Gabe Newell: Valve will release its own console-like PC

Consoles tend to be in the 5-10 million unit range by year two. At what threshold would a Steam box need to reach for third party pubs to jump on board with day one support?

Hard to say, but if they could move 1M units of a PC that's bought only by people who BUY games, they'd probably get very good support.
 
I suspect they will rev it annually or semi-annually, and it will just maintain cost, and there will be some policy about having to run on say the last 2 versions of the hardware.
It's pretty much what Apple does with idevices.

Not sure I'd want to buy a game console every 2 years. Maybe if they did a two-year refresh and asked for "certified" games to run on the two most recent generations, giving you a 4 year life for your console.

Kind of getting way ahead of the real info (almost nothing) at this point though.
 
I suspect they will rev it annually or semi-annually, and it will just maintain cost, and there will be some policy about having to run on say the last 2 versions of the hardware.
It's pretty much what Apple does with idevices.

Whats the point of that? The Steam box isn't going to limit linux gaming like consoles do for console gaming. The initial hardware being the minimum specs for linux based gaming for a number of years shouldn't stop others from continually upgrading their linux machine at their leisure or as much as their wallet will allow.

I doubt the Steam box in terms of hardware sales becomes a major source of revenue for Valve since any premium they create to grow profit margin will only encourage consumers to piece meal together cheaper alternatives or simply buy cheaper linux based machines with that rival the specs of the Steam box. Based on economy of scale Valve is not going to be able to compete with PC vendors in terms of hardware costs especially initially, if ever.

Making money off the hardware shouldn't be a huge motivation for Valve, nor should wasting money which constantly upgrading the console would encourage. Unless Valve plans to limit Steam to proprietary linux based hardware, iOS and its devices do not serve as a comparable model. Steam in general is more comparable to Android, where the service drives revenue not the hardware.
 
Last edited by a moderator:
Hard to say, but if they could move 1M units of a PC that's bought only by people who BUY games, they'd probably get very good support.

How attractive to third party devs are consoles with only 1 million users after a year or two on the market? There is no way that those million users will support every linux based title that appears on Steam.

A million users may encourage ports, but most pubs will probably limit 1st day release support for well known high selling franchises. New ips will probably have to show success on other more mature and proven platforms before being ported to linux. Therefore, Steam on linux won't be as attractive as Steam on PC due to a stronger available library and priority when it comes to release schedule.

Under that scenario Valve has to hope that Windows basically collaspes on itself with Windows 8 and that people will flock to the Steam on linux as a viable alternative.
 
I think the thinking is that Valve sells the console, but users are free to build their own linux machines to game on, if they want better hardware. So if devs start making linux games, the market is larger than just the valve console, especially if you port your game to OSX or Windows (more difficult). And I imagine they hope to lure a number of those Windows devs to Linux if they see there are a few million console owners waiting for games to play.

Who knows. Maybe the thing runs OS/2 Warp and allows complete hardware customization. Definitely interested in hearing more about it.

I still think the easiest way to lure gamers over to Linux would be for them to say, "Hey, you know that Counter-Strike/Day of Defeat/Left 4 Dead/Portal/Team Fortress/Half-Life sequel you've been waiting for? Well, it's coming out for Linux first, and it'll be out on all the other platforms in six months. If you don't want to wait, dual boot your PC to Linux (for free) or buy this nice little console we built where the game has been tested extensively."
 
I've been waiting half a decade to play Counter Strike and DoD actually.

I hope it succeeds. But I won't invest in a Linux based gaming PC unless the vast majority of third party PC games come with day one support on linux.

There's not much investing, unless you don't own a desktop or laptop computer at this point. Most of the early games will be low on resources (Valve games, indie games).
The trick though is you need one of the right graphics cards else you're screwed : I think the minimum is Sandy Bridge, Radeon 5000 or Geforce 8.
You would be able to play the same games on Windows XP or 7 with a geforce 6 or 7, or a Radeon 2000 to 4000 but you won't on Linux, because of AMD dropping driver support quite early, geforce 6/7 not supporting some fringe OpenGL feature etc.
 
