Spin-Off: AMD & Intel Business Practices

[piesquared]

Another uncompetitive product?

You must be thinking of FailTrail. You know, the one that intel has to subsidize with > $1 Billion to get any interest lol. The one that Kabini already outperforms on a 28nm node, at similar power consumption. If this is Beema/Mullins, intel will need a couple Billion more to 'compete'.

Spun off from http://forums.anandtech.com/showthread.php?t=2380234.

And pie, it may have been your thread but you can still get infracted for thread crapping in it.

-ViRGE

 

[Arachnotronic]

You must be thinking of FailTrail. You know, the one that intel has to subsidize with > $1 Billion to get any interest lol. The one that Kabini already outperforms on a 28nm node, at similar power consumption. If this is Beema/Mullins, intel will need a couple Billion more to 'compete'.

They're not subsidizing Bay Trail-M and Bay Trail-D...you know, the chips that are now taking share from AMD in the low end at high margins.

Bay Trail-T will get Intel's foot nicely in the door, bloated BoM or not, paving the way for profitable, extremely competitive Cherry Trail and Broxton.

 

[piesquared]

They're not subsidizing Bay Trail-M and Bay Trail-D...you know, the chips that are now taking share from AMD in the low end at high margins.

Bay Trail-T will get Intel's foot nicely in the door, bloated BoM or not, paving the way for profitable, extremely competitive Cherry Trail and Broxton. But I wouldn't expect somebody who thinks that AMD runs a great business to understand the kinds of tactics required for long-term business success.

Kabini already outperforms Fail Trail, Beema and Mullins will just increase the margin to 200%-250% per/watt (on a 28nm node! incredible). Fairy Tale and Broxton will also likely need subsidies.

"kind of tactics". lol yes, everyone is aware of 'intel's tactics', which they resort to when their products are technologically inferior. Cat cores are a problem for intel, but I wouldn't expect somebody invested in intel/ short AMD to admit that.

 

[Arachnotronic]

Kabini already outperforms Fail Trail, Beema and Mullins will just increase the margin to 200%-250% per/watt (on a 28nm node! incredible). Fairy Tale and Broxton will also likely need subsidies.

"kind of tactics". lol yes, everyone is aware of 'intel's tactics', which they resort to when their products are technologically inferior. Cat cores are a problem for intel, but I wouldn't expect somebody invested in intel/ short AMD to admit that.

Let's see AMD win some designs in notebooks/tablets that matter, and I'll happily believe AMD's continued claims of technological superiority and business viability.

How many Temash tablets made it to market? How many "Hondo" tablets made it to market? How many of these were any good?

Mock Bay Trail all you'd like, but I have a strong suspicion that it will be in much more desirable devices this year than any AMD part will be. Qualcomm and Samsung are Intel's problem...not AMD.

p.s. unlike many of the AMD optimists/shareholders when discussing their company, I am often quite critical of Intel's efforts in the mobile space. Merrifield was a disaster, Moorefield is too late, and Broxton is still unclear (but for the record I am very hopeful). Bay Trail-T's contra-revenue requirement was unfortunate, but it is a byproduct of poor design decisions at the platform level. Intel's not perfect, but don't talk to me about companies like AMD (the financial results don't lie - how much share did AMD lose in PCs this quarter?) when as an Intel investor my eyes are on the companies with fat R&D budgets and a track record of strong execution (i.e. Qualcomm).

 

[piesquared]

Let's see AMD win some designs in notebooks/tablets that matter, and I'll happily believe AMD's continued claims of technological superiority and business viability.

How many Temash tablets made it to market? How many "Hondo" tablets made it to market? How many of these were any good?

