It almost surprises me how much Intel is losing in Mobile. How much are they losing per atom chip sold? Atom is certainly a decent chip, although its getting a bit old now, so I wonder how much market share they're gaining. That would really help put this in perspective. If they've gained a decent share of the market, it's certainly not too bad.
It's a total crap chip that they are giving away for free. I think we now know for fact why AMD's Mullins haven't been in about anything by now and everyone is using crap Z3735 chips.
"It's a total crap chip that they are giving away for free." Right, total crap chip, that is why the Atom 3775 scores 2863 in geekbench while the new Apple A8 scores 2881.
Have you even used something with one of the Atom in, like a Asus T100. It feels as fat as a core laptop.
What is total crap is what you have written, that all...
"Have you even used something with one of the Atom in, like a Asus T100. It feels as fat as a core laptop.
What is total crap is what you have written, that all..."
It is obvious that you have no idea what I am talking about. You better start reading a few more articles and google a few unknown names to you, like AMD and Mullins, before posting nonsense. Apple and AMD may have a common first letter in their names but, and this is going to shock you, they are two totally different companies.
So, first you are comparing a quad core x86 chip that has a turbo speed at 1.8 with a dual core ARM chip that has a 1.4GHz frequency. This is a huge facepalm.
you might learn something, but I don't believe so. From your comment I am pretty sure you are 12 years old, mentally at least, or you are just trolling.
Did you even look at what you posted in support of your agrument. In some of the test, the top of the range A10 6700T manages to beat the middle of the range 3740 by around 20%, but then loses to the 3770 when it is included in the benchmarks. So while I agree the AMD chip is great for graphicxs it is certainly no better in CPU tests.
The reason I assess your 'crap' comment against the A8 is because that is the benchmark for tablet chips. Nobody cares about some AMD tablet chip that will never make it in the market.
And thanks for expalining that about dual core and quad core to me. I have just been out there buying octacore smartphone because I was sure they had to be faster than dual core ones. I mean it is like having 8 engines in a car instead of just 2.
Crap is what descibes almost everything that AMD puts out at the moment, cheap but still crap...
Did you? I don't think so. In every tablet out there you will find a z3735, a little slower than the z3740 in T100 that loses in ALL benchmarks except PCMark where other parameters could be the reason for this(storage performance).
Fortunately for the Atom, there where no GPU benchmarks in that article. To be fair the site would had to add an age verification page before giving access to those results. That's how bad those results would have been for Intel.
z3770 is not only the top, it is also the most expensive. And I don't think Intel is giving it as a gift. The problem here is that Intel gives away it's lowest SoCs for free. If you had the mentality to understand how bad monopolists are, I would had spent a little more time to explain to you. Just to help you a little, this is the reason why you will never see an AMD chip in a tablet. Not an x86 anyway.
At least you saw that mentioning A8 was stupid. It would have been better of course if you had the guts to accept your mistake and not just try to find excuses for it talking about your new octa core smartphone(congrats).
PS. More cores is not always better.
PS 2. LEARN TO READ, then you may accuse others about crap comments.
--- I stop here. Nothing else to say. Lost enough time with you.
Try ranting less. You compare Ghz like it's the single defining factor in a mobile chip. Try chilling the fck out and maybe someone might be able to understand the point you're trying to make.
Yannigr is right. I have a T100 in front of me and I can tell you it is not *nearly* as powerful as a laptop. The z3775 atom is equivalent to a 1ghz ulv C2D at best with half the grunt per core. Don't try to sell it as something equivalent to the A8 (or A7) because it absolutely is not even in the same league in terms of responsiveness.
Also, only a few devices even use the z3775. It's mostly z373x, 374x models in the market.
It's not that crappy but if Intel wasn't buying their way to market share through contra revenue you'd probably not see many x86 tablets at all. So you can pick your poison, either it won't sell because of Intel or it won't sell because of ARM, either way Mullins is pretty much dead on arrival. It's no secret what Intel is doing:
AMD is rather screwed, but then this was rather predictable. Everyone knew Intel and ARM (Samsung, Qualcomm, nVidia, Apple+++) was going to clash and when giants fight the little guys are likely to get stomped. I understand the two-way race with Intel was going badly for AMD, but they jumped out of the frying pan and into the fire by going into ARM territory themselves.
There would have been Mullins in the market. Not many, only a few, but there would have been products with Mullins enough to justify the R&D cost, or part of it.
