Predict: The Next Generation Console Tech

[erick]

Think about even if extra ram only added average cost of 5$ per console at the end of console cycle that would have been 5$*100million equalling to 500 million dollars of lost money(not even counting if the extra ram happened to cause more consoles to break and require warranty repairs).

What.

First, are there even any consoles that have made it to year 5 and not become profitable to manufacture?

Second, if discussing RAM, it loses its price the quickest out of all performance components, since it is the easiest to manufacture and the first technology to take advantage of new process nodes due to it's simplicity compared to processors.

We are dealing with a 10 year life-cycle here, and its not like sales numbers pan out evenly over all the years. The first 2-3 years, when the manufacturer stands to make the most losses and console prices are high, are usually the ones that make the smallest sales. I mean, really small sales. By end of year 2 X360 was still around 5 million units sold. Microsoft itself said that expects X360 to become profitable to manufacture by 2007.

So by today we should have 5 million X360 consoles sold at loss, 50 million sold at even or at profit and more to come?

The following is a very simplified and non-reasearched example of why it makes sense to lose money in the first years at all, and why it makes sense to lose money on RAM:

Let's be very pessimistic and say that you will make a $50 loss on 8GB of GDDR5 (2014 prices, remember) on every $499 console sold. So, you sell 5 million in first 2 years. That's $250 million lost.

But then as prices come down, you hit even for the year 3, and incur no extra losses.

By the beginning of the 4th year you launch a new SKU that makes a $50 profit at a $249 price tag. In the years 4 and 5, you sell 30 million consoles due to the reduced price tag, (and because your console's games run smoother and look better due to 8GB RAM compared to the competition, which is clearly getting long in the tooth with just 4GB and where developers have to devise clever ways to overcome the hardware handicap... some do an acceptable job that gets "almost the same, with pop-in and reduced detail here and there", but most won't even bother and just tune down texture detail until it fits, meaning your console's version is the one recommended by all critics and gamers alike for almost every multiplatform AAA game).

You make $1,500 million, and cover all the losses taken during the initial 2 years, with plenty left over.

RAM is now the cheapest performance component of your system at $6.25 for the total of 8GB, as prices have been halved 3 times during the 5 years since launch. And that's just at the half-way mark of your console's life cycle.

 

[liolio]

Damned which manufacturer is planning on selling 5 millions units in 2 years? I didn't even read further.

 

[Shifty Geezer]

First, are there even any consoles that have made it to year 5 and not become profitable to manufacture?

No console is profitable to manufacture. It's profitable only when you sell it for more than it cost to manufacture and distribute. An $800 BOM console can still be profitable is sold for $850, and a $99 BOM console can lose money if you have to sell it at $79 to have anyone buy it. They are sold are landmark pricepoints $399, $399, $249, etc. An added $15 cost means you either have to delay when you reach a pricepoint, risking your competitor getting in before you, or you lose more money at each pricepoint. At no point can the cost of a component be ignored as immaterial. This is why even super-cheap components like an extra USB port are cut from designs.

By end of year 2 X360 was still around 5 million units sold. Microsoft itself said that expects X360 to become profitable to manufacture by 2007.

So by today we should have 5 million X360 consoles sold at loss, 50 million sold at even or at profit and more to come?

Okay, not wanting to sound rude, but you're really showing your ignorance of the console space now. XB360 sold over 15 million in its first two years

The following is a very simplified and non-reasearched example of why it makes sense to lose money in the first years at all, and why it makes sense to lose money on RAM:

Right. Now look at the real console market and how loss leading hardware with significant specs has been a financial disaster, and then make a case that adding even more loss to the hardware is a Good Idea. XB was a fabulous loss-leader, and it was a financial hole. PS3 was a significant loss leader, and Sony likely won't make back their money on this generation. 360 was a loss-leader, and the returns on this gen aren't going to be amazing. In fact they could have done with spending an extra $15 per unit on better cooling, instead of enduring the $billion loss for the RROD fiasco, let alone lose even more by including more RAM. The console space is damned hard to do well in. Your oversimplified view is very well reasoned, very academic, and not very practical.

Edit: I'll go further:

Let's be very pessimistic and say that you will make a $50 loss on 8GB of GDDR5 (2014 prices, remember) on every $499 console sold. So, you sell 5 million in first 2 years. That's $250 million lost.

Let's be really realistic and factoring in all the other components that are already incuring a $150 loss at a $399 launch price and you add your $50 extra loss for an extra 4GBs GDDR5. That's $200 loss per unit. You sell 15 million units, same as other successful consoles, for a grand loss of $450 million, hoping to make a lot of that up with software sales and licensing deals because that's where the money comes from in this game.

