• Happy holidays, folks! Thanks to each and every one of you for being part of the Tom's Hardware community!

AMD Announces Proposed Plan to Sell Singapore Facility

Status
Not open for further replies.


The profit on delivering console chips may not be that high. Remember that they have to support this chip for many years,
that costs money too.
 


Which is why Nvidia didn't bother with consoles this time around, unless you listen the fanATIc's.
 
Someone with deep pockets needs to buy AMD. I hope they can make it by themselves, but things haven't been looking so good the past 2 1/2 years.
 
Like AMD would jump into both consoles so they can take a loss. I'm sure they are making a little money on it, otherwise they wouldn't have jumped in. They may not be able to beat intel but they aren't stupid.
 


I do think they will make money from it, but it won't be the kind of money to keep them afloat, without having some other strong products too. Nvidia has been down this path, and i do believe that they chose not to do it, rather then being overlooked.
 


marketing usually costs money, AMD being in the consoles is more marketing than anything else. Miniscule profit + free marketing tool = win.
 


Nvidia reckoned it was next to impossible to recoup the development costs from the sale of console chips which is why they passed up this round and if that is true then it's going to be interesting to see how AMD get around that. Maybe we'll just see a lot more interviews from AMD employees slagging Nvidia off as that seems to be their current idea of "marketing".
 


These are custom SOCs that both Sony and Microsoft paid for with non-recurring engineering fees (NRE), support contracts, and part orders. They didn't lose money developing these so there is nothing to recoup.

NVidia wasn't even in the running this round if Sony and Microsoft wanted single chip solutions. NVidia only makes some low power ARM parts that are not powerful enough for consoles like these. They could have been in the running if they wanted an Intel CPU and discrete graphics like the original XBox.
 


I'm sure I saw an article a while back where Nvidia said that they were asked but declined thus taking themselves out of the running which is quite a bit different than they way you and others are portraying things.
 
First there was nothing stopping NV from making a 100watt box SOC. Wisely it seems now, they passed due to lacking console sales (see wiiu off 50% in sales already and VITA can't even break 7mil last I checked) and single digit margins for AMD. AMD says model for low double digits (that's less than 15% margin even then right? LOW DOUBLE?) but that may not happen for a while and even if sales are excellent (say 10mil for each of xbox1 & ps4) they won't make more than 100mil on them in the next 12 months (maybe 150mil? $100 or less chip at single digits is what like $10 or less per chip? 20million units total in a year won't make you rich). That won't get anywhere near the 600mil or so NV already makes yearly and this includes LOSING money on tegra still (I think T4/T4i/Kepler changes this shortly). So I'd say they passed for good reason. I think console sales will tank (already are) as the mobile stuff takes over. There are so many competitors coming this year or next (and shield, ouya etc say revving yearly, so kepler shield next year), that there will be a LOT of lost sales from casual gamers (the ones who buy in years 4-8 of a console's lifespan) causing the audience to only be hardcore which in turn will be a massive FAILURE. Run on? :)

Ouya, Shield, Gamepop, Wikipad, Madcatz, google, apple, and more I'm probably forgetting. They are all consoles. I suspect Google/Apple (or something coming soon) will be in an xbox1 sized box and run at 100w or so. Either of these two can customize a soc from NV etc. IF you don't have to run in a box the size of shield or a tablet I'd imagine 2-4 LOGAN's (kepler mobile coming Q1) would be very impressive and shouldn't cost more than $80-100. Add a $40HD (250-500GB?) a 100w PSU, 8-16GB of ram, and a fan/heatsink and you have a serious REAL console competitor for $250-300 which neither google nor apple need to make a dime on (no need for kinect crap or bluray, pure download boxes). Their appstore sales to these units (and ads sold on them etc) pays for it all allowing some aggressive hardware sold at cost. All they hope to do with these rev1 models is kill some console xmas sales etc and evolve them yearly to finish them off in the next 3-4 years. Kill the console sales, and you kill dev interest. No games=no console sales...see the vicious circle here? Consoles are dead and thank god, as mobile+PC will bring 4K gaming instead of sticking us in 1080p for 8yrs.