Many people are happy with an Android tablet or a chromebook, many are with an iPhone/iPad, many others want a cheap portable device with x86 Windows on it. If Intel wasn't giving away free chips, yes, x86 devices would have been a little more expensive and not a threat for the ultra cheap ARM tablets. But people who wanted x86 Windows would be buying those. Not in great numbers, but they would be buying them. And AMD would have a small part of that x86 market, even a very small one.
What Intel is doing now is not just trying to create a market share that will have to keep spending money forever to maintain, but it is also killing anything AMD can create and not give away for free of course. So, this poison is much worst for AMD than the other one.
As for AMD jumping on the fire. Nope. You said it yourself. Mullins end up DOA because of Intel's tachtics. So, what do you do then? You go to the ARM platform and do what Nvidia is doing. Use your graphics advantage to try to take a part of the market. At least there you could sell something for money, without Intel sabotaging you in any way possible.
Nvidia abandoned smartphones, in the future Tegra will move from tablets and chromebooks to Steam machines. AMD will follow with both low power x86 and ARM cores. Both companies have an advantage there over Intel and when ARM cores become powerful enough, and Windows less important, they could turn the tables against Intel. Intel of course has billions and billions for R&D and it can close the gap in graphics to defend it's huge market share thanks to integrated graphics that it has today. The future is not x86 only, so no, AMD isn't jumping on the fire.
You believe the only reason why AMD Mullins NOT selling is due to Intel's contra revenue? To me, that's very hard to believe. Unless you are saying there's no / not much difference between Bay Trail and Mullins. Performance / feature differentiation I THINK will always over-power cost difference if there's a market need for it. I think it's that there's really no real Tablet market need for Mullins (or Bay Trail) when it is already saturated by ARM counterparts.
I argue exactly opposite actually. All of the money Intel pours into tablet market to enable x86 share will benefit AMD in the future versions of Mullins (don't know their code names.) If anything AMD should thank Intel for helping to establish x86 Tablet market. So this should be a good news for fanboy like you. :)
AMD usually let Intel create a market and then comes in, I agree. They do it constantly with RAM for example. But this time because of ARM a market is created and sustained in a way that AMD, at least today, can not enter in any meaningful way.
A free chip and an Intel logo on a tablet that it's cost is already low, is way too much marketing value for an AMD product to become an option, even if it performs better in some areas like for example graphics.
Don't worry - Intel won't be able and will not keep up the contra revenue for too long. Next generation of x86 based products need to be free of contra revenue (NOT free of charge of course.) That's when AMD needs to target market entry for Tablet - better make sure whatever lined up to match Intel performance is below Intel's chip cost so value proposition is better than Intel. Cross your fingers. :)
Well, if I was Intel shareholder, I would have been certainly asking, what is their long-term proposition of that business, that they needed to sink half their profits into it ...
NV abandoned commodity phones ($300 and down), not the high end. We will see what Denver gets in this area soon, as it's in house cpu should be even better at power than K1 32bit, which should easily enable their own modem or even qcom's to be included and do a nice job vs the competition especially since S808/810 LOSE this advantage as they are off the shelf IP.
We know maxwell's power kicks the crap out of kepler, so next june we should see an M1@20nm which will surely garner some high end phones if not Denver itself shortly. Nexus 9 and shield tablet have LTE so you can already do it, just need to see how Denver's battery is, and if it's capable of running with decent life in a high end phone. Qcom loses perf and battery advantages of IN HOUSE next year so we'll see how that affects their phone sales. I'd bet NV wins at least a few with denver, or at worst the maxwell version. I don't think NV will put out a 20nm K1, but that would surely do it too (I'm guessing 20nm will be M1 and beyond), but who knows.
how phones was Tegra 4 (not 4i, 4i is a completely different beast and is aimed at lower cost phones) again?
I think it was one, and, that one was part of 2 models, one with Qualcomm... I believe the Tegra one was projected to ship less due to which carriers got it in China and some junk like that.
Only potential reason to but Tegra K1 in a phone is the GPU power, but, in a phone TPD/heat limit, you would have to downclock the GPU so much that it wouldn't end up much faster than its competitors.
I don't buy Tegra getting back into phones anytime soon, at least, not unless they have a Tegra 4i successor, which, I would love to see... Maybe the Maxwell Tegra parts will have high end part with 2 SMM and low end part with 1? I think it would be awsome in a mid-high end phone than.
