Intel Q213 Results

[ShintaiDK]

You are not looking.

Neither are you.

 

[pablo87]

Wake me up when they do something about it. Everything is going as expected. But who thinks baitrail is going to change that pattern? Good grief.

Yeaa i am sure bail trail will walk all over an furure a12. But really this old time bm positioning is kind of childish competing with samsung and apple.

What the world needs is a53. And intel is comming to market with a solution that makes a15 look dirt cheap.

Doesn't Intel have a few design wins? Or did they pay for them?

 

[krumme]

And a strong case can be made that Intel's lackadaisical approach and arrogance is part of their undoing.

Intel is now a reactionary company to external forces, not an innovator pushing the market in new directions. That honor goes to companies like ARM, Apple, Samsung, Qualcomm, and Google. There are lots of indicators that their process tech lead, which really is their last bastion of strength, may not last much longer either.

I don't hate intel. Being an American company I'd like them to succeed vs Samsung or ARM, but they need to do something fantastic here to avoid becoming a has-been in a few years. Ditto for Microsoft.

A company in Intels position ought to have far higher value and profit. Its always easy to look in the mirror, but Intel is doing the same errors right now. Its textbook bad, and will be an historical example.

Seen from here the non technical management simply looks to lack class, and some intellectual weight. How can they continue to segment the market, eroding ther last strongholes, like we are in the nienties?

They make worldclass processes taking years of research and planning, and yet the business side looks line 2 years planning.

 

[krumme]

Danger: Capex alert

 

[krumme]

Doesn't Intel have a few design wins? Or did they pay for them?

Didnt you read the post where i wrote what Intel inside means; i need money fast, and i am looking for a new job.

 

[StrangerGuy]

Neither are you.

Meh, his Dailytech was June, yours was May.

Besides, process advantage means nothing when the market doesn't want your products and TSMC/Samsung at worst will only be half a node away. If the "world leading" 22nm didn't destroy TSMC why would anything change by 14nm?

Isn't that roadmap glorious though, filled with model numbers even enthusiasts will have trouble wrapping their head around let alone layman consumers.

 

[NTMBK]

You are not looking.

http://www.dailytech.com/Report+Int...les+Haswell+Refresh+for+2014/article31770.htm

http://www.dailytech.com/Report+Sam...2014+Make+14+nm+iPhone+7+SoC/article31963.htm

http://www.brightsideofnews.com/new...mce28099s-16nm-finfet-process-technology.aspx

The "Broadwell is delayed" stories are based on the desktop roadmap- Broadwell is still coming out on time in mobile, and 14nm Atom will be coming out in 2014 as well.

 

[krumme]

Well what does it really matter none wants intel phone or tablet cpu, as its a loss either way; if they get into phones or not? Its a lose/lose strategy.

 

[ShintaiDK]

Well what does it really matter none wants intel phone or tablet cpu, as its a loss either way; if they get into phones or not? Its a lose/lose strategy.

Is that something you can backup? Or is just your view?

Even Samsung moves to Intel CPUs in their tablets.

 

[SiliconWars]

Is that something you can backup? Or is just your view?

Even Samsung moves to Intel CPUs in their tablets.

Qualcomm is basically mocking Intel on their lack of design wins - http://www.bloomberg.com/news/2013-07-17/qualcomm-says-tablet-chip-will-build-lead-over-intel.html

The company has begun selling upgraded versions of its Snapdragon processor line and anticipates the new chips will appear in 200 phones and tablets.

How many design wins, total, has intel had in phones and tablets? Even with Bay Trail they'll still be lagging far behind. Niow Qualcomm is saying "we're going to beat you easily in tablets as well". Do you disagree with this? If so then I'd like to see where the design wins are.

Also, that Samsung tablet is getting pilloried in the press. It's an overpriced, under-specced piece of crap - http://gigaom.com/2013/07/16/galaxy-tab-3-10-inch-vs-nexus-10/

Based on Intels other architecture group loss ($608 million? That's crazy and shows just how far away from Qualcomm they are) this quarter it appears they must be giving Clover Trail away for free.

 

[blackened23]

That's pretty funny. I remember a time when AMD was equally as arrogant as ARMH / Qualcomm, mocking intel (during the Athlon and pre-conroe era), and we all know what happened there.