Mock Bay Trail all you'd like, but I have a strong suspicion that it will be in much more desirable devices this year than any AMD part will be. Qualcomm and Samsung are Intel's problem...not AMD.

p.s. unlike many of the AMD optimists, I am often quite critical of Intel's efforts in the mobile space. Merrifield was a disaster, Moorefield is too late, and Broxton is still unclear. Bay Trail-T's contra-revenue requirement was unfortunate, but it is a byproduct of poor design decisions at the platform level. Intel's not perfect, but don't talk to me about companies like AMD (the financial results don't lie - how much share did AMD lose in PCs this quarter?) when as an Intel investor my eyes are on the companies with fat R&D budgets and a track record of strong execution (i.e. Qualcomm).

You talk as if AMD's product capabilities are the reason for their market share. I think it's clear that AMD's market share is a result of intel's 'tactics', and not AMD's technological capabilities. We've seen this for over a decade now, from white boxed motherboards to outright bribes. Design wins are a result of backroom deals and market clout obviously, or we'd see AMD with a deservedly far higher market share. I want products that bring the capabilities that AMD IP offers, and I blame and accuse intel's 'tactics' for limiting my choices of the best products possible.

Yep Qualcomm, Samsung, and a host of other ARM players are serious threat to intel, but I doubt very much they are taking their eye off the AMD ball. Complacency will kill them, but they aren't strong enough to fight on so many fronts. That $1 Billion dollar 'contra revenue' will morph into some other name next year when they continue to fall further and further behind the big boys in mobile.

Do you think a %200-%250 is an insignificant deficit that warrants more products being sold? If that is the case then a serious red flag will be raised.

 

[ShintaiDK]

Yes, everyone but AMD is to blame for AMDs shortcomings....

If you're going to refute someone, take the time and effort to do it right. This is just thread crapping
-ViRGE

 

[2is]

Yes, everyone but AMD is to blame for AMDs shortcomings....

Please edit your post! Don't let the cat out of the bag... We need people who think this to keep thinking this. They help keep AMD in business which is good for YOU and ME!

 

[Arachnotronic]

You talk as if AMD's product capabilities are the reason for their market share. I think it's clear that AMD's market share is a result of intel's 'tactics', and not AMD's technological capabilities. We've seen this for over a decade now, from white boxed motherboards to outright bribes. Design wins are a result of backroom deals and market clout obviously, or we'd see AMD with a deservedly far higher market share. I want products that bring the capabilities that AMD IP offers, and I blame and accuse intel's 'tactics' for limiting my choices of the best products possible.

Yep Qualcomm, Samsung, and a host of other ARM players are serious threat to intel, but I doubt very much they are taking their eye off the AMD ball. Complacency will kill them, but they aren't strong enough to fight on so many fronts. That $1 Billion dollar 'contra revenue' will morph into some other name next year when they continue to fall further and further behind the big boys in mobile.

Do you think a %200-%250 is an insignificant deficit that warrants more products being sold? If that is the case then a serious red flag will be raised.

The contra-revenue is to offset a pretty serious issue with the platform level bill of materials. Intel has promised its shareholders that this is gone by 2015 as SoFIA/Broxton roll-out. If this ISN'T the case, then Intel will have some pretty serious explaining to do to its shareholders.

Also, I will believe the Beema/Mullins numbers when I see them but more importantly, I want to see design wins. That's the only thing that matters in this game, and why Intel's smartphone efforts to date have been largely a disappointment. Merrifield/Moorefield have been touted as great and awesome but not a single design based on them despite Intel's claims has been announced or even leaked.

My nonsense detector is equal opportunity and I make sure to let my analysis guide my stock positions, not to let my stock positions guide my analysis. Intel's execution in phones has been very poor, but its push in tablets has been better, albeit the Bay Trail mishap is far from ideal.

 

[piesquared]

The contra-revenue is to offset a pretty serious issue with the platform level bill of materials. Intel has promised its shareholders that this is gone by 2015 as SoFIA/Broxton roll-out. If this ISN'T the case, then Intel will have some pretty serious explaining to do to its shareholders.