"The Internet of Things group, which is a relatively new division focused on embedded platforms for retail, automotive, home, and transportation, had revenues for the quarter of $530 million, up 12% year-over-year. Revenues were only up $1 million over last year, with $153 million in revenue for this quarter."
One wonders whether the IoT numbers are as fake as the previous Mobile numbers were. It's pretty obvious that the previous Mobile quarters were playing some sort of games, booking revenue as early as possible and delaying accounting for the bribes until they could be delayed no longer (ie this quarter). After all Atom is actually not a completely crappy chip, unlike Quark, so if they have to bribe the hell out of anyone to use Atom, god only knows how large the bribes are to use Quark.
Atom is only a good chip because of its 22nm FinFET process compared to 28nm planar for ARM chips. But even then it barely competes, especially in GPU performance.
Put Atom and the most cutting edge ARM chip (such as Nvidia's Denver) on the same process, and you'll just how pathetic Atom really is, and why the x86 bloat is very much alive and well, despite Anand's nonsense about the "x86 myth being busted".
The myth was never busted. Anand just conveniently ignored the fact that Atom was using 2-generation jump in process technology to EQUAL ARM chips.
The good news is that we will soon get to see that, as TSMC and Samsung will go to 14/16nm FinFET (still behind Intel's 14nm FinFET a bit, but the gap is much smaller now than before) next year, and we'll also get to see Atom chips made by Rockchip and Spreadtrum at 28nm planar, which will then reveal how weak Atom really is, even compared to 20nm planar (low-end) Cortex A53 OEMs will be using by then at the same level.
I don't think that's right. $25.79 trillion seems an order of magnitude or two too high. The US's GDP is 16.8 trillion, and the world's is 74.9 trillion (google estimates). I don't think a third of the money in the world passes into the hands of the PC Client Group. But if it did, <insert terminator joke here>
Can't wait to see how intel pulls off goin below 10nm. Even 10nm is going to be a rly huge challenge but after 10nm the current lazers just wont do it and they need to use EUV combined with other exotic techniques to get down to 7nm. Wait and see the 14nm transition may of took a little extra but the 10nm transition will be even longer and the 7nm transition is going to be superrrr long. Then we may see 5nm but after that something radical is going to have to happen. No more silicon wafers we need some alien element from planet SA 924
Actually Intel has indicated that 14->10nm will not actually be as hard a shift as 22->14nm was. In large part because some (a lot?) of the changes Intel needed to make to move from 22 to 14nm are carried over for 14->10nm. It is below 10nm that things change up significantly again.
indeed, I believe Intel managed to do single-patterning 193nm light for "22nm" process.
14nm and 10nm both do double patterning 193nm light.
Granted, I think 10nm will have some hidden challenges, given Rock's law (~4 years cost of fabs + R&D + etc. doubles) given that Intel's R&D has been near flat for 2 maybe 3 years.
Sub 10nm I am guessing will be on hold until EUV AND/OR 450mm wafers.
It's not accurate to say that the PC is recovering since it's still declining. It might be stabilizing but that's most likely on available tablets being very poor.If we had a great Nexus and a 4k ipad or any exciting tablets (outside China) things would be different. Intel and M$ managed to make a mess of the Android market this year, Apple doesn't have anything good enough to boost sales, Qualcomm doesn't een have a compelling offering for SoCs without integrated baseband and Google doesn't seem to care much about the tablet market or Wintel trying to cripple Android in tabs. If the tab market gets some exciting products, the PC market will suffer and from signs of stability it would go back to harsh declines. Intel's PC units numbers are not proportional with the market for 2 main reasons, They gained some share from AMD lately and in Q3 the inventory in the channel was up.
My understanding is that only the Broadwell Mobile parts are shipping to customers. Desktop and laptop parts aren't due until Q2, and that has already slipped a number of times. I wouldn't be too surprised to see these slip again.
I am not sure that Intel's name recognition is going to carry over into the mobile market, they just don't have the market recognition of Qualcom in that sector. I guess we will find out soon enough with Christmas coming along soonish.
I wouldn't be surprised by ANY semiconductor process node improvement slip from this point on. I am a believer but the physics is up against timeline for everyone it seems. :)
Broadwell gadgets seem interesting: Core performance in smaller things. But it's worrying that nobody bought a Surface, and folks might see faster ARMs as good enough.