But then as prices come down, you hit even for the year 3, and incur no extra losses.

At the same price-point. Or you drop the price for more loss-leading. Or you wait at a higher price and get undercut by your cheaper rival. Or you wait at your higher price to see some new device like an iPad come along that competes for people's disposable income.

RAM is now the cheapest performance component of your system at $6.25 for the total of 8GB, as prices have been halved 3 times during the 5 years since launch. And that's just at the half-way mark of your console's life cycle.

That's about the only bit you've got right, except you miss possible costs of implementing extra RAM, and the fact that your rival spent that money not on extra RAM but on securing services and iOS interaces and goodness-knows-what-else comes along. Your marketeers tell the world that your console has twice the RAM of the rival, but Joe Public can't see much difference because COD looks the same on both boxes. All their mates have bought RivalBox because they can play it wirelessly on their tablets, so Joe Gamer also buys RivalBox. Still at least you'll have a couple of exclusive flagship titles to show off what 8GBs could have achieved if it was used, and those couple of games sell well to your install base, making them profitable, but you are nowhere near recovering the losses on RAM.

There are lots of hypothetical scenarios that can be played through where more RAM is good for the platform or not. All I want from this is for you to recognise that more RAM isn't the only option to invest in, isn't the certain benefit you believe it to be, and that the console space is already occupied with lossy hardware that loses companies money and they have to balance a load of factors that make a difference. Then we might be able to get back to discussing the next-gen hardware specs instead of the business issues of the console market.

 

[function]

First, are there even any consoles that have made it to year 5 and not become profitable to manufacture?

Console history is strewn with the wreckage of platforms that never made it to 5 years and where the hardware (even the entire business) never became profitable. Excluding the failures (basically what you're doing) is not a great basis for your argument.

Second, if discussing RAM, it loses its price the quickest out of all performance components

You keep saying this, but how are you measuring performance?

We are dealing with a 10 year life-cycle here, and its not like sales numbers pan out evenly over all the years. The first 2-3 years, when the manufacturer stands to make the most losses and console prices are high, are usually the ones that make the smallest sales. I mean, really small sales. By end of year 2 X360 was still around 5 million units sold. Microsoft itself said that expects X360 to become profitable to manufacture by 2007.

You might want to check that 5 million figure.

The last few years - the twilight years - are usually the ones where you make the least sales.

So by today we should have 5 million X360 consoles sold at loss, 50 million sold at even or at profit and more to come?

You might want to check that 5 million figure.

 

[erick]

Sorry, my mistake, I was basing my numbers on US sales only - looking at the graph.

As for memory prices, look here. An example:

In November 2005, 1GB DDR2 cost $119.
In February 2009, that is 26 months later, 1GB DDR2 cost $11 (as part of a 2x1GB DDR2 kit at $21.99).

That is a price reduction of almost 11x over 26 months. In the 26 month period the price was halved 3 times. Questions?

 

[Shifty Geezer]

That is a price reduction of almost 11x over 26 months. In the 26 month period the price was halved 3 times. Questions?

Why haven't you responded to any of the other points raised? You keep saying more RAM will result in a better console, and the RAM price will drop dramatically. Not one person here disagrees with you on these points. So why do you keep repeating them ad-nauseum without addressing a single one of the other factors raised, other than brushing them aside with a simple "they are shortsighted" response?

 

[erick]

There are lots of hypothetical scenarios that can be played through where more RAM is good for the platform or not. All I want from this is for you to recognise that more RAM isn't the only option to invest in, isn't the certain benefit you believe it to be, and that the console space is already occupied with lossy hardware that loses companies money and they have to balance a load of factors that make a difference. Then we might be able to get back to discussing the next-gen hardware specs instead of the business issues of the console market.

Shifty, recognized. Naturally 8GB of VRAM is not the only option to invest in. It's just the cheapest and best way to futureproof the graphics of your system. If futureproofing for the period of 2015 - 2025 is not important, and you're willing to take a chance and try something else, sure.

 

[function]

Sorry, my mistake, I was basing my numbers on US sales only - looking at the graph.

As for memory prices, look here. An example:

In November 2005, 1GB DDR2 cost $119.
In February 2009, that is 26 months later, 1GB DDR2 cost $11 (as part of a 2x1GB DDR2 kit at $21.99).

That is a price reduction of almost 11x over 26 months. In the 26 month period the price was halved 3 times. Questions?

November 2005 to February 2009 is 39 months.

 

[Gerry]

November 2005 to February 2009 is 39 months.

He's getting closer. He was 200% off with his Xbox stat. Now he's only 50% off with this one.

