Intel Q213 Results

[Arachnotronic]

Exophase,

I don't think it's fair that every time Intel releases a mobile platform people get to say they weren't really trying or it was just a practice run just because they didn't sell that many of them.

If the 22nm generation falls flat, then I would become significantly less optimistic about Intel's ability to compete in this space.

That crap may have flown for Moorestown, but Medfield? Clover Trail+? How are these not real efforts? Just because they didn't change the CPU uarch doesn't mean they didn't work like crazy on everything else, think of all the effort that went into power management, integration, even migrating their fabs to SoC capability.

Sure, but without competitive CPU and GPU blocks, that effort is an important learning experience, but certainly not ready for prime-time. Clover Trail (not +) has seen success (Intel is now in the top 5 tablet SoC players) because of the home-field advantage on Windows.

In my mind the move from Moorestown to Medfield was much more significant than the move from Medfield to Merrifield will be, even if one came with much sexier CPU PPT slides.

Agree, but that is because Moorestown was just so incredibly off the mark. Medfield/CT got Intel into the SoC space, but Merrifield/Bay Trail should make these SoCs legitimately good.

 

[SiliconWars]

So I just had a look through Intel's past results and found that they have leaked a staggering $3 billion loss from their Other Intel Architecture group during the last two years.

This shows the sheer magnitude of the task facing them as they attempt to break into mobile. It also shows why AMD are staying out of it, given Nvidia are also just failing to compete and losing hundreds of millions as well.

 

[blackened23]

So, basically, the Intel doubters think that they don't have a chance in the ultra portable market. The problem here is that in the past, that they were limited to a part which received little thought, under performed, and had underwhelming efficiency. This is changing. Intel is going all-in, so to speak, with their new low power architectures; Bay Trail is the start of this. They will use their latest lithography processes (which are superior to that of TSMC by years), they will price them competitively, and they are serious about updating these parts on a twice-yearly basis. Again, you can't basis future success (or lack of) with their prior Atom part - they clearly weren't focused on mobile and the changing nature of computing. Again, this has changed. They know what they must do now, and their new CEO is commited to just that.

You (naysayers) can predict doom and gloom all you want, but if intel has a part that is competitively priced (and they WILL price it on a competitive basis), has great performance and efficiency ahead of the best ARM SOCs - there will be a market for it. There was not a market for Atom because frankly - it sucked. Intel didn't take mobile serious at that time.

Of course, Intel was doubted prior to Conroe's release as well. Meanwhile, intel knew what they had to do in order to win and they quietly executed. They didn't have their marketing department work over time with ridiculous slides ala AMD style, they delivered on a part that destroyed the competition. While performance remains to be seen, you can't discount an intel that is hellbent on winning the market. Their node advantage with 14nm later this year cannot be understated - that will be absolutely huge. They are YEARS ahead of TSMC. That combined with leading performance, efficiency, and competitive pricing.....will definitely create market demand. I'm not saying ARM SOCs will be obsolete, but if intel is competitive with Bay Trail (Atom wasn't), there will be a market and manufacturers will absolutely buy it in large numbers.

 

[krumme]

You mean "Intel doubters" or "naysayers"?

 

[frozentundra123456]

AMD is not causing Intel major problems, AMD is finding their own niche and avoiding Intel. They are giving Intel problems to some limited extent with low-end laptops and tablets (Jaguar products), and with low-end servers (again, Jaguar), but that's besides the point.

The cuts Intel are receiving now are cuts that AMD has already avoided. AMD is going up quarter by quarter (cutting expenses, marginal increases in sales, etc.) while Intel is starting to lose its grip and beginning to post losses. Intel needs to cut it's fat and find a way to make it's very, very expensive process lead work for other markets.