$1 million revenue - $1 BILLION loss - per QUARTER. How much poorer can you run a business Intel?
This is why Intel quit the ARM race a decade ago, too. Intel just doesn't KNOW how to compete with ARM. If they wouldn't have billions of dollars from other more profitable businesses, we would never even hear about an "Intel" chip in the mobile market, just like we aren't hearing about AMD in the mobile market right now.
In fact, I'm surprised Qualcomm, Nvidia, AMD and others don't report Intel to EU and other governments for selling chips WELL below the cost/virtually for free to gain customers.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
37 Comments
Back to Article
Drumsticks - Tuesday, October 14, 2014 - link
It almost surprises me how much Intel is losing in Mobile. How much are they losing per atom chip sold? Atom is certainly a decent chip, although its getting a bit old now, so I wonder how much market share they're gaining. That would really help put this in perspective. If they've gained a decent share of the market, it's certainly not too bad.yannigr2 - Wednesday, October 15, 2014 - link
It's a total crap chip that they are giving away for free. I think we now know for fact why AMD's Mullins haven't been in about anything by now and everyone is using crap Z3735 chips.Speedfriend - Wednesday, October 15, 2014 - link
"It's a total crap chip that they are giving away for free."Right, total crap chip, that is why the Atom 3775 scores 2863 in geekbench while the new Apple A8 scores 2881.
Have you even used something with one of the Atom in, like a Asus T100. It feels as fat as a core laptop.
What is total crap is what you have written, that all...
yannigr2 - Wednesday, October 15, 2014 - link
"Have you even used something with one of the Atom in, like a Asus T100. It feels as fat as a core laptop.What is total crap is what you have written, that all..."
It is obvious that you have no idea what I am talking about. You better start reading a few more articles and google a few unknown names to you, like AMD and Mullins, before posting nonsense. Apple and AMD may have a common first letter in their names but, and this is going to shock you, they are two totally different companies.
So, first you are comparing a quad core x86 chip that has a turbo speed at 1.8 with a dual core ARM chip that has a 1.4GHz frequency. This is a huge facepalm.
Second, who cares about A8? Who mentioned A8?
Have a look at this
http://www.anandtech.com/show/7974/amd-beema-mulli...
you might learn something, but I don't believe so. From your comment I am pretty sure you are 12 years old, mentally at least, or you are just trolling.
Speedfriend - Wednesday, October 15, 2014 - link
Did you even look at what you posted in support of your agrument. In some of the test, the top of the range A10 6700T manages to beat the middle of the range 3740 by around 20%, but then loses to the 3770 when it is included in the benchmarks. So while I agree the AMD chip is great for graphicxs it is certainly no better in CPU tests.The reason I assess your 'crap' comment against the A8 is because that is the benchmark for tablet chips. Nobody cares about some AMD tablet chip that will never make it in the market.
And thanks for expalining that about dual core and quad core to me. I have just been out there buying octacore smartphone because I was sure they had to be faster than dual core ones. I mean it is like having 8 engines in a car instead of just 2.
Crap is what descibes almost everything that AMD puts out at the moment, cheap but still crap...
yannigr2 - Wednesday, October 15, 2014 - link
Did you? I don't think so. In every tablet out there you will find a z3735, a little slower than the z3740 in T100 that loses in ALL benchmarks except PCMark where other parameters could be the reason for this(storage performance).Fortunately for the Atom, there where no GPU benchmarks in that article. To be fair the site would had to add an age verification page before giving access to those results. That's how bad those results would have been for Intel.
z3770 is not only the top, it is also the most expensive. And I don't think Intel is giving it as a gift. The problem here is that Intel gives away it's lowest SoCs for free. If you had the mentality to understand how bad monopolists are, I would had spent a little more time to explain to you. Just to help you a little, this is the reason why you will never see an AMD chip in a tablet. Not an x86 anyway.
At least you saw that mentioning A8 was stupid. It would have been better of course if you had the guts to accept your mistake and not just try to find excuses for it talking about your new octa core smartphone(congrats).
PS. More cores is not always better.
PS 2. LEARN TO READ, then you may accuse others about crap comments.
---
I stop here. Nothing else to say. Lost enough time with you.
