make that personal choice.

Importantly, the console business isn't unique. People need to stop thinking of it like it is. Its historical paradigm was, but ultimately it's an entertainment business to provide games, and will follow exactly the same economic principles as any other business. Changes in technology will allow different hardware and software models. The businesses need to choose those that maximise profits. There's a new option now at launch to provide tiered performance SKUs which any and every other consumer product has, so it's definitely not something to be discounted out of a sense of tradition.

 

[Gemüsepizza]

The trick will be getting the right top-end device, not stupidly expensive but different enough to be worth the added price. I have no idea what sort of premium gamers might pay for. If the baseline is $400, is $500 worth it for 10% more juice? 20% more juice? Would they go as high as $600 and what for?

Idk, I'm not sure a higher specced version makes sense...

How many people would buy a device that's only 10-20% more powerful? For 25-50% more money?

-> Probably not many, imo. Even PS4 Pro and Xbox One aren't that popular, and those machines have 200-350 % more power than the base machines. Why would anyone care about 10-20%?

What meaningful things can you do with something like 10-20% more power?

-> Maybe 10-20% higher resolution? Is that meaningful? What else could they do?

What resources are devs willing to spend on features for only a fraction of their user base?

-> Probably not much. Some will probably increase a few values here and there, like resolution. I think an important factor is probably optimization. Increasing certain aspects of graphics requires optimization, to make sure that you can hit your performance target. I guess this is usually pretty time-consuming. And if they don't want to spend that much time / resources for optimization, they will choose conservative values for improvements. Which means that 10-20% power increase will yield even less graphic increases in reality.

I think if they make a premium console at launch, it could make more sense to focus on things like an Elite controller, or increased HDD size, maybe even an SSD or something like that. That way developers don't have to spend resources, and the value is immediately clear.

 

[Shifty Geezer]

Idk, I'm not sure a higher specced version makes sense...

How many people would buy a device that's only 10-20% more powerful? For 25-50% more money?

-> Probably not many, imo. Even PS4 Pro and Xbox One aren't that popular, and those machines have 200-350 % more power than the base machines. Why would anyone care about 10-20%?

You're missing the point - it's about profits.

Let's say only 10% of first year consumers are interested. If you can get $100 more profit from those consumers and sell 10 million consoles, 1 million of those will be the Elite and that's $100 million extra money. If you don't have that option, you lose $100 million. Unless you can gain more money elsewhere without the tiered hardware, why would you decline it?

Everything else you say applies exactly to the GPU market. Consumers are given a wide choice of how many ALUs and at what clocks. They are offered options of only $20 difference, with no noticeable change realistically, yet some people choose to spend a little more for a little faster. That's how consumers work. They don't all number crunch equations to get the best value purchase and justify every dollar spent. Offer a faster model and some people will buy it. There's also zero extra effort from the devs, at least in the simplest model of a higher tier box. Just with binned components and a bit more RAM, you can have better framerates and higher quality textures without doing anything different.

 

[AlNom]

You're missing the point - it's about profits.

Let's say only 10% of first year consumers are interested. If you can get $100 more profit from those consumers and sell 10 million consoles, 1 million of those will be the Elite and that's $100 million extra money. If you don't have that option, you lose $100 million. Unless you can gain more money elsewhere without the tiered hardware, why would you decline it?

Everything else you say applies exactly to the GPU market. Consumers are given a wide choice of how many ALUs and at what clocks. They are offered options of only $20 difference, with no noticeable change realistically, yet some people choose to spend a little more for a little faster. That's how consumers work. They don't all number crunch equations to get the best value purchase and justify every dollar spent. Offer a faster model and some people will buy it. There's also zero extra effort from the devs, at least in the simplest model of a higher tier box. Just with binned components and a bit more RAM, you can have better framerates and higher quality textures without doing anything different.

What might be interesting to try and sweeten the pot is also throwing in more RAM as MS also demonstrates that the extra RAM in Scorpio can be automagically used by unenchanted titles for caching (IIRC), so it becomes part of the general user experience. For example, let's say a hypothetical Elite/Pro/Super/Enchanted/Ludicrous SKU is more-or-less just the same as devkits - fully enabled APU + double the RAM.

At the end of the day, it may be niche, but so are the cherry picked APUs that function at full spec. xD

 

[anexanhume]

You're missing the point - it's about profits.

