bbc article here showing how pc sales in 6th year successive fall
http://www.bbc.co.uk/news/technology-25771760
http://www.bbc.co.uk/news/technology-25771760
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
intel no profit this year
- Thread starter ~>Dg<~
- Start date
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,558
- Location
- Dormanstown.
LOL @ TITLE.
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,558
- Location
- Dormanstown.
Not at all.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,901
- Location
- Planet Earth
It is considering process nodes are costing more and more,and the fact they are spending at least a billion dollars to get more companies to use Bay Trail in the way of subsidies,ie, essentially reducing margins. That comes on the heels of another billion dollars spent on trying to get companies to make more Ultrabooks.
Remember,one new fab costs billions of dollars - even upgrading an existing one is expensive,and its not getting any cheaper.
The reason Intel could justify spending loads on process nodes,was decent profits,but if profits are falling,it means they need to make a bigger return on existing investments.
You can already see the effects of this:
http://hexus.net/business/news/components/64973-intel-indefinitely-postpones-opening-fab-42-arizona/
Moreover,desktop chips(and probably a reasonable number of laptop chips),will still be 22NM Haswell for remainder of the year. I expect Broadwell will mostly be ULV laptop,some special parts for Apple and Atom at some point.
Profits falling also means investors get a lower return too. Intel is considered a stable stock,with stable profits,etc. One way to keep the share price stable is to reduce the number in circulation.
Intel has been doing this for the last year:
http://www.theinquirer.net/inquirer/news/2229539/intel-will-issue-bonds-to-buy-back-its-own-shares
Edit!!
Also,it appears 14NM will be more optimised towards density than previous Intel nodes. I expect this is so they can get more chips from each wafer,and hence reduce the cost of each chip.
Second Edit!!
It seems,the Intel division which makes chips for phones,tablets and industrial devices has had a loss of $620 million in Q4 2013 alone. It seems the same division made a $495 million loss in Q4 2012.
I do wonder if the Atom division ever really made a profit?? That extra billion dollars in incentives is not helping.
Last edited:
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
Cat, profits will take into account R&D spending throughout the year, so that is still the profit being made after upgrading fabs, finishing the building on fab 42. Had they actually paid for and bought the equipment to fill the fab they would have probably dropped 2-3billion off the profit, but still been in heavy profit with all the spending.
Either way it's the general spend to make money model so many companies follow. Pay companies to use Atom over 5 years while you actually finally get around to making a top notch low end chip and they might finally be relevant.
The biggest mistake Intel has been making is sticking to quad core chips in mainstream.
When they dropped dual core into higher mainsteam pricing(the 2500-2700k type bracket) then a significant number of sales, particularly to gamers/other higher end desktop users, are performance chips. This drives software to use said performance, then those chips aren't seen as that fast with newer software so people buy faster chips, which once again drives faster software.
This just has not happened since, pretty much i5/i7 hit the market. The thing is Intel has just reduced the die size of every chip since then, there are increases in transistor count but they are so far below the capacity of die size they can make with a hefty profit.
The effects of this are pretty much two fold, they've stagnated performance to minimal yearly gains rather than drastic gains every couple years with more cores. This has drastically slowed software development(in games in particular) for higher performance which has drastically dropped demand to upgrade. Up to maybe 2008-09, every thread with "will a new cpu help my FPS in game X,y,z" was met with a "get a new chip, that is so last year". For the past 3-4 years the same threads get the answer "meh, if you have a Sandy you'll barely even notice a Haswell, buy an ssd or better gpu".
We end up with 181mm2 haswell and that is with a large gpu that didn't use to be there. If this chip was 250-300mm2, their high end chips can simply be 100W cpu, 80W gpu, or 100-130W both, most people don't use the igpu much at all. Considering the amount of that 181mm2 core that is gpu that almost no one uses, it's insane. We don't really encode any more due to online content delivery/size of blurays, etc, so quicksync is pretty useless compared to a decade ago. Real gamers are using discrete gpu's and barely touch what, 30-40% of the transistors they are buying from Intel.
Even if they were making non gpu 181mm2 cores we'd have had vastly increased cpu performance in mainstream, if they were making 250mm2 cores still, we'd have 8-12 cores already, in mainstream, Intel would have higher fab utilisation, performance would have enabled more/better software which would drive the industry to upgrade more frequently, making more sales.
Tiny cores with limited extra performance has stalled software, stalled the PC industry, stalled development, stalled upgrade, stalled Intel sales....... and made their fabs severely underutilised. The PC industry stagnating is almost entirely Intel's refusal to make vastly better chips.
Everyone I knew at school/college, they'd buy a new desktop every 2-3 years because their old one felt slow. When Intel stopped pushing the industry forward, software stopped pushing forward, and peoples 3 year old desktops don't seem slow any more.... so instead of upgrading every 2-3 years, my parents computer is now, well it's got my old Q6600 in it, I used to pass on upgraded systems every couple years max, but I don't upgrade that often either.