RMSe17 - Wednesday, October 15, 2014 - link
lol.djscrew - Wednesday, October 15, 2014 - link
Try ranting less. You compare Ghz like it's the single defining factor in a mobile chip. Try chilling the fck out and maybe someone might be able to understand the point you're trying to make.ppi - Thursday, October 16, 2014 - link
With revenue just only $1M, they must be giving ALL of them for free.garbagedisposal - Wednesday, October 15, 2014 - link
Yannigr is right.I have a T100 in front of me and I can tell you it is not *nearly* as powerful as a laptop.
The z3775 atom is equivalent to a 1ghz ulv C2D at best with half the grunt per core.
Don't try to sell it as something equivalent to the A8 (or A7) because it absolutely is not even in the same league in terms of responsiveness.
Also, only a few devices even use the z3775. It's mostly z373x, 374x models in the market.
TheJian - Wednesday, October 15, 2014 - link
Yet the K1 smoked T100 in everything. See the review of the shield tablet here at anandtech ;)Kjella - Wednesday, October 15, 2014 - link
It's not that crappy but if Intel wasn't buying their way to market share through contra revenue you'd probably not see many x86 tablets at all. So you can pick your poison, either it won't sell because of Intel or it won't sell because of ARM, either way Mullins is pretty much dead on arrival. It's no secret what Intel is doing:http://www.pcworld.com/article/2089421/how-intel-i...
AMD is rather screwed, but then this was rather predictable. Everyone knew Intel and ARM (Samsung, Qualcomm, nVidia, Apple+++) was going to clash and when giants fight the little guys are likely to get stomped. I understand the two-way race with Intel was going badly for AMD, but they jumped out of the frying pan and into the fire by going into ARM territory themselves.
yannigr2 - Wednesday, October 15, 2014 - link
There would have been Mullins in the market. Not many, only a few, but there would have been products with Mullins enough to justify the R&D cost, or part of it.Many people are happy with an Android tablet or a chromebook, many are with an iPhone/iPad, many others want a cheap portable device with x86 Windows on it. If Intel wasn't giving away free chips, yes, x86 devices would have been a little more expensive and not a threat for the ultra cheap ARM tablets. But people who wanted x86 Windows would be buying those. Not in great numbers, but they would be buying them. And AMD would have a small part of that x86 market, even a very small one.
What Intel is doing now is not just trying to create a market share that will have to keep spending money forever to maintain, but it is also killing anything AMD can create and not give away for free of course. So, this poison is much worst for AMD than the other one.
As for AMD jumping on the fire. Nope. You said it yourself. Mullins end up DOA because of Intel's tachtics. So, what do you do then? You go to the ARM platform and do what Nvidia is doing. Use your graphics advantage to try to take a part of the market. At least there you could sell something for money, without Intel sabotaging you in any way possible.
Nvidia abandoned smartphones, in the future Tegra will move from tablets and chromebooks to Steam machines. AMD will follow with both low power x86 and ARM cores. Both companies have an advantage there over Intel and when ARM cores become powerful enough, and Windows less important, they could turn the tables against Intel. Intel of course has billions and billions for R&D and it can close the gap in graphics to defend it's huge market share thanks to integrated graphics that it has today. The future is not x86 only, so no, AMD isn't jumping on the fire.
AnakinG - Wednesday, October 15, 2014 - link
You believe the only reason why AMD Mullins NOT selling is due to Intel's contra revenue? To me, that's very hard to believe. Unless you are saying there's no / not much difference between Bay Trail and Mullins. Performance / feature differentiation I THINK will always over-power cost difference if there's a market need for it. I think it's that there's really no real Tablet market need for Mullins (or Bay Trail) when it is already saturated by ARM counterparts.I argue exactly opposite actually. All of the money Intel pours into tablet market to enable x86 share will benefit AMD in the future versions of Mullins (don't know their code names.) If anything AMD should thank Intel for helping to establish x86 Tablet market. So this should be a good news for fanboy like you. :)
yannigr2 - Wednesday, October 15, 2014 - link
AMD usually let Intel create a market and then comes in, I agree. They do it constantly with RAM for example. But this time because of ARM a market is created and sustained in a way that AMD, at least today, can not enter in any meaningful way.A free chip and an Intel logo on a tablet that it's cost is already low, is way too much marketing value for an AMD product to become an option, even if it performs better in some areas like for example graphics.