Also, I will believe the Beema/Mullins numbers when I see them but more importantly, I want to see design wins. That's the only thing that matters in this game, and why Intel's smartphone efforts to date have been largely a disappointment. Merrifield/Moorefield have been touted as great and awesome but not a single design based on them despite Intel's claims has been announced or even leaked.

My nonsense detector is equal opportunity and I make sure to let my analysis guide my stock positions, not to let my stock positions guide my analysis. Intel's execution in phones has been very poor, but its push in tablets has been better, albeit the Bay Trail mishap is far from ideal.

What would it suggest to you, if the 200%-250% perf/watt lead for Beema/Mullins is indeed accurate, yet design wins are low? Or let's lowball it for argument's sake and suggest a 150%-200% lead. intel's abysmal smartphone efforts have failed because their products there are failures, contrary to what intel was telling it's investors. That's not the case with Cat cores, they are very competitive and leaders in the majority of metrics, so logic says that something else must be influencing design wins.

I dabbled with 100 shares a while back just to feel like I supported fair play, but after witnessing the corruption and greed that is wall street, and the propaganda medium that it has turned tech forums around the internet into, I won't give them a dime to manipulate and then take. Greed is such an unadmirable trait.

 

[cbn]

The contra-revenue is to offset a pretty serious issue with the platform level bill of materials. Intel has promised its shareholders that this is gone by 2015 as SoFIA/Broxton roll-out. If this ISN'T the case, then Intel will have some pretty serious explaining to do to its shareholders.

I could see SoFIA not needing contra revenue......but I don't yet understand how Broxton will not need contra revenue?

 

[cbn]

Merrifield/Moorefield have been touted as great and awesome but not a single design based on them despite Intel's claims has been announced or even leaked.

Could Merrifield be used as a low cost desktop chip? On some kind of low cost motherboard with the appropriate amount of I/O?

 

[ShintaiDK]

I could see SoFIA not needing contra revenue......but I don't yet understand how Broxton will not need contra revenue?

Increased volume is the key point.

 

[cbn]

Increased volume is the key point.

Increased volume on a chip with 16EUs?

How many people need something with such a large iGPU?

I could see Intel possibly getting away with the overkill iGPU if the SOC's integration in other areas was best of class, but they are not there yet.

SoFIA, on the other hand, yes I could see that one hitting some good volume.

 

[Vesku]

Let's see AMD win some designs in notebooks/tablets that matter, and I'll happily believe AMD's continued claims of technological superiority and business viability.

That's part of AMD's freefall issue, it's pretty obvious they haven't been able to pay to gain OEM design wins for at least 6 perhaps 9 years now. If Beema/Mullins is as big of an improvement in efficiency as AMD claims they may at get a few designs through sheer merit, even then it's just a maybe.

 

[Arachnotronic]

That's part of AMD's freefall issue, it's pretty obvious they haven't been able to pay to gain OEM design wins for at least 6 perhaps 9 years now. If Beema/Mullins is as big of an improvement in efficiency as AMD claims they may at get a few designs through sheer merit, even then it's just a maybe.

LOL, do you think Intel runs a charity? You think that all of those BILLIONS in PC profits are derived from Intel *paying* OEMs to use its chips?

AMD's SoCs are missing quite a few critical blocks to make them suitable for tablets. This will get worse as AMD has neither connectivity nor comms in-house and integration of these IPs into SoCs for most tablets will be essential for long-term success (and this is why Qualcomm is doing such a good job).

AMD took a netbook chip without particularly effective power management/low power states and tried to down-clock it and shove it into a tablet. The end result is poor performance, a bloated bill of materials, and low battery life.

...but yeah, it's all about AMD not "paying" for its designs.

 

[Vesku]

LOL, do you think Intel runs a charity? You think that all of those BILLIONS in PC profits are derived from Intel *paying* OEMs to use its chips?