Let's say only 10% of first year consumers are interested. If you can get $100 more profit from those consumers and sell 10 million consoles, 1 million of those will be the Elite and that's $100 million extra money. If you don't have that option, you lose $100 million. Unless you can gain more money elsewhere without the tiered hardware, why would you decline it?

Everything else you say applies exactly to the GPU market. Consumers are given a wide choice of how many ALUs and at what clocks. They are offered options of only $20 difference, with no noticeable change realistically, yet some people choose to spend a little more for a little faster. That's how consumers work. They don't all number crunch equations to get the best value purchase and justify every dollar spent. Offer a faster model and some people will buy it. There's also zero extra effort from the devs, at least in the simplest model of a higher tier box. Just with binned components and a bit more RAM, you can have better framerates and higher quality textures without doing anything different.

Profits don’t come from the handles. These guys are in the razor blade business. Services, games, recurring revenue sources are where the money is made.

 

[Shifty Geezer]

That isn't an argument against a profitable SKU. That's, quite frankly, medieval legacy thinking. You aren't presenting logical justification against having a profitable box and are instead just going back to tradition - "we've never done it before."

If a business can make money on the hardware, why would they choose not to? I presented a realistic case where $100 million was on the table for a higher tier SKU. Why would you turn that down? Is it realistic to expect that your fans would refuse to buy your console because they are confused or felt snubbed?? "I like PlayStation, and I want a PlayStation 5, but because there's a high end model available, I'm going to buy Xbox." Doesn't sound plausible to me.

 

[AlNom]

It's a tough situation all around since even 2x Scorpio won't buy us much headroom for true 4k and appreciable improvements to shading simultaneously (IMO), but I haven't thought things through yet because allergies.

(Devs are still largely targeting ~ 1080p @ 1.84TF)

3x3 SSAA of XO titles would put a baseline @ just under 12TF. That'd be a cute thing. They'd probably have to only do forced AF for Scorpio-enchanted titles since there's not enough POW for 2x2 SSAA there.

 

[chris1515]

https://www.windowscentral.com/microsoft-plans-announce-next-xbox-next-year

While discussing the Xbox One's struggle to maintain traction in Korea, Xbox Business Lead for Asia, Jeremy Hinton, has reportedly expanded on Microsoft's next-generation console hardware. Speaking with major Korean gaming outlet, GameMeca, Hinton claims engineers are aiming to announce the console "next year." Below we have a translated excerpt from the article, provided to Windows Central by a professional translator.

"Lastly, Hinton made an announcement in regards to the surprise news about the next Xbox which was announced on the 10th during the Xbox showcase. Many engineers are working hard to announce the new console by next year and doing their best to announce it as soon as possible. However, he said if there is no awareness of crisis or clear policy to improve the Xbox market in the country, it should not be forgotten that releasing the new console could be an obstacle to success

He is in trouble

Microsoft US has reached out to say that Jeremy Hinton's comments were misrepresented and that Microsoft isn't yet officially announcing any timeframe for the next Xbox.

 

[AlNom]

Controlled leak?

 

[cheapchips]

I doubt they'd go for a higher spec/cherry picked APU for a SKU tier. There's not going to be an appriciable on screen difference between the base model and any higher spec given node/power constraints.

I think options will be the next console, next console + VR/AR headset and a cheap streaming mini console.

 

[Wynix]

I don't see a pro version launching along side the standard.
It rules out double dipping and as previously mentioned, they won't be able to offer a large enough leap to even be noticeable.
They need at least a jump to 5nm.
Because they can only offer one pro version per generation, they have to get it right.

 

[bitsandbytes]

I don't see a pro version launching along side the standard.
It rules out double dipping and as previously mentioned, they won't be able to offer a large enough leap to even be noticeable.
They need at least a jump to 5nm.
Because they can only offer one pro version per generation, they have to get it right.

So Scarlet being a family of products like a console, streamer and portable for example and Phil Spencer mentioning consoles just means there will be a mid-gen upgrade again 3/4 years in?

 

[steveOrino]

Importantly, the console business isn't unique.

I get where you are coming from but I disagree. While consoles have shifted to commodity hardware instead of one off designs like in the past its still a price sensitive industry with a large upfront cost. Comparing game consoles to phones (or really any other consumer computing device) is dubious because the base usage (text/talk) of a phone doesn't change much from $100 - $1000. Consoles are exclusively built to play games with a large portion of the software designed to get max utilization from the memory and ALUs and that alone is why I would disagree with your assessment that the console business isn't unique. I would also say your point about consumers wanting premium SKUs for higher end gaming already is filled by the PC because if high resolutions, super IQ and 16~8ms refresh rates really really matter to a consumer then I doubt a console is ever going to fit that niche.