Intel quad cores at 45nm in 2009 were 296mm2, with no gpu. Today Intel is pushing 181mm2 cores, with a gpu taking up around 1/3rd of that.... Intel can and should be making 250-300mm2 chips in mainstream, they've made it vastly unattractive to upgrade by offering what used to be budget sized chips in the premium price bracket with very little extra performance per generation.
Why is it 5 years ago 300mm2 quad cores cost £200, and today a 40% smaller chip with gpu, where the non gpu size of the chip is closer to 1/3rd the size, costs the same, or even more? Is it any wonder people aren't buying?
How many people on the forum would upgrade from their 2500k's if they could get an 8 core new gen chip in the same £150-200 bracket? I sure as **** would, but another quad core that is 20% faster at that price.... no thanks.
Either way it's the general spend to make money model so many companies follow. Pay companies to use Atom over 5 years while you actually finally get around to making a top notch low end chip and they might finally be relevant.
The biggest mistake Intel has been making is sticking to quad core chips in mainstream.
When they dropped dual core into higher mainsteam pricing(the 2500-2700k type bracket) then a significant number of sales, particularly to gamers/other higher end desktop users, are performance chips. This drives software to use said performance, then those chips aren't seen as that fast with newer software so people buy faster chips, which once again drives faster software.
This just has not happened since, pretty much i5/i7 hit the market. The thing is Intel has just reduced the die size of every chip since then, there are increases in transistor count but they are so far below the capacity of die size they can make with a hefty profit.
The effects of this are pretty much two fold, they've stagnated performance to minimal yearly gains rather than drastic gains every couple years with more cores. This has drastically slowed software development(in games in particular) for higher performance which has drastically dropped demand to upgrade. Up to maybe 2008-09, every thread with "will a new cpu help my FPS in game X,y,z" was met with a "get a new chip, that is so last year". For the past 3-4 years the same threads get the answer "meh, if you have a Sandy you'll barely even notice a Haswell, buy an ssd or better gpu".
We end up with 181mm2 haswell and that is with a large gpu that didn't use to be there. If this chip was 250-300mm2, their high end chips can simply be 100W cpu, 80W gpu, or 100-130W both, most people don't use the igpu much at all. Considering the amount of that 181mm2 core that is gpu that almost no one uses, it's insane. We don't really encode any more due to online content delivery/size of blurays, etc, so quicksync is pretty useless compared to a decade ago. Real gamers are using discrete gpu's and barely touch what, 30-40% of the transistors they are buying from Intel.
Even if they were making non gpu 181mm2 cores we'd have had vastly increased cpu performance in mainstream, if they were making 250mm2 cores still, we'd have 8-12 cores already, in mainstream, Intel would have higher fab utilisation, performance would have enabled more/better software which would drive the industry to upgrade more frequently, making more sales.
Tiny cores with limited extra performance has stalled software, stalled the PC industry, stalled development, stalled upgrade, stalled Intel sales....... and made their fabs severely underutilised. The PC industry stagnating is almost entirely Intel's refusal to make vastly better chips.
Everyone I knew at school/college, they'd buy a new desktop every 2-3 years because their old one felt slow. When Intel stopped pushing the industry forward, software stopped pushing forward, and peoples 3 year old desktops don't seem slow any more.... so instead of upgrading every 2-3 years, my parents computer is now, well it's got my old Q6600 in it, I used to pass on upgraded systems every couple years max, but I don't upgrade that often either.
Intel quad cores at 45nm in 2009 were 296mm2, with no gpu. Today Intel is pushing 181mm2 cores, with a gpu taking up around 1/3rd of that.... Intel can and should be making 250-300mm2 chips in mainstream, they've made it vastly unattractive to upgrade by offering what used to be budget sized chips in the premium price bracket with very little extra performance per generation.
Why is it 5 years ago 300mm2 quad cores cost £200, and today a 40% smaller chip with gpu, where the non gpu size of the chip is closer to 1/3rd the size, costs the same, or even more? Is it any wonder people aren't buying?
How many people on the forum would upgrade from their 2500k's if they could get an 8 core new gen chip in the same £150-200 bracket? I sure as **** would, but another quad core that is 20% faster at that price.... no thanks.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,901
- Location
- Planet Earth
The problem is that the division which handles tablets,phones and industrial(basically Atom) computing has lost billions of dollars since 2011/2012 AFAIK. I think it is over $3 billion so far. That does not include the latest $1 billion incentives this year.
Intel cannot keep making losses like that,and no wonder they are making their higher end chips smaller - they are trying to subsidise the low end. If they keep doing that for another 5 years,companies will walk over Intel,because by then ARM based SOCs will have reached the performance level for more intensive tasks and be good enough for them. With more and more stuff being shifted to the cloud(for the average computer user),I doubt the average user will even need a mega-CPU at home.