AnakinG - Wednesday, October 15, 2014 - link
Don't worry - Intel won't be able and will not keep up the contra revenue for too long. Next generation of x86 based products need to be free of contra revenue (NOT free of charge of course.) That's when AMD needs to target market entry for Tablet - better make sure whatever lined up to match Intel performance is below Intel's chip cost so value proposition is better than Intel. Cross your fingers. :)ppi - Thursday, October 16, 2014 - link
Well, if I was Intel shareholder, I would have been certainly asking, what is their long-term proposition of that business, that they needed to sink half their profits into it ...TheJian - Wednesday, October 15, 2014 - link
NV abandoned commodity phones ($300 and down), not the high end. We will see what Denver gets in this area soon, as it's in house cpu should be even better at power than K1 32bit, which should easily enable their own modem or even qcom's to be included and do a nice job vs the competition especially since S808/810 LOSE this advantage as they are off the shelf IP.We know maxwell's power kicks the crap out of kepler, so next june we should see an M1@20nm which will surely garner some high end phones if not Denver itself shortly. Nexus 9 and shield tablet have LTE so you can already do it, just need to see how Denver's battery is, and if it's capable of running with decent life in a high end phone. Qcom loses perf and battery advantages of IN HOUSE next year so we'll see how that affects their phone sales. I'd bet NV wins at least a few with denver, or at worst the maxwell version. I don't think NV will put out a 20nm K1, but that would surely do it too (I'm guessing 20nm will be M1 and beyond), but who knows.
testbug00 - Friday, October 17, 2014 - link
how phones was Tegra 4 (not 4i, 4i is a completely different beast and is aimed at lower cost phones) again?I think it was one, and, that one was part of 2 models, one with Qualcomm... I believe the Tegra one was projected to ship less due to which carriers got it in China and some junk like that.
Only potential reason to but Tegra K1 in a phone is the GPU power, but, in a phone TPD/heat limit, you would have to downclock the GPU so much that it wouldn't end up much faster than its competitors.
I don't buy Tegra getting back into phones anytime soon, at least, not unless they have a Tegra 4i successor, which, I would love to see... Maybe the Maxwell Tegra parts will have high end part with 2 SMM and low end part with 1? I think it would be awsome in a mid-high end phone than.
name99 - Wednesday, October 15, 2014 - link
"The Internet of Things group, which is a relatively new division focused on embedded platforms for retail, automotive, home, and transportation, had revenues for the quarter of $530 million, up 12% year-over-year. Revenues were only up $1 million over last year, with $153 million in revenue for this quarter."One wonders whether the IoT numbers are as fake as the previous Mobile numbers were.
It's pretty obvious that the previous Mobile quarters were playing some sort of games, booking revenue as early as possible and delaying accounting for the bribes until they could be delayed no longer (ie this quarter).
After all Atom is actually not a completely crappy chip, unlike Quark, so if they have to bribe the hell out of anyone to use Atom, god only knows how large the bribes are to use Quark.
Krysto - Wednesday, October 22, 2014 - link
Atom is only a good chip because of its 22nm FinFET process compared to 28nm planar for ARM chips. But even then it barely competes, especially in GPU performance.Put Atom and the most cutting edge ARM chip (such as Nvidia's Denver) on the same process, and you'll just how pathetic Atom really is, and why the x86 bloat is very much alive and well, despite Anand's nonsense about the "x86 myth being busted".
The myth was never busted. Anand just conveniently ignored the fact that Atom was using 2-generation jump in process technology to EQUAL ARM chips.
The good news is that we will soon get to see that, as TSMC and Samsung will go to 14/16nm FinFET (still behind Intel's 14nm FinFET a bit, but the gap is much smaller now than before) next year, and we'll also get to see Atom chips made by Rockchip and Spreadtrum at 28nm planar, which will then reveal how weak Atom really is, even compared to 20nm planar (low-end) Cortex A53 OEMs will be using by then at the same level.
iwod - Tuesday, October 14, 2014 - link
It continues to baffles me why the hell Intel ever bought McAfee in the first place.staiaoman - Tuesday, October 14, 2014 - link
wow! The PC Client group has contributed $25.79 TRILLION in revenue! Great work, the PC Client group!.Also, if something is down "around 100%" wouldn't that mean it went from X to 0, rather than from X to X/2?