It's been proven they pay to get design wins, they did it privately during P4 days they are publicly doing it now to get Baytrail into tablets. I'd be surprised if they didn't make some payments and price concessions even during their return to technical superiority with Core2 to present for desktop and notebooks. Intel takes some of the money their customers give them and uses it to ensure competitors find it that much harder to sell their own versions to those customers, you are welcome to try to prove this statement wrong. In no way did I imply they pay out so much they are unprofitable as a company. Although, Intel has publicly admitted this year that they are on track to be unprofitable in mobile due in large part to paying for design wins.

Back on the main topic:

Those two die pics are not the same. The black areas are different.

Everything seems to match up, the only potential differences I see are in this circled region. But it could just be the different method they used to highlight the transistor blocks.

If those are real differences they are in the GPU portion of the die. Is Kabini GCN 1.0, 1.1 without true audio? If so that could be the difference, Beema/Mullins may be latest GCN versus Kabini not quite latest.

 

[Arachnotronic]

It's been proven they pay to get design wins, they did it privately during P4 days they are publicly doing it now to get Baytrail into tablets. I'd be surprised if they didn't make some payments and price concessions even during their return to technical superiority with Core2 to present for desktop and notebooks.

During the Pentium 4 days, AMD took pretty significant share from Intel in both PCs and servers because Intel's chips weren't competitive at all on performance, power, or integration. It was a mess for Intel and a wonderful time for AMD.

"Bribes" didn't get Intel out of the mess that it was in, only the roll-out of unequivocally better products did that.

Intel is in a similar mess today with its mobile parts, except the CPU isn't the problem - it's the total SoC package. If Intel can put out an unequivocally leadership product and continue to extend this lead, then it will be fine. If it can't ever do that, then I would think the company's long-term viability would be in very serious question as the world is going mobile whether Intel wants to admit it or not.

Although, Intel has publicly admitted this year that they are on track to be unprofitable in mobile due in large part to paying for design wins.

They're providing a bill of materials off-set in order to make their chips more competitive on a "cost to implement" basis for the OEM. They're not "paying OEMs to use their chips" in the way that you are implying.

 

[Abwx]

During the Pentium 4 days, AMD took pretty significant share from Intel in both PCs and servers because Intel's chips weren't competitive at all on performance, power, or integration. It was a mess for Intel and a wonderful time for AMD.

"Bribes" didn't get Intel out of the mess that it was in, only the roll-out of unequivocally better products did that.

That s not the thread but you are unaware of the real scale of these illegal activities , likely that 10-15bn werepoured in a matter of 5 years.

Intel is in a similar mess today with its mobile parts, except the CPU isn't the problem - it's the total SoC package. If Intel can put out an unequivocally leadership product and continue to extend this lead, then it will be fine. If it can't ever do that, then I would think the company's long-term viability would be in very serious question as the world is going mobile whether Intel wants to admit it or not.

Their real problem is that they have the money but are totaly without imagination, for instance back in the late 70s Samsung did only product very low cost gear and currently they are 5 times Intel revenue wise, real investors are just hedging the stock within a convenient range while others are left waiting for a rise in value that will likely never happen.

They're providing a bill of materials off-set in order to make their chips more competitive on a "cost to implement" basis for the OEM. They're not "paying OEMs to use their chips" in the way that you are implying.

Their product is so uncompetitive that giving it for free is not enough to convince OEM to produce Intel based items, hence Intel is subsiding part of the BOM, wouldnt you call that paying OEMs since Intel get massive losses to grant such subsides.?..

 

[Vesku]

During the Pentium 4 days, AMD took pretty significant share from Intel in both PCs and servers because Intel's chips weren't competitive at all on performance, power, or integration. It was a mess for Intel and a wonderful time for AMD.

"Bribes" didn't get Intel out of the mess that it was in, only the roll-out of unequivocally better products did that.