Take a good look at the GDC 2013 USA (2500+) and Euro devs (300+) and you can see they have little interest in anything console until some serious sales happen. They are focusing all efforts on mobile and PC (YAY finally!). Also note that now PC game sales EQUAL to total game sales on PS3+Xbox360. This is why they have turned, and also why they are waiting on console Q1 sales to see if they are DOA or not. Why make a game if you have to wait for a few years for sales to materialize on a console that may never hit even 30mil units sold each? Wasted efforts until they hit 10mil+ each. This is why MS announced 1Billion for games for xbox1 THEMSELVES. They have to get some games out there to hopefully spur sales since nobody else is planning games (only EA/Activision and even they have stated a wait and see approach for Q1-Kotick himself said he's watching for Wiiu like dropoff after xmas for the xbox1/ps4).

This is the classic Chicken/Egg syndrome, but this time they have a 1/2 dozen or more competitors not to mention phones/tablets revving yearly and they are a Billion unit+ audience already there to sell to and heading to over 2Billion sold yearly in the next 2yrs. Game devs go where the numbers are. Right now and for the next year (unless consoles prove otherwise Q1 etc) the numbers are ALL on mobile/PC.

Nothing is stopping qcom or ARM themselves from making this chip for the imagined console from google/apple either. Any of these can fund a dual quad A15 paired with massive gpus and slap them in 100w boxes especially if they wait for 20nm and jack them up to 3-4ghz with a fan/heatsink on them (this isn't just an NV ballgame, all can do this with google or apple funding the chip dev cost). Next xmas will be very interesting for anyone doing this with 20nm SOCS. I could see 2-4 quad cores (A15 quads) strapped to say 4 kepler SMX's per soc or something to up the gpu power instead of the SINGLE SMX coming on kepler mobile. Again, Arm/Qcom can be paid to do something similar with google/apple funding it all (heck apple can do it themselves with powervr backing it for gpu-whatever you get the point). Consoles are not alone this year :) This story will get worse every year going forward as they rev the competition yearly.
 
Nvidia was in Xbox1 and they switched to ATI for the 360, there was a whole blow up between MS and Nvidia over cost of the GPU once the Xbox1 got more mature, MS expected to get cheaper chips but Nvidia would not play ball.

Atleast for the MS; Nvidia would not even get a moment of consideration.

For Sony I have no idea, they probably were aware of the MS Nvidia problem and wrote some solid pricing contracts for the PS3 so not one got the short end of the stick, so relations should be ok.

If Nvidia gave both MS and Sony the cold shoulder, which I am sure they did not, it would be a huge error. Console developers will be living in an APU world trying to fully exploit the synergy between CPU and GPU, now certainly this will carry momentum into the PC world. By next year, there will probably be PC versions of console games that will be natively optimized for an AMD APU and leave Intel in the dust. Well that is what I hope for the sake of AMD.

AMD has the right product at the right time, and power consumption is only marginally important in a console.
 


AMD are used to selling their chips at a loss......they need to realize they cant do this forever.
 
Wow did that degrade fast, and welcome to the article of TL😀R Let me sum it up for everyone.

"AMD>Nvidia.. no Nvidia>AMD..."

On Topic, its probably time for AMD to leave the fab business and concentrate on their core. Since switching to TSMC they have had a lower scale manufacturing processes and better yields than when they were creating the chips and fabing them themselves.

 
No surprises really, they've been shrinking for years. It would appear no CEO can successfully turn them around to the rising power house they were 8 years ago.

A 15 year challenge that AMD has yet to crack in the UK is, using our biggest computer retailer as an example, out of 29 desktop computer brands & models only 4 are sold with an AMD processor. Similarly with laptops 15 / 123 different makes & models have AMD processors in.
In total AMD processors represent 12.5% of this retailers pre-built computer and laptop stock, I didn't include their Apple Mac Book stock in that quick equation which would throw the % down even further. While I can only speak for the UK, I suspect that this isn't limited to here.
 
Status
Not open for further replies.