It would be nice to see the rest of the specs, just for a real comparison. My guess is the rest of the system won't be quite up to par with the PS4/One, especially since the manufacturer, the retailer, AND Valve will be getting a cut of the sales.  As far as retailers go, don't expect them to take the miniscule profit they do on console HW, since they won't be seeing any sales/profit after the initial purchase.

Of course, I don't see this doing very well. If it does okay, it will definitely be a very niche section of the market. The average console gamer isn't going to go for this since they will have probably already made their mind up to go Nintendo, Sony, or MS. Avid PC gamers will most likely have PCs that are at least on par with these specs. So really, they are aiming at console gamers that have no interest in what either the Big 3 are offering.

There are also two main things that I think are going to make this a hard sell. One, the controller looks atrocious, and I think it is really going to turn people off when they see it on the box. Especially those who are used to the PS3/PS4 and 360/One controllers. And two, your average consumer is going to be really confused in a year or two, when there are 3-4 different companies making different Steam Machines with different cases and specs. In the end, they'll most likely go with the cheaper, easier option of a console.

Ill take a wait and see approach. Software is key.

@Turkish:

Where did he claim to be from the US?
I mean, I'm not sure what your point is since Americans quite often have atrocious English,
(anybody studying English as a second language should have realized that by now)
but other than coming off as passive aggressive, what is the point of that conjecture?

And yeah, he chose to use a different name for a corporate product rather than blindly follow what PR directed him to. So? You SOMEHOW managed to infer the topic he was discussing, so I don't see how his terminology is any worse than the corporate conformist approach...? What does not blindly following corporate dictate, but choosing to use one's one folk terminology have to do with bad grammar? Product names have nothing to do with grammar, many of them would in fact be poor grammatic constructions if evaluated on that level.

FYI: It's somewhat incoherent to spell out "lawl" since that derives from an acronym "lol" meaning "laugh out loud", thus changing the spelling destroys the reference. I mean, if one is inclined to waste one's brain space attacking strangers for comprehensible spelling errors, one should be concerned about such things, right? You might want to catch up on English refresher lessons as well, you missed that the contraction of "it is" is "it's" NOT "its". Likewise, you should have written "how the fuck do people not know THAT" not "not know it". Isn't English lovely?

I don't see this thing even shipping for that price. By the time the build is complete, the shipping costs, the write off for failed builds and returns...

Totalling the individual parts costs and then assuming a total shipped cost isn't going to work.

R9 270 has 20 CUs at 925 MHz. PS4 has 18 CUs at 800 MHz. Both have same amount of ROPS. R9 270 doesn't share its memory bandwidth with a CPU, the PS4 does.

Let's wait and see how low level can you go on this STEAM OS. If something like Mantle can be used to code for games in Valve's new box, then I expect the hardware to be just as efficient as it is in consoles. In that scenario even the vanilla R9 270 will outpace the PS4.

Indeed Mantle should work fine, although not really any better than Windows.  It needs to be remembered that Valve is NOT standardizing on AMD here, in fact the first Steam Machines were announced with NVIDIA collaboration... Both AMD and NVIDIA GPUs will be going in Steam Boxes, so no SteamOS games can be written solely for one GPU vendor, they will need to be compatable with both, cutting into optimization budgets.  NVIDIA's G-Sync certainly looks to offer nice benefits, and so I see both vendors having a continued presence in this 'platform'.  SteamOS certainly will have lower overhead than Windows, although I'm not sure why it really would compared to at least PS4's OS, if things like Vent are to be available, as well as on-line manuals, browser-based match making forums and leaderboards, etc, SteamOS will need roughly the same heft as PS4 with PS4 in fact able to shed more extraneous OS cruft since SteamOS will otherwise be standards compatable Linux with OpenGL etc. 