It would take a lot of convincing for me to believe that Intel isn't a massive company which rides upon a shrinking market, with no success at branching out so far. AMD is able to cover tablets, laptops, desktops, servers, consoles, GPU's etc. all under a much smaller budget, while Intel is left with just x86 laptops, desktops and servers, unless Bay Trail actually gains them any traction. (I am not trying to argue AMD is better off, they may be too unfocused for their own good, what I am saying is that Intel only has one real market)

Well, I would rather be "going down" from 2b profit than "going up" from less than zero profit. I am not saying Intel does not have problems, mainly pricing and entrenched ARM ecosystem, but you yourself are basically admitting that AMD is resigning itself to be a low margin bit player.

 

[CHADBOGA]

It would take a lot of convincing for me to believe that Intel isn't a massive company which rides upon a shrinking market, with no success at branching out so far. AMD is able to cover tablets, laptops, desktops, servers, consoles, GPU's etc. all under a much smaller budget, while Intel is left with just x86 laptops, desktops and servers, unless Bay Trail actually gains them any traction.

It is mind boggling to think that Bay Trail won't gain them any traction.

And tell me all about AMD's current and future success in tablets and how they have tablets covered and Intel does not.

Tell me about AMD's great future in GPU's, when their GPU division isn't even profitably and that segment will be rapidly shrinking over the next few years.

Graphics Division Now Unprofitable

The first problem I saw with the report is that in the year ago period, AMD's Graphics and Visual solutions segment reported net revenue of $367M and operating income of $31M. In the most recent quarter, the company reported net revenue of $320M and zero operating income. While I have been critical for some-time of AMD's strategy of giving away a non-trivial amount in game bundles in an attempt to gain market-share against Nvidia (NVDA) in the discrete GPU business, it is clear that this strategy is not particularly sustainable.​

 

[Arachnotronic]

So I just had a look through Intel's past results and found that they have leaked a staggering $3 billion loss from their Other Intel Architecture group during the last two years.

This shows the sheer magnitude of the task facing them as they attempt to break into mobile. It also shows why AMD are staying out of it, given Nvidia are also just failing to compete and losing hundreds of millions as well.

It costs money to break into an industry. With that kind of spending, I expect significant fruits from that labor.

Let's see how Bay Trail does when it launches later this year, and we'll have 2014 to see just how good Merrifield is. Until then, it is premature to claim that these investments have not been worth it.

 

[sushiwarrior]

So, basically, the Intel doubters think that they don't have a chance in the ultra portable market. The problem here is that in the past, that they were limited to a part which received little thought, under performed, and had underwhelming efficiency. This is changing. Intel is going all-in, so to speak, with their new low power architectures; Bay Trail is the start of this. They will use their latest lithography processes (which are superior to that of TSMC by years), they will price them competitively, and they are serious about updating these parts on a twice-yearly basis. Again, you can't basis future success (or lack of) with their prior Atom part - they clearly weren't focused on mobile and the changing nature of computing. Again, this has changed. They know what they must do now, and their new CEO is commited to just that.

You (naysayers) can predict doom and gloom all you want, but if intel has a part that is competitively priced (and they WILL price it on a competitive basis), has great performance and efficiency ahead of the best ARM SOCs - there will be a market for it. There was not a market for Atom because frankly - it sucked. Intel didn't take mobile serious at that time.

Of course, Intel was doubted prior to Conroe's release as well. Meanwhile, intel knew what they had to do in order to win and they quietly executed. They didn't have their marketing department work over time with ridiculous slides ala AMD style, they delivered on a part that destroyed the competition. While performance remains to be seen, you can't discount an intel that is hellbent on winning the market. Their node advantage with 14nm later this year cannot be understated - that will be absolutely huge. They are YEARS ahead of TSMC. That combined with leading performance, efficiency, and competitive pricing.....will definitely create market demand. I'm not saying ARM SOCs will be obsolete, but if intel is competitive with Bay Trail (Atom wasn't), there will be a market and manufacturers will absolutely buy it in large numbers.