Meanwhile intel just quietly does their own thing without talking smack through the press.

 

[ShintaiDK]

That's pretty funny. I remember a time when AMD was equally as arrogant as ARMH / Qualcomm, mocking intel (during the Athlon and pre-conroe era), and we all know what happened there.

Meanwhile intel just quietly does their own thing without talking smack through the press.

Yep. If anyone wants a hillarious travel back in time, you can find it here:
http://shintai.ambition.cz/amd.pdf

On pure MPU sales alone, Qualcomm is suddenly not so big.

 

[MichaelBarg]

That becomes a chicken and egg argument.

Ie, does Intel create more powerful CPUs first, then developers learn to use that power ( the proverbial killer app ) and thus drive hardware upgrades? Or does the killer app that demands more power from CPUs come first?

Historically it has been the former, just look at all the upgrading in the 90s centered around Quake, GLQuake, Quake 2. Recently it has been neither.

I am increasingly skeptical of this argument. For enterprise users especially, compute is just not a factor anymore. I have an I7-2600k on my desk at work and if someone snuck in overnight and swapped it for a SB I3 I'm not sure I'd ever notice. Many desktop buyers have lots of extra compute and have for a while.

 

[StrangerGuy]

Qualcomm is basically mocking Intel on their lack of design wins - http://www.bloomberg.com/news/2013-07-17/qualcomm-says-tablet-chip-will-build-lead-over-intel.html

How many design wins, total, has intel had in phones and tablets? Even with Bay Trail they'll still be lagging far behind. Niow Qualcomm is saying "we're going to beat you easily in tablets as well". Do you disagree with this? If so then I'd like to see where the design wins are.

Also, that Samsung tablet is getting pilloried in the press. It's an overpriced, under-specced piece of crap - http://gigaom.com/2013/07/16/galaxy-tab-3-10-inch-vs-nexus-10/

Based on Intels other architecture group loss ($608 million? That's crazy and shows just how far away from Qualcomm they are) this quarter it appears they must be giving Clover Trail away for free.

Not to mention getting design wins is one thing, actually selling those wins to consumers is another. No. of design wins is no guarantee for success with Samsung and Apple taking up the bulk of the market.

That's pretty funny. I remember a time when AMD was equally as arrogant as ARMH / Qualcomm, mocking intel (during the Athlon and pre-conroe era), and we all know what happened there.

Meanwhile intel just quietly does their own thing without talking smack through the press.

Right, because we now are obviously living in a past era where computing is defined by Wintel or die.

 

[blackened23]

You're right. We're in a new era defined by efficiency and I think intel will do just fine there - look at how intel delivered with Haswell ULV in that respect. If Haswell is any indication, Bay Trail will be everything intel claims it to be. That's the thing about intel. They deliver on promises while ARMH boasts about unreleased products from 2015 through marketing spin and powerpoint slides. Rather than fighting battles through the press, intel just delivers the goods. I expect nothing less with Bay Trail.

 

[pelov]

You're right. We're in a new era defined by [efficiency and price] and I [don't] think intel will do just fine

FTFY.

ARM isn't winning on the low end because it's got amazing performance or incredible efficiency, it's winning because it offers IP to everybody and that consequently has opened up a massive market for selling small and efficient chips at rock-bottom prices.

With a process advantage, I have no doubt that Intel can match ARM in efficiency, but that's not the question here. The question is, can Intel match Qualcomm/Samsung/Apple/etc. in efficiency, performance, price, and volume over an extended period of time? The first two? Sure. The latter two? I just don't see it happening. If you're relying on a process advantage to put your products on an even keel against your competitor then you're also going to have to charge more for those same products, otherwise that process advantage will very quickly disappear.

It all revolves around money. If Intel can't charge a high margin on costly chips they won't maintain a process advantage. I don't care how many great engineers and architects you have, unless you've got billions upon billions to spend for fabs, you're going to have it rough. -30% in profits means less cash to spend on process

 

[blackened23]

I would disagree there. ARMH took the mobile world by storm because their efficiency was far beyond what anything else on the market offered. Period.