How attractive to third party devs are consoles with only 1 million users after a year or two on the market? There is no way that those million users will support every linux based title that appears on Steam.

Not very but the steambox wouldn't be a console, PC margins are better and sales of that sort would give you a sizeable active buying population.

A million users may encourage ports, but most pubs will probably limit 1st day release support for well known high selling franchises. New ips will probably have to show success on other more mature and proven platforms before being ported to linux. Therefore, Steam on linux won't be as attractive as Steam on PC due to a stronger available library and priority when it comes to release schedule.

I agree, I think Valve would have to move a LOT (comparable to consoles) of units before a Linux box would demand day and date releases with Windows, let alone day and date with console releases.

Under that scenario Valve has to hope that Windows basically collaspes on itself with Windows 8 and that people will flock to the Steam on linux as a viable alternative.

I'm still not sold on the fact it's even a Linux box.
 
Whats the point of that? The Steam box isn't going to limit linux gaming like consoles do for console gaming. The initial hardware being the minimum specs for linux based gaming for a number of years shouldn't stop others from continually upgrading their linux machine at their leisure or as much as their wallet will allow.

I doubt the Steam box in terms of hardware sales becomes a major source of revenue for Valve since any premium they create to grow profit margin will only encourage consumers to piece meal together cheaper alternatives or simply buy cheaper linux based machines with that rival the specs of the Steam box. Based on economy of scale Valve is not going to be able to compete with PC vendors in terms of hardware costs especially initially, if ever.

Making money off the hardware shouldn't be a huge motivation for Valve, nor should wasting money which constantly upgrading the console would encourage. Unless Valve plans to limit Steam to proprietary linux based hardware, iOS and its devices do not serve as a comparable model. Steam in general is more comparable to Android, where the service drives revenue not the hardware.

Not sure I even think it's going to be a linux box.

As soon as you start upgrading components, it's no longer a steam box. All you have is a valve branded PC with potential compatibility problems possibly running linux.

The reason to rev it at some fixed frequency is so the hardware doesn't become stagnant.
Once you do that you need some policy, so that 2 years old hardware isn't obsolete.
 
I've been waiting half a decade to play Counter Strike and DoD actually.



There's not much investing, unless you don't own a desktop or laptop computer at this point. Most of the early games will be low on resources (Valve games, indie games).
The trick though is you need one of the right graphics cards else you're screwed : I think the minimum is Sandy Bridge, Radeon 5000 or Geforce 8.
You would be able to play the same games on Windows XP or 7 with a geforce 6 or 7, or a Radeon 2000 to 4000 but you won't on Linux, because of AMD dropping driver support quite early, geforce 6/7 not supporting some fringe OpenGL feature etc.

Last time I dropped a discrete gpu into my computer I think it was ati 9600 with BF2 and F.E.A.R being the last PC games I remember purchasing. I kind of floated away from PC gaming after going strictly laptops at home (outside of using a mac mini for an htpc) and mainly depending on console gaming.

I always felt like I would come back once discrete add-on gpu boards became a mainstream feature of laptops but that never happened.

So the ideal of a Steam box appeals to me only because I wouldn't mind getting back into PC gaming and having a linux box at home. But I will need it to be fully supported because the thought of waiting for an answer of when or even if a port will come to linux is in no way appealing.
 
Not sure I even think it's going to be a linux box.

As soon as you start upgrading components, it's no longer a steam box. All you have is a valve branded PC with potential compatibility problems possibly running linux.

The reason to rev it at some fixed frequency is so the hardware doesn't become stagnant.
Once you do that you need some policy, so that 2 years old hardware isn't obsolete.

The is no outright admission from Valve that the Steam box will be linux based but Newell still refer to the hardware as a PC. Furthermore, Nvidia, Valve and some devs getting together and spending a year working to put out the R310 drivers seems a little much for just putting out Steam on Linux given the market size of linux based gaming. But Valve seems to have big plans for Steam for Linux and it makes sense if the Steam Box is part of that plan.