 

[hoho]

You do understand that memory bandwidth can be as wide as you want but once you have to go after your texture data in the slower system RAM or in case of consoles, the HDD/optical drive, the performance drops to unplayable?

Depends how well your streaming system works. Also, just adding extra RAM won't help nearly as much for IQ if you don't pair it up with higher bandwidth and that costs an arm and a leg compared to just using bigger memory chips.

And how does computing and bandwidth tie into this? As long as we are dealing with texture sizes less than 320MB the 8800 GTS 320MB computes just fine and shows exactly as good numbers as its bigger brother.

The amount of unique data you can see on screen per-frame is pretty much only limited by vram bandwidth. Capacity is nice too but not nearly as important or useful, at least providing you have enough for getting the basics to work ( = as much as competitors).

This is what I've tried to explain from the start - bandwidth and computing are very, very nice but once you hit the physical barrier of not having enough VRAM you basically have to choose between reducing image quality or having unplayable fps.

True but that will NEVER happen in a console world where you have clearly defined hardware. How many games had significantly worse IQ on PS2 than XB considerng the latter had tons more RAM? How big part of IQ difference on PS3 vs XB360 was caused by lack of RAM* and not by differently allocated performance**?

*) IIRC PS3 reserved more RAM to OS and it also has two memory banks instead of one unified chunk.
**) PS3 having generally weaker GPU and stronger CPU

Therefore it makes a lot of sense not to skimp on having twice the RAM for something like1/30th of the cost of the whole box if you want it to last!

Again, let's assume MS or Sony had doubled the amount of RAM their machine had vs the competitor. Do you think they could have had significantly higher IQ as a result and not be bottlenecked for having to stream twice as much data through their GPU?

It completely negates the multiplatform point - if one platform is more powerful, it will be the lead platform.

This was addressed before but it's quite funny how THE most powerful platform, PC, hasn't been the lead platform for pretty much the entire lifespan of latest gen consoles

Laa-Yosh, the whole idea of having enough RAM is so you can push the need to reduce IQ as far forward into the future as possible. Obviously the platform which is forced to reduce IQ first will suffer greatly at this disadvantage.

Texture resolution is only a tiny part of IQ. Shading is much more important and bigger textures don't help you much there.

On the previous page someone said that bandwidth is more important than RAM size, again. You'd think that I put that dog to rest with the 8800 GTS 320MB vs 8800 GTS 640MB example.

You were using a deeply flawed example. Consoles are not comparable to PCs and as has been repeated several times already simply having more ram without increasing bandwidth and processing power will not give you higher IQ. Only reason why your 320 vs 640M example works is that no one expects any random PC GPU to handle anything thrown at it.

Second, if discussing RAM, it loses its price the quickest out of all performance components, since it is the easiest to manufacture and the first technology to take advantage of new process nodes due to it's simplicity compared to processors.

While the price of the chip is dropping the cost of providing N-bit wide channel to it is pretty much constant and overall costs significantly more than the memory chip itself (more complex motherboard, memory controller, power usage, ...) so again all you could get is just more RAM at same bandwidth for less loading screens and less need for streaming but you won't get much better IQ.

In November 2005, 1GB DDR2 cost $119.
In February 2009, that is 26 months later, 1GB DDR2 cost $11 (as part of a 2x1GB DDR2 kit at $21.99).

How nice of you to bring up a clearly undercut price point. In the next ~22 months the cheapest RAM was still significantly more expensive than that DDR2 there.
Also, in the same list there was a 512MB SDR for $39 or 1G for $78 on sale on August 2003. That doesn't quite match your nice prediction of price reduction

It's just the cheapest and best way to futureproof the graphics of your system

Find me a game that is GPU computing-limited* on a 560GTX and compare 1G vs 2G cards. I'd love to see the IQ and performance difference between them when you crank up the texture resolutions.

*) because that's what is really limiting game IQ and that's what gets botttlenecked first on a console. Anecdotical "evidence" with a game that doesn't have any kind of half-decent streaming is not an example of what happens in a real world on consoles.

 

[Shifty Geezer]

I'm not entirely sure what you mean by 'futureproofing'. The consoles exist to sell and make money. Long legs helps with that, but long legs doesn't necessitate high performance. PS2 is the longest selling console and has a pitiful 32MBs. As I say, you cannot futureproof technology for ten years. The consoles need to be relevant for a few years of launch, then they drop behind the curve but ride their software and experience for a few years, and then they drop down to being cheap CE devices, well behind the tech curve but offering a cheap platform for entertainment with a new flagship platform being released to offer something new-and-improved. If some new tech comes out that stretches what your current platform can do (like Kinect), you make do the best you can until the next generation. Any console releasing in 2015 will look terrible in specs compared to cheap deivces in 2025, even with several hundred bucks of RAM on launch day!