Intel is in a similar mess today with its mobile parts, except the CPU isn't the problem - it's the total SoC package. If Intel can put out an unequivocally leadership product and continue to extend this lead, then it will be fine. If it can't ever do that, then I would think the company's long-term viability would be in very serious question as the world is going mobile whether Intel wants to admit it or not.



They're providing a bill of materials off-set in order to make their chips more competitive on a "cost to implement" basis for the OEM. They're not "paying OEMs to use their chips" in the way that you are implying.

Am I implying anything other than Intel pays money, some of it directly to electronic manufacturers, to ensure its products end up in the hands of as many consumers as possible? No need to drive the discussion into the weeds to explain each payout.

There s a block above the left end of the ellipse that was added, seems that the GPU was improved and there are possibly GCN 2.0 CUs, the CPUs look the sames but a few mods at very low level would be indeed invisible on a die comparison.

Ah, yes I see some possible tweaks in that upper left block. Last block in my main elipse and that block you point out should be part of the Fusion Controller Hub (FCH). So guessing from what is visible, assuming it actually is a new die shot and not just a differently masked kabini image, updated GCN GPU and tweaked external IO compared to Kabini.

 

[Abwx]

Now that AMD has a competing product in the subsided BT segment this point become relevant for further discussions and as pointed by a member in another thread the so called contra revenue campaign is obviously a breach in the AMD/Intel deal about anti competitive practices, curious to see the outcome.

 

[Edgy]

From my experience these companies (Intel AND AMD both) deal with large OEMs very similarly with the exception of scope.

product x official selling price set at (for example) $100/unit.

Large OEMs get special discounts due to overall purchasing "volume" pricing (flat or tiered based on actual total units purchased quarterly or yearly) and/or even further discounts for any special and very specific business opportunity upon approval (such as government contracts, specific projects, education etc.,).

Without going into "special opportunity" discounts, sticking to OEM discounts only, if an OEM get pricing of $50/unit (a 50% discount).

Intel/AMD would report sales at $50/unit.

Everything's all "legal" up to this point.

Where things get fuzzy is when something called "marketing funds" come in. Intel/AMD (yes they BOTH do it) supports OEMs and large customers with these funds to push their products. And these marketing funds are supposed to be used for well... marketing (such as "Intel Inside" commercials, product designs, etc.,) but more often than not certain % of these funds are used to pad the bottom line (by internally "lowering" cost per unit) or just applied to profits.

These marketing funds are typically not big (certain in AMD's case they can't afford it) but if you remember Dell/Intel fiasco where Dell would not use AMD chips during the Athlon dominance era even when Dell's customers (mostly server side) were clamoring for AMD solutions - well, Dell was caught trying to explain additional billion or so in profits that seemed to have no valid source. Yep - that's Intel throwing their marketing money around - so much so that Dell couldn't find legit ways to "spend" it to make it look good (this is my opinion btw).

They (Intel/Dell) were stupid enough to have kept records/communications on this and other "pressuring" tactics with OEMs which were subpoenaed for trial and had to pay AMD for their anti-competitive actions.

Bottom line is that OEM design win is and always will be MUCH tougher for AMD than Intel - at least while Intel has the marketing money to throw around that's greater than total AMD revenue.

That's not to say AMD is a saint (you should hear some of things they do to shoot themselves in the foot) but to me, this is the main reason why much of AMD's design "wins" are hardly ever innovative products and are either low cost segment or are just ones of convenience for the OEMs - meaning the OEM in question probably did not have much invest much resource/time.

Having a competitive product is not enough in my opinion to overcome the marketing $$ disparity.

 

[VirtualLarry]

Everything's all "legal" up to this point.

Where things get fuzzy is when something called "marketing funds" come in. Intel/AMD (yes they BOTH do it) supports OEMs and large customers with these funds to push their products. And these marketing funds are supposed to be used for well... marketing (such as "Intel Inside" commercials, product designs, etc.,) but more often than not certain % of these funds are used to pad the bottom line (by internally "lowering" cost per unit) or just applied to profits.