The Steam Box won't be upgradeable but Steam for Linux will serve as a viable alternative with the the ability to upgrade. One of priority right now for Valve is getting Big Picture to work with Steam for Linux.

The fixed frequency is to provide devs with a fixed enviroment and gamers with assurances that their purchase will remain viable for an acceptable period of time. Apple devices can have relatively short period of turnover because one the profit margin on their hardware is high, their userbase thrives on short product cycles and their annual sales are almost an order of magnitude higher than consoles. Consoles' life cycles are relatively long because of their lack of margins and it take longer for the console market to mature.

If the Steam box is linux based and its games are compatible with Steam for Linux, there is no point in having a short upgrade cycle. As it makes the hardware more expensive as cost savings based on volume will be undermine and there is Steam for Linux for those that want to upgrade on a more frequent basis. If a console can last 7-10 years without upgrades, a linux based Steam box can last as long with an additional caveat of allowing those who want improved graphics to have at it without causing additional cost to Valve. Who would be yammering about the next Xbox or PS if you could simply upgrade current consoles yourself with PC parts? Most of us wouldn't care about the XB3 if you could shove a new gen AMD gpu in a 360 along with Ivy Bridge.
 
The is no outright admission from Valve that the Steam box will be linux based but Newell still refer to the hardware as a PC. Furthermore, Nvidia, Valve and some devs getting together and spending a year working to put out the R310 drivers seems a little much for just putting out Steam on Linux given the market size of linux based gaming. But Valve seems to have big plans for Steam for Linux and it makes sense if the Steam Box is part of that plan.

The market size is near zero, but there's chicken-and-egg, no games because there's no market and no market because there's no games. The technical problems have been strong too : linux is a moving target. Unlike Windows or even commercial Unix you can't run an old games, libraries versions have changed, whole subsystems are replaced etc. There were a few attempts a decade ago (Loki and Id games), those games were unplayable just a few years after release.
Terrible drivers is the other problem (also suffering from API and ABI instability)

But a whole decade later, it's maybe slightly more possible. There have been two popular long term Ubuntu versions, supported for three years : 8.04 and 10.04, now the 12.04 version is being supported for five years (i.e. security updates till 2017), also the whole OS is better quality overall and nominally usable by inexperimented users.

Linux market share is about 1%, if there are about 500 million PC computers that's about 5 millions. (in reality numbers are unknowable, and there are a lot of corporate PCs and Pentium 3 etc. which I try to account by giving a lowish number of PCs)

It's not big a market but it's a completely untapped one, which will grow. (more people will use linux or have it as main OS if they can game on it)


So the ideal of a Steam box appeals to me only because I wouldn't mind getting back into PC gaming and having a linux box at home. But I will need it to be fully supported because the thought of waiting for an answer of when or even if a port will come to linux is in no way appealing.

Linux users can be desperate for games, I basically have chess, SNES emulator (I don't even play them anymore) and flash games. And Quake Live, a glitchy version of a 1999 game.
There is Wine (running Windows games and apps on Linux) but you can only run cracked games known to run properly and you still suffer sound bugs or random issues.
I would gladly like to be able to even run a subset of five-year-old games and more recent games, and indie games, if it's all supported and at Steam prices.

With a Steam console (or box, with keyb/mouse as well as gamepad), well it would be like any console, you only get to play the subset of games that exist on it. You might not need it if you're invested into X360 etc. instead. Or you may be interested in it if it has PC-like games as well as console-like games. (plus doing media center and browsing, documents stuff)
If you want all AAA games and on day 1 though, yes you won't get that.

Linux gaming on self-built PC by users who maintain their own hardware and software, that may be an additional, "bonus" market. With Valve maybe barely breaking even from the additional sales vs support expenses, but they will be getting a captive market.
 
The reason to rev it at some fixed frequency is so the hardware doesn't become stagnant.
Once you do that you need some policy, so that 2 years old hardware isn't obsolete.