The thing is though it is the higher performance X86 laptop and desktop CPU market which is still a decent percentage of Intel profits(with servers). However,just as Intel is trying to move downwards to less powerful and more efficient chips,ARM based alternatives are slowly moving upwards too,and are increasingly starting to encroach onto the traditional Intel consumer markets,ie,laptops and the like. We are now seeing tablets and hybrid tablets eating into laptop sales,just like laptops ate into the desktop market.
The problem is Intel is not just fighting AMD now(with its eternal lack of resources),its everyone with a MIPS and ARM license. Companies like Qualcomm have many times the free cash in hand Intel has and are making profits on the chips in phones and tablets,not losses. Plus there are companies like Samsung and Apple who are increasingly going into the chip business. These companies have tens of billions of dollars in the war chest.
Intel is fully reliant on their process lead ATM. If they falter even once,it will be worrying and each process node is getting more and more expensive. This is why they are sticking with 22NM for as long as they can.
On top of this the companies like Mediatek and Chinese equivalents are starting to flood the lower end of the market too. We have started to see the big performance increases with the newer 64 bit ARM cores. Once these filter down to companies like Mediatek and the like,even Qualcomm(let alone Intel) will start feeling the heat. Why bother with a £250 tablet when a £100 will end up having 70% of the performance??
Its a race to the bottom.
Heck,even Intel is pushing more and more of its focus to commercial computing now,since the profit margins are much higher. IBM did this since they saw where the consumer market was going.
Intel cannot keep making losses like that,and no wonder they are making their higher end chips smaller - they are trying to subsidise the low end. If they keep doing that for another 5 years,companies will walk over Intel,because by then ARM based SOCs will have reached the performance level for more intensive tasks and be good enough for them. With more and more stuff being shifted to the cloud(for the average computer user),I doubt the average user will even need a mega-CPU at home.
The thing is though it is the higher performance X86 laptop and desktop CPU market which is still a decent percentage of Intel profits(with servers). However,just as Intel is trying to move downwards to less powerful and more efficient chips,ARM based alternatives are slowly moving upwards too,and are increasingly starting to encroach onto the traditional Intel consumer markets,ie,laptops and the like. We are now seeing tablets and hybrid tablets eating into laptop sales,just like laptops ate into the desktop market.
The problem is Intel is not just fighting AMD now(with its eternal lack of resources),its everyone with a MIPS and ARM license. Companies like Qualcomm have many times the free cash in hand Intel has and are making profits on the chips in phones and tablets,not losses. Plus there are companies like Samsung and Apple who are increasingly going into the chip business. These companies have tens of billions of dollars in the war chest.
Intel is fully reliant on their process lead ATM. If they falter even once,it will be worrying and each process node is getting more and more expensive. This is why they are sticking with 22NM for as long as they can.
On top of this the companies like Mediatek and Chinese equivalents are starting to flood the lower end of the market too. We have started to see the big performance increases with the newer 64 bit ARM cores. Once these filter down to companies like Mediatek and the like,even Qualcomm(let alone Intel) will start feeling the heat. Why bother with a £250 tablet when a £100 will end up having 70% of the performance??
Its a race to the bottom.
Heck,even Intel is pushing more and more of its focus to commercial computing now,since the profit margins are much higher. IBM did this since they saw where the consumer market was going.
Last edited:
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,901
- Location
- Planet Earth
Intel announces over 5000 job cuts this year:
http://money.cnn.com/2014/01/17/technology/intel-jobs/
http://money.cnn.com/2014/01/17/technology/intel-jobs/
Last edited:
Soldato
- Joined
- 6 Feb 2010
- Posts
- 14,583
If Mantle really works out on reducing CPU limitation in games, that would be bad news for Intel as well. Just imagine all the people that are still on 1st gen i5/i7 that are waiting to upgrade suddenly realise they don't need to anymore lol...
Soldato
- Joined
- 25 Sep 2009
- Posts
- 10,361
- Location
- Billericay, UK
Troll thread title. Can a mod please amend it.
Soldato
- Joined
- 21 Jul 2005
- Posts
- 21,259
- Location
- Officially least sunny location -Ronskistats
Great point!
I like this post drunkenmaster, but I think you may have the cause and effect the wrong way around. In the last 5-7 years the majority of PC users haven't needed to upgrade because everything they do is either through a web browser or on office software. Intel have noticed this and responded by going where the majority of the market is.
The fact is that most users don't need hundreds of GFLOPS at all, the software people use already has all the power it needs in a tiny mobile chip (or in a server somewhere). The minority of users who do need a powerful solution have simply parallelised existing (energy efficient) chips, i.e. clusters and/or GPUs. The market for single high performance CPUs has disappeared.