It remains to be seen how much of a success Bay Trail will be. I don't think Intel ever said "JK guys that last Atom was just for s***s and giggles lawl", and I don't think Intel has ever meant that. And also, just because Intel WANTS it to succeed doesn't mean it will either.

Intel's node advantages are not what they are claimed to be. Look at their M1 (metal pitch 1) measurements over the last few nodes. TSMC 28nm is similar to Intel's 22nm. Intel may be on "14nm", but in essence it may be more similar to TSMC's 22nm than TSMC's eventual 14nm.

I have no doubt Intel can succeed with Bay Trail, but we can't go around claiming victory already just because Intel wants Bay Trail to succeed "for real this time". To me, that is a BS and apologist excuse. Companies cannot succeed by releasing products that they "don't actually care about".

Well, I would rather be "going down" from 2b profit than "going up" from less than zero profit. I am not saying Intel does not have problems, mainly pricing and entrenched ARM ecosystem, but you yourself are basically admitting that AMD is resigning itself to be a low margin bit player.

Damn straight I would rather be too I'm not saying Intel is doomed, I'm just saying these are the warning signs that I think require Intel to change their strategy in some capacity. I'm just referring to AMD as an example, because they took action (trimmed spending, focused on certain markets) and as a result they are on an upswing (which isn't saying much, because they were doing terrible before, but they're doing better). Intel looks to be on the downswing of a long and steady rise.

It is mind boggling to think that Bay Trail won't gain them any traction.

And tell me all about AMD's current and future success in tablets and how they have tablets covered and Intel does not.

Tell me about AMD's great future in GPU's, when their GPU division isn't even profitably and that segment will be rapidly shrinking over the next few years.

Well of course Bay Trail is going to sell units, it just remains to be seen whether or not it will be the success Intel needs. Not gaining any traction != not gaining enough traction.

Temash is a very viable tablet product, all that Intel sells for tablets right now are outdated. Bay Trail could change this.

GPU division is spending lots of R&D money for next gen GPU developement and this quarter is historically bad for GPU's due to the incoming BTS and holiday seasons where all of the games are released and GPUs are sold. Also this thread has nothing to do with GPUs, so frankly I'm not sure why that has any relevance here. We are talking about Intel, or at the very least processors, not GPUs

 

[CHADBOGA]

Temash is a very viable tablet product, all that Intel sells for tablets right now are outdated.

So where are all the Temash Tablets then?

Also this thread has nothing to do with GPUs, so frankly I'm not sure why that has any relevance here. We are talking about Intel, or at the very least processors, not GPUs

We are talking about GPU's because it was one of the areas of great financial windfall you saw AMD reaping.

So you brought it up, then wonder why we are talking about it.

Don't you even know what you are saying?

 

[sushiwarrior]

So where are all the Temash Tablets then?



We are talking about GPU's because it was one of the areas of great financial windfall you saw AMD reaping.

So you brought it up, then wonder why we are talking about it.

Don't you even know what you are saying?

AMD sucks at dealing with OEMs. One of their major weaknesses. Hence their financials.

AMD sells GPUs, Intel does not. Hence Intel is not as diverse as AMD. AMD's GPU division could be a multi billion dollar sinkhole every year for all I care, that doesn't change the fact that AMD is more diverse than Intel, irregardless of how successful they are.

 

[ShintaiDK]

AMD sucks at dealing with OEMs. One of their major weaknesses. Hence their financials.

AMD sells GPUs, Intel does not. Hence Intel is not as diverse as AMD. AMD's GPU division could be a multi billion dollar sinkhole every year for all I care, that doesn't change the fact that AMD is more diverse than Intel, irregardless of how successful they are.

Intel is far more diverse than AMD. Not sure how you ever could come to the conclusion of the opposite.

 

[SOFTengCOMPelec]

Intel is far more diverse than AMD. Not sure how you ever could come to the conclusion of the opposite.

You are SO correct!

Intel sells a really wide, diverse range of items.