I would push this to at least three years to ensure that there will be at least one step down in lithography. Without any lithographic gains hardware advances at a given power draw is very modest these days. For the same reasons, a two step interval may actually make more sense.
 
Simply question: how many of these need to be sold to be viable? 1 million? 5? 10? I ask because any high end enthusiast is far more likely to build their own box than to buy a SteamBox as there is no way a hard ore gamer will game with specs like a Core i3 and mid-range GPU.

But what about the average person? Do they really want a large beige box in their living room? The major reason why consoles are so customized is to get a small foot print in an entertainment center.
 
That depends on how many people use steam on Linux. Steambox is like PSPGo, it is not a new platform, only a new form factor.
 
I ask because any high end enthusiast is far more likely to build their own box than to buy a SteamBox as there is no way a hard ore gamer will game with specs like a Core i3 and mid-range GPU.

Right there you're asking the wrong question about the wrong people. This isn't about targeting the "enthusiast" who likely turns their nose up at the PS3/X360 due to them being "inferior machines."

But what about the average person? Do they really want a large beige box in their living room? The major reason why consoles are so customized is to get a small foot print in an entertainment center.

Here, you are getting closer to the mark, but still manage to miss it entirely.

Would you consider the original Xbox a "large beige box"? Because the original Xbox was basically just a slightly above average PC at the time.

You can build a dead sexy HTPC in a small sleek attractive box that can play games right now. There's absolutely no need for anything large or beige. It doesn't even have to be rectangular if they don't want it to be. Until we know the level of performance they are targetting, we won't know what their price target is or what hardware constraints they'll have to work with.

Even going for a large powerful box with say a Core i7 and a GTX 680 or Radeon 7970 could fit into a relatively small and attractive HTPC box. Cooling might be a little loud but it's certainly doable. In fact, I have one that I cart around in a backpack to do LAN gaming at a friend's house.

So making it attractive in the living room isn't a problem no matter what hardware is used. The question comes down to what level of performance are they targetting at what level of cost? Because cost is going to be potentially very important if they want to sell to a large audience. And cost is going to limit what you can do on the hardware and design front.

Then again a living room Steam "console" only has to target 720p or 1080p. So the hardware bar isn't set terribly high.

Regards,
SB
 
Simply question: how many of these need to be sold to be viable?
Quite possibly 1. If it's not a closed platform that needs devs to target it specifically, then it doesn't need a minimum install base. It'll be a 'DVD player' for Steam games. Microsoft would love that model for their OS!
 
Quite possibly 1. If it's not a closed platform that needs devs to target it specifically, then it doesn't need a minimum install base.

I am expect some kind of certification process if you want to see your game on SteamBox. Depending on how this is calculated you would need some more systems to cover the cost. But I think the number of systems that need to be sold is more cirtical if you look at it from Valves side. The need to cover development and ongoing support costs.

It'll be a 'DVD player' for Steam games. Microsoft would love that model for their OS!

They have this model already. It is called Windows Media Center.
 
I don't think that is what shifty means. More like the valve box will be, just like dvd players, a box that can come in many variaties. But one thing will be certain: If you put in the dvd (steam game) it will work.
 
They have this model already. It is called Windows Media Center.
No, for games systems adopted in the living room. We've had a few stabs at this idea before with MSX and 3DO hoping to be open standards supported by different hardware manufacturers. MS have wanted this before, but without a hardware vendor really behind the idea, they've ended up with XBox as the living room platform and Windows on the desktop and the game platform fractured. When Windows is on the compact living room box as well as the PC, running the same games across devices, on hardware from all sorts of other companies, MS will be happy. However, if Valve make that push and produce a mass-market user-friendly TV box that doesn't ruun Windows but which spawns an open platform, MS will be mighty miffed (same if Google do that). I think that's very low probability, but it's definitely what MS want and seemingly Valve too - Valve would much rather have other HW companies worry about margins and they just sell software.
 
Back
Top