Choosing 8GBs over 4 is a cheap option long term dollar wise, but it's quite possibly not the best ROI to make a console a success. Hence this thread should be speculating types of RAM, probable amounts given range of options, etc., and not trying to convince that a particular memory amount or clock speed or anything else is going to be making an appearance.

 

[Laa-Yosh]

It's just the cheapest and best way to futureproof the graphics of your system.

You keep repeating that theory but fail to supply any arguments.

Several of us have listed reasons why the lack of RAM can be compensated relatively easily with better software technology; but lack of a fast background storage or general programmability or fast internal buses are far more important weaknesses, that can't be dealt with using more advanced software architecture.

Then there are the financial aspects; a serious lack of "futureproof graphics" has not stopped the Wii in seriously outselling both HD consoles and making more money than the two combined together. Why the hell should the vendors care about graphics at all?

 

[Gubbi]

A little memory/storage system speculation:

I expect MS to pursue a loss-leading model again since they have recurring revenue from Live Gold members to offset the costs. I expect two SKUs, one with and one without HDD. I also expect the same launch price points for these as the 360, $399 and $299, even though the dollar is 20% weaker than in october of 2005.

I expect a similar BOM for the premium as for the premium 360, $525 (according to iSupply). Around 25% of the BOM of the launch premium 360 was spent on storage (RAM, flash and HDD). This gives us $131 to play around with.

Current prices are:
$4 per GB of DRAM (spot market)
$1 per GB of Flash (spot market)
$100 per TB of 2.5" HDD (E-tail, Newegg)

For the Premium this would yield:
4GB DRAM ($16)
64 GB FLASH ($64)
500 GB HDD ($50)

Q: Why have 64GB FLASH when you have a HDD ?
A: Cache a full Bluray Disc.

For the 'Arcade':
4GB DRAM ($16)
64 GB FLASH ($64)
Total: $80

The Arcade is the bigger loss-leader, but DRAM and FLASH cost-reduces over time, the HDD does not (or very little).

If next gen launches late next year (before the holidays) I expect the above. If we get a late 2013 launch, double all capacities.

Cheers

 

[hoho]

Q: Why have 64GB FLASH when you have a HDD ?
A: Cache a full Bluray Disc.

Good in theory but in practice I would think most players get bored of a game and switch to another one before you finish the buffering. In a regular linear game you could get by with much less than 64G by only buffering whatever gets loaded in upcoming loading screen

 

[Gubbi]

Good in theory but in practice I would think most players get bored of a game and switch to another one before you finish the buffering. In a regular linear game you could get by with much less than 64G by only buffering whatever gets loaded in upcoming loading screen

Few games will be a full Bluray disc (at least initially). So your 50GB caching area could cache multiple games (the same way the 360 caches up to 3 different games' content on the HDD now).

Cheers

 

[function]

Q: Why have 64GB FLASH when you have a HDD ?
A: Cache a full Bluray Disc.

I guess MS would have the option of limiting usable disk space in a similar manner to the 360. This might not go down well, but it could allow them to limit the size of a full-disk cache.

 

[Tahir2]

Hi Gubbi

I would think a smart cache system like that Intel promotes with its Z68 motherboard would be more useful than the brute force attempt of just caching a whole Blu-ray disc. In any case whilst load times do decrease they still take a while even on an SSD.

As to erick's post it would imho be wiser for MS and Sony to increase cache on their processor and GPU rather than waste money on oodles of RAM. However I know erick - he won't listen

The opinion that 8GB of RAM is the cheapest and best way to increase performance is just that, an opinion based on guesses that are not very educated (the numbers you posted erick were incorrect.)

 

[AlNom]

I expect a similar BOM for the premium as for the premium 360, $525 (according to iSupply).

What's the price difference between a DVD drive and a blu ray one?

 

[Squilliam]

Oi Gubbi I would preferably estimate 1/2 that cost for Arcades simply because you simply dont need much more than 4GB for your basic SKU.

 

[tunafish]

Yes, you need to find a balance, but does 8GB of GDDR5 at 224GB/s sound bandwidth starved? I don't think so.

How the ¤%&# do you plan to get 224GB/s out of a console? Seriously?

Looking for something that would ship in the 2013/14 timeframe, the most likely value I can think of is around 70GB/s (for 128bit 1100ishMHz), and the absolute top end is about 150GB/s, if they for some weird reason decide to go 256bit, and use a little more expensive ram chips.

Look at the past consoles, for ¤%&#s sake. No console has ever shipped with state of the art ram. The 360 has the bandwidth of a 9800pro, which shipped nearly 2.5 years before it.