;-P
All in all, appreciate the summary, as well as the quick turnaround!
feeblegoat - Wednesday, October 15, 2014 - link
I don't think that's right. $25.79 trillion seems an order of magnitude or two too high. The US's GDP is 16.8 trillion, and the world's is 74.9 trillion (google estimates). I don't think a third of the money in the world passes into the hands of the PC Client Group. But if it did, <insert terminator joke here>stadisticado - Wednesday, October 15, 2014 - link
Yeah, pretty sure that was meant as a European comma and not an American one.Drumsticks - Wednesday, October 15, 2014 - link
I noticed that. I think you're right, although it's interesting because everywhere else used American notation in the article.AnakinG - Wednesday, October 15, 2014 - link
Intel has beaten apple and google as first trillion company? :)Laststop311 - Wednesday, October 15, 2014 - link
Can't wait to see how intel pulls off goin below 10nm. Even 10nm is going to be a rly huge challenge but after 10nm the current lazers just wont do it and they need to use EUV combined with other exotic techniques to get down to 7nm. Wait and see the 14nm transition may of took a little extra but the 10nm transition will be even longer and the 7nm transition is going to be superrrr long. Then we may see 5nm but after that something radical is going to have to happen. No more silicon wafers we need some alien element from planet SA 924azazel1024 - Wednesday, October 15, 2014 - link
Actually Intel has indicated that 14->10nm will not actually be as hard a shift as 22->14nm was. In large part because some (a lot?) of the changes Intel needed to make to move from 22 to 14nm are carried over for 14->10nm. It is below 10nm that things change up significantly again.testbug00 - Friday, October 17, 2014 - link
indeed, I believe Intel managed to do single-patterning 193nm light for "22nm" process.14nm and 10nm both do double patterning 193nm light.
Granted, I think 10nm will have some hidden challenges, given Rock's law (~4 years cost of fabs + R&D + etc. doubles) given that Intel's R&D has been near flat for 2 maybe 3 years.
Sub 10nm I am guessing will be on hold until EUV AND/OR 450mm wafers.
jjj - Wednesday, October 15, 2014 - link
It's not accurate to say that the PC is recovering since it's still declining. It might be stabilizing but that's most likely on available tablets being very poor.If we had a great Nexus and a 4k ipad or any exciting tablets (outside China) things would be different. Intel and M$ managed to make a mess of the Android market this year, Apple doesn't have anything good enough to boost sales, Qualcomm doesn't een have a compelling offering for SoCs without integrated baseband and Google doesn't seem to care much about the tablet market or Wintel trying to cripple Android in tabs. If the tab market gets some exciting products, the PC market will suffer and from signs of stability it would go back to harsh declines.Intel's PC units numbers are not proportional with the market for 2 main reasons, They gained some share from AMD lately and in Q3 the inventory in the channel was up.
danjw - Wednesday, October 15, 2014 - link
My understanding is that only the Broadwell Mobile parts are shipping to customers. Desktop and laptop parts aren't due until Q2, and that has already slipped a number of times. I wouldn't be too surprised to see these slip again.I am not sure that Intel's name recognition is going to carry over into the mobile market, they just don't have the market recognition of Qualcom in that sector. I guess we will find out soon enough with Christmas coming along soonish.
AnakinG - Wednesday, October 15, 2014 - link
I wouldn't be surprised by ANY semiconductor process node improvement slip from this point on. I am a believer but the physics is up against timeline for everyone it seems. :)testbug00 - Friday, October 17, 2014 - link
Check out Rock's law and Intel's R&D/fab spending and you'll figure out the issue pretty quickly.twotwotwo - Wednesday, October 15, 2014 - link
Broadwell gadgets seem interesting: Core performance in smaller things. But it's worrying that nobody bought a Surface, and folks might see faster ARMs as good enough.TiGr1982 - Wednesday, October 15, 2014 - link
Some did (e.g. I bought Surface Pro 1 128 GB this year in March for $500 on sale. Its a very nice device overall in its class).Krysto - Wednesday, October 22, 2014 - link
$1 million revenue - $1 BILLION loss - per QUARTER. How much poorer can you run a business Intel?This is why Intel quit the ARM race a decade ago, too. Intel just doesn't KNOW how to compete with ARM. If they wouldn't have billions of dollars from other more profitable businesses, we would never even hear about an "Intel" chip in the mobile market, just like we aren't hearing about AMD in the mobile market right now.
In fact, I'm surprised Qualcomm, Nvidia, AMD and others don't report Intel to EU and other governments for selling chips WELL below the cost/virtually for free to gain customers.