Cost is one thing, but this is the reason ARMH SOC's have done so well. They have not been contested up until this point, but it seems that Bay Trial will have even better efficiency than the best ARM SOC's. Heck, even ULV Haswell has equal or better battery life than most ARM SOCs and we obviously know Haswell destroys them in performance. I expect the same from Bay Trail.

 

[pelov]

ARMH sells most of their chips for embedded devices and has a very long history of doing that. There have been other efficient designs that competed, but ultimately it boils down to both price and efficiency.

And ARM has been contested. I don't think you understand the difference between the x86 landscape and ARM...

When AT forum-goers speak of ARM, they literally mean ARM Holders, the company that designs the IP and licenses the ISA, but that's not where the competition is. Intel, and x86, isn't facing ARM Holders, they're facing the plethora of massive corporations that license ARM. That's where the competition is and will be for the foreseeable future. Intel has to go up against Samsung producing A15s on their own fabs and making Apple's A6 SoCs. Intel has to face Qualcomm and the army of SoCs and IP they've got up for offer. Intel has to tackle the lower end of that spectrum as well, with Rockwell and the like. But it doesn't stop there, because these corporations are facing off against themselves as well. ARM isn't about tiny efficient cores. If you really think that, you're ignoring the most important factor: price and competition. What can I get for as cheap a price as possible?

Intel's Haswell sells at prices that can be 10x higher than the entire ARM SoC - and that's the entire SoC, not just the CPU+GPU. More importantly, it's selling in products that nobody wants and at product prices that are too high. Intel would never be able to operate selling their Haswell cores at bargain bin prices and especially not the ULV stuff that has to be strictly binned to get there.

Anandtechers just don't get it. It's like a foreign language to most people here. Here it's been about paying a premium for premium performance. We view computing, and great CPU performance more succinctly, as a luxury. The world outside doesn't really care about GFLOPs and views computing as a necessity that should be cheap, portable and disposable. ARM has a long history of adapting to that and thriving in that landscape whereas Intel is hopelessly lost.

Intel's new CEO, Krzanich, at least seems to understand it:

In the call, he promised lower price points for chips and cheaper devices -- particularly on the low end.

II. $150 USD Tablets With Intel Processors? Intel CEO Promises Big

Key to that effort will be the upcoming sixth major release to the Atom platform. Manufactured on the latest 22 nm node the quad-core tablet-geared Atoms (core: Silvermont; SoC: ValleyView; chipset: Bay Trail) are expected to be shipping in tablets and hybrid notebook/tablets this holiday season.

Mr. Krzanich boldly predicted the following price points:
Convertibles (hybrids) : $400 USD
Laptops (w/ touch) : $300 USD
Tablets : $150-200 USD

http://www.dailytech.com/Intel+CEO+...op+Promises+150+Atom+Tablets/article31996.htm

but what are the margins and what are the Atoms selling for? If the OEMs have to keep skimping on the display or battery or build quality, then hitting those price points won't matter. We're in an era where a tablet might cost $200-300 to build and the SoC should cost only a tenth of that. Intel's been living off margins that go anywhere from 30-50% of the BoM for a $500-$750 laptop/desktop.

Everyone knows what they need to do, the more pertinent question is can they actually compete with ARM on price and follow through? Nowhere in recent history has Intel offered x86 chips for cheap, and I don't see that changing.

 

[Genx87]

Meh, his Dailytech was June, yours was May.

Besides, process advantage means nothing when the market doesn't want your products and TSMC/Samsung at worst will only be half a node away. If the "world leading" 22nm didn't destroy TSMC why would anything change by 14nm?

Isn't that roadmap glorious though, filled with model numbers even enthusiasts will have trouble wrapping their head around let alone layman consumers.

I have to admit all the model numbers have become confusing even for somebody like me who enjoys the industry. I have to look up model numbers to figure out their specs. The result? I have lost interest.

 

[StrangerGuy]

Anandtechers just don't get it. It's like a foreign language to most people here. Here it's been about paying a premium for premium performance. We view computing, and great CPU performance more succinctly, as a luxury. The world outside doesn't really care about GFLOPs and views computing as a necessity that should be cheap, portable and disposable. ARM has a long history of adapting to that and thriving in that landscape whereas Intel is hopelessly lost.