These marketing funds are typically not big (certain in AMD's case they can't afford it) but if you remember Dell/Intel fiasco where Dell would not use AMD chips during the Athlon dominance era even when Dell's customers (mostly server side) were clamoring for AMD solutions - well, Dell was caught trying to explain additional billion or so in profits that seemed to have no valid source. Yep - that's Intel throwing their marketing money around - so much so that Dell couldn't find legit ways to "spend" it to make it look good (this is my opinion btw).

They (Intel/Dell) were stupid enough to have kept records/communications on this and other "pressuring" tactics with OEMs which were subpoenaed for trial and had to pay AMD for their anti-competitive actions.

Bottom line is that OEM design win is and always will be MUCH tougher for AMD than Intel - at least while Intel has the marketing money to throw around that's greater than total AMD revenue.

That's not to say AMD is a saint (you should hear some of things they do to shoot themselves in the foot) but to me, this is the main reason why much of AMD's design "wins" are hardly ever innovative products and are either low cost segment or are just ones of convenience for the OEMs - meaning the OEM in question probably did not have much invest much resource/time.

Having a competitive product is not enough in my opinion to overcome the marketing $$ disparity.

And that, my friends, is the PC OEM market in a nutshell.

 

[AtenRa]

Food for thought question,

What stops Qualcomm, Samsung, Nvidia, and others with big pockets from having a Contra-Revenue tactics in mobile next year ???
What Intel will try next then ?? donor its mobile SoCs for free for charity reasons ?? Clearly there is a problem with this behavior and if one starts to do it others will follow.

 

[mrmt]

These marketing funds are typically not big (certain in AMD's case they can't afford it) but if you remember Dell/Intel fiasco where Dell would not use AMD chips during the Athlon dominance era even when Dell's customers (mostly server side) were clamoring for AMD solutions - well, Dell was caught trying to explain additional billion or so in profits that seemed to have no valid source. Yep - that's Intel throwing their marketing money around - so much so that Dell couldn't find legit ways to "spend" it to make it look good (this is my opinion btw).

I think you are missing a point here. Marketing funds will only work if your product is good enough to have a very lean cost structure and good enough to fetch a good price structure, enough to cover all the costs and still make a reasonable profit.

The primary indicator of this is the gross margins, and this is the reason Intel seems to be really focused on it. If Intel has a good gross margin they might even have an inferior product, but they will have a lot of maneuvering room with OEMs and in the channel.

AMD does not need high revenues to get into these same sort of schemes, they needed high margins and lean cost structure, something they did achieve some times in the past. They were really obsessed in making the K8 and K10 die size small, and the main reason to die salvaging X3 and X2 chips were to improve the cost structure of the business, and give the OEM all the money they need. That's why they were able to offer OEMs, for example, 1 million of chips for free (but still were turned down in the specific case). In the end, AMD market share grew even with Intel marketing practices.

Today AMD is taking a battering in the OEM market because they threw every cost consideration aside with Bulldozer. You simply cannot compete selling chips twice the size of your competitor, and this is why the thing has been losing share all the way in every market segment it is in. Even if the chips were to provide good performance, AMD would still have a bad cost structure and business would suffer, albeit less because the better price structure would soften the effects. With such a cost structure, you lost one of the pillars of your business, you can't really blame your competitor when you paint yourself into a corner like that. Bulldozer poor performance is just the icing in this poison cake AMD baked for itself.

 

[NTMBK]

I think you are missing a point here. Marketing funds will only work if your product is good enough to have a very lean cost structure and good enough to fetch a good price structure, enough to cover all the costs and still make a reasonable profit.

Well that's patently untrue. Intel lost ~$1bn in mobile in the last quarter- that's not "a reasonable profit", that's pouring money into the segment to buy design wins and mind share. The only reason it is possible is because Intel is flush with cash from their other, very successful business units.