"Enthusiastic" PC gamers are exactly what will make Steam Machine/OS work, and that's why I really see it as a development of PC gaming, more so than a new console per se.  Valve seems to be of the same opinion, given their controller design seems to be built around appealing to mouse+keyboard gamers in terms of accuracy, etc.  I don't see any impediment to running SteamOS games on SteamOS/otherLinux on other PCs (not Steam Machines) via dualboot/whatever, and I might guess that many PC gamers might like the idea of buying a low-mid end Steam Machine as gaming box rather than deal with upgrading their rig and all that.  But all it's doing really is cutting Microsoft Windows out of the picture, the hardware profile is still basically just as varied as Windows PCs.  Lower Linux overhead in performance (as well as licence cost) with good drivers and dual AMD Mantle/NVIDIA OpenGL graphics paths can yield better performane than the status quo in Windows but will just not have the optimization of the closed consoles.  Of course, in a year's time the $500 Steam Machine will be given more CPU/GPU grunt and that can counter the optimization advantage to a large extent... Although PS4/XBone are supposed to be able to quickly drop their prices to boot, so the 'same price point as consoles' will be a moving target as well, and optimization should certainly give a good advantage there in the end for exclusive titles, espeially on the stronger console platform (ahem).  At a certain point hardware advancements will overwhelm optimization even with price reductions on the closed consoles, but with simpler development costs and better profit margin (rather than losing money out of the gate) for PS4/XBone, I expect them to do a console refresh more quickly than last time, plausibly in a way that is 'backwards compatable' with the PS4/XBone generation.

Your point re: retailer interest is pretty spot on, as far as Valve's cut it probably will be zero: they are getting every Steam Machine owner to buy all their games thru Steam after all, with zero used game market.  They aren't making the units other than the controllers for which a reasonable margin is fine as that is supporting the entire platform.  Valve is also happy if people dual-boot regular PCs with Linux to play these games, no real difference to them, only the Steam Machine OEM partners.

It'd be great if Mantle could achieve console level of optimization, but it simply can't, although it can come closer than DX/GL can.
Just using Mantle does not allow for all the memory optimization that PS4 does with unified memory especially great for hybrid compute,
and there's a whole host of other optimizations possible with PS4 (and XBone) that are just not possible with open/variable HW config.
Not to mention PS4's disproportionate amount of ACE's will enable it to 'play above it's rating' in terms of similar size/cost GPUs,
that amount of ACEs will become norm on the HIGH END of PC GPUs but PS4 will be able to leverage that like the PS3's SPUs.
With both consoles using AMD, it seems likely that crossplatform games will have a shared level of optimiation that is Mantle compatable for the most part,
but as I wrote, Mantle and AMD is not the exclusive GPU for Steam Machine, it also uses NVIDIA, just like Windows that factor is not constant.

This is probably true... And more generally, it seem unrealistic to be comparing some no-name "IBuyPower" company assembling Linux microPCs

and just make a direct price comparison... People can (and some do) home-build PCs from components, and no-name companies offer lowest cost PCs,

but for some reason the vast majority of PCs sold are actually from major brands with warranties and customer service people trust.

If this is to be considered on a mass market basis, that is the context it needs to be understood in, and all those things add cost.

(the type of costs that Sony/MS consoles also need to account for)

In 2 years Steam Boxes will demolish the performance of the PS4/XB1 with newer GPUs and the OS will likely be as user friendly as either the PS4/XB1 to boot.

Then things will get interesting.

meh. of course more expensive hardware can out power lower price hardware.
in two years PS4 may well be at $250. will same price Steam Boxes really demolish it?
is PC GPU and CPU designs really going to get that much more efficient in power/transistor count over that time?
Architecture doesn't change that quick. Sony and MS will be taking advantage of the same process shrinks, etc,
and will be getting bigger volume discount on all the components than the motley crue of Steam Box builders.
With hardware optimization kicking in major for PS4 (and XBone) while the hardware agnostic Steam platform
can only access half the level of optimization by going for vendor level APIs like Mantle/NVIDIA GL (splitting their optimization budget) but not optimize for specific setups, I don't see the comparison level shifting significantly, and in fact PS4 in particular should even pull ahead in terms of the level of optimized exclusive games (not to mention those games being exclusive to begin with). But on a general level, especially for multiplatforms, it should at least stay on par which certainly is an opportunity for SteamOS to flourish as a software platform on it's other merits.