Let me tell you why, some people here have this baseless opinion that as long as Intel can simply put out a superior chip than ARMy and then every ARM user and their mother will drown Intel with money. When the market simply doesn't work that way.

 

[Genx87]

That's pretty funny. I remember a time when AMD was equally as arrogant as ARMH / Qualcomm, mocking intel (during the Athlon and pre-conroe era), and we all know what happened there.

Meanwhile intel just quietly does their own thing without talking smack through the press.

I would say there is a difference however. AMD never had the majority of the market in terms of profits or marketshare. AMD was a small barking dog at the slower mastiff. Eventually the Mastiff woke up and smoked the small dog. Qualcomm is the mastiff right now in this market. Intel has the resources and ability to get it done. I just question whether they can stomach the low margins over the next decade to do it.

 

[StrangerGuy]

I would say there is a difference however. AMD never had the majority of the market in terms of profits or marketshare. AMD was a small barking dog at the slower mastiff. Eventually the Mastiff woke up and smoked the small dog. Qualcomm is the mastiff right now in this market. Intel has the resources and ability to get it done. I just question whether they can stomach the low margins over the next decade to do it.

The difference in situations is massive:

2003: Intel deliberately went for a lousy microarch for desktops, allowing AMD to win with A64 but still dominates laptops with Pentium M. People HAVE to buy PCs to anything done, only choice was whether it had an AMD or Intel chip. PC market then was exploding, many people didn't have one at all or stuck with slow pre-2000 stuff.


2013: PC market shrinks by double digits consecutively for 2 quarters, people are satisfied even with 2008 PCs because Core was so good right from the start that it had amazing longevity, and general software doesn't need so much CPU power. Most people doesn't care about AMD, Intel or MS anymore, because iPad and Android stuff are now the hot shit, and they also don't care what SoC is inside those either. Many can even get by without touching anything from x86 or MS these days.

 

[SammichPG]

I would disagree there. ARMH took the mobile world by storm because their efficiency was far beyond what anything else on the market offered. Period.

Cost is one thing, but this is the reason ARMH SOC's have done so well. They have not been contested up until this point, but it seems that Bay Trial will have even better efficiency than the best ARM SOC's. Heck, even ULV Haswell has equal or better battery life than most ARM SOCs and we obviously know Haswell destroys them in performance. I expect the same from Bay Trail.

When people are happy with 200$ tablets they won't pay 300$ for the intel inside sticker just to keep intel's margins.

Intel can't compete on price, the other competitors are happy to live with lower margins and turn those savings into OEM profits or a lower street price, either way Intel is not the first choice.

The good enough computing is in full swing and microsoft looks like another giant too slow to adapt.

 

[blackened23]

Intel can and will compete on price. Intel has re-iterated at investor conference calls that they will provide bay trail chips for the sub 200$ market, using both Windows 8.1 and Android. Furthermore, keep in mind that there are a wide array of chips both for the low end and high end - this applies to ARM SOCs as well. Intel will provide both - obviously the low end will be low margin (but will sell more) and the high end chips will be high margin. High end ARM SOCs also are not cheap, don't think for a second that 100% of consumers are obsessed with ultra low end android junk. This is why apple sells a gabillion macbook airs and high end ipads in the 500$+ price range - high end chips absolutely are still relevant. But, again, intel will provide bay trail chips for the low end as well.

But, think what you want. Intel knows what has to be done, they know they have to compete on price (and have stated as such, and WILL do it). Just because low end products exist, doesn't mean that high end spectrum products don't sell - they do. I personally think the sub-200$ tablet market is trash. I want something more than that, i'm fully willing to pay 500$+ for a high quality product, and others do as well. Again, see how many people buy 600-800$ ipads for proof of this. But that being said, again, intel will provide for the sub 200$ trash market that you seem so obsessed with.

As for all the doubters, I just find it funny. We'll just wait and see what happens I suppose, it will be quite funny to see ARMH's hubris catch up to them. AMD also had similar press release wars with intel, back when they had a clear performance and efficiency lead. What is AMD doing these days?

 

[sontin]

Intel has no high end SoC. They have only BayTrail. And you can't go into the high end with 1/2 half of the GPU performance of nearly every other high-end ARM SoC.