 

[Silenti]

https://wccftech.com/exclusive-amd-navi-gpu-roadmap-cost-zen/

"Here is a fun fact: Vega was designed primarily for Apple and Navi is being designed for Sony – the PS5 to be precise."

 

[bitsandbytes]

https://wccftech.com/exclusive-amd-navi-gpu-roadmap-cost-zen/

"Here is a fun fact: Vega was designed primarily for Apple and Navi is being designed for Sony – the PS5 to be precise."

The source isn't the best but must admit I've asked elsewhere back in April whether AMD being almost completely silent on Navi could be because it is leading on PS5 and AMD don't want to give details away?

 

[anexanhume]

That isn't an argument against a profitable SKU. That's, quite frankly, medieval legacy thinking. You aren't presenting logical justification against having a profitable box and are instead just going back to tradition - "we've never done it before."

If a business can make money on the hardware, why would they choose not to? I presented a realistic case where $100 million was on the table for a higher tier SKU. Why would you turn that down? Is it realistic to expect that your fans would refuse to buy your console because they are confused or felt snubbed?? "I like PlayStation, and I want a PlayStation 5, but because there's a high end model available, I'm going to buy Xbox." Doesn't sound plausible to me.

Of course it is. I firmly believe that the console makers view early market share as crucial to establishing their platform. You can easily find quotes where the first to 10M or first to 20M is considered the winner. Once the market has settled out, then they can play with the supply and demand curve by adjusting their pricing.

MS outright acknowledges their box was too expensive to start this gen. BOMs estimated in the $470 range to construct, and PS4 was in a similar region. There’s no reason to suggest that will be different this gen when they both see the market in the US and Europe as very much in play.

 

[AzBat]

Like cheapchips and bitsandbites mentioned I could see a multi SKU launch but not a higher binned super APU on the high-end & a regular version at the normal end. Basically you'll get the normal generational leap console at $400-$500 but they will probably take a page from Nintendo's playbook & have a cheaper slower APU(ARM?) that will play all the same XB1 & XB2 games as the expensive XB2 console but at 1080p through backward compatibility and/or streaming. The cheaper APU could be put in separate portable & streaming devices or it could be put in a device that does both. $300 & $500 launch prices? I could see me buying the $300 model for portability & upgrading later to get 4K. They could later do an mid-gen 8K SKU.

Tommy McClain

 

[eastmen]

https://wccftech.com/exclusive-amd-navi-gpu-roadmap-cost-zen/

"Here is a fun fact: Vega was designed primarily for Apple and Navi is being designed for Sony – the PS5 to be precise."

The source isn't the best but must admit I've asked elsewhere back in April whether AMD being almost completely silent on Navi could be because it is leading on PS5 and AMD don't want to give details away?

Doesn't really track. So AMD is going to just dip out of the graphics card market for all of 2018 and some of 2019 dispite them having a 7nm Vega which would sell decently for little cost since they are already producing it for the professional market ?

Also AMD is focusing on Sony which is great for the both of them but I guess they are willing to loose MS as a partner since I doubt MS will take a current vega chip or wait for whatever is after navi in 2021 and get sony's left overs . The article kinda portrays AMD's graphics side as to small to handle the professional market , pc gaming and sony consoles so how would they handle xbox also.

 

[dobwal]

https://wccftech.com/exclusive-amd-navi-gpu-roadmap-cost-zen/

"Here is a fun fact: Vega was designed primarily for Apple and Navi is being designed for Sony – the PS5 to be precise."

That article makes no sense. AMD based consoles and AMD discrete graphic hardware basically share a common architecture. The semi custom business allows AMD to leverage its core designs outside of its traditional market which is basically PC hardware. It’s why it’s called “semi-custom”.

“Navi” has been AMD’s PC roadmap for years. If AMD wanted cards for PC gamers then taking the design used for consoles and creating discrete PC gaming cards would be a simple straight forward low cost endeavor. You don’t need a ton of engineers to pull that off. But PC gpus are becoming less about gaming and more about compute and datacenters. It would make more sense if leaders in the RTG groups felt they didn’t have enough man power to create cryptomining or AI monsters that competed readily with Nvidia.

Navi isn’t delayed because of Sony, if anything it’s delayed because of 7 nm. GCN based PC cards came well before GCN based consoles and it literally had no negative effect on the console market. There would be no point in delaying an architecture for a console.