From memory.

Cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, and even cpus.

Compared, e.g. to Samsung, Intel is NOT diverse, in my opinion.

Intel does sell the odd item, that is NOT a cpu, but they are NOT major market players with those items, except perhaps chipsets (maybe).

Answer this question, if Intel stopped producing cpus, tomorrow, what would they have left to sell ?

 

[sushiwarrior]

Intel is far more diverse than AMD. Not sure how you ever could come to the conclusion of the opposite.

Intel has very limited product lines. They sell processors and SSD's.

Tablets, Laptops, Desktops, Servers (from low-end to HPC, including Atom servers), and let's say phones to be generous. On top of this, they sell SSD's. They tried to get into graphics, they failed. They left the motherboard market.

AMD, while not nearly as successful as Intel, covers a lot more markets.

Tablets, Laptops Desktops, Servers (low-end to high-end), game consoles, semi-custom, GPUs (desktop, mobile, workstation), desktop chipsets, desktop RAM, TV boxes (lol, would matter if anyone ever bought one).

Intel makes a lot more money than AMD. Intel makes better processors than AMD. Intel also pretty much ONLY makes processors. And very specific processors. They are tied to x86, which can either be like a soaring rocket or a sinking anchor... I am worried it is starting to look like the rocket has hit it's peak.

 

[blackened23]

Intel's node advantages are not what they are claimed to be. Look at their M1 (metal pitch 1) measurements over the last few nodes. TSMC 28nm is similar to Intel's 22nm. Intel may be on "14nm", but in essence it may be more similar to TSMC's 22nm than TSMC's eventual 14nm.

Do you really believe this? Since intel invested in FinFET and double patterning many years ago (TSMC hasn't),Intel is being more honest with their nodes. As honest as you can get in the silicon industry, anyway - TSMC's 20nm will be substantially worse than Intel's 22nm. The statement that TSMC 28nm is similar to intel's 22nm is hilariously incorrect. Intel's 22nm will still be better than TSMC's 20nm with better power characteristics due to the aforementioned issues.

 

[ShintaiDK]

You are SO correct!

Intel sells a really wide, diverse range of items.

From memory.

Cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, cpus, and even cpus.

Compared, e.g. to Samsung, Intel is NOT diverse, in my opinion.

Intel does sell the odd item, that is NOT a cpu, but they are NOT major market players with those items, except perhaps chipsets (maybe).

Answer this question, if Intel stopped producing cpus, tomorrow, what would they have left to sell ?

Not sure if you are serious. Because your post is simply embarrassing.

Intel sells, among other things:
Desktop and server motherboards, SSDs, network chips and cards, Wifi chips and cards, ethernet switch controllers, server chassis, NUC systems, RAID controllers, I/O expansion modules, different types of software, modem, tuner and demodulation chips and so on.

Intel has very limited product lines. They sell processors and SSD's.

Tablets, Laptops, Desktops, Servers (from low-end to HPC, including Atom servers), and let's say phones to be generous. On top of this, they sell SSD's. They tried to get into graphics, they failed. They left the motherboard market.

AMD, while not nearly as successful as Intel, covers a lot more markets.

Tablets, Laptops Desktops, Servers (low-end to high-end), game consoles, semi-custom, GPUs (desktop, mobile, workstation), desktop chipsets, desktop RAM, TV boxes (lol, would matter if anyone ever bought one).

Intel makes a lot more money than AMD. Intel makes better processors than AMD. Intel also pretty much ONLY makes processors. And very specific processors. They are tied to x86, which can either be like a soaring rocket or a sinking anchor... I am worried it is starting to look like the rocket has hit it's peak.

Its amazing you can find all those products for AMD. Yet none for Intel besides CPUs and SSDs....
AMD doesnt make desktop memory. They rebrand it. And almost all the rest you list is their CPU or GPUs. Amd has almost zero diversity. They got 3 things more or less. CPUs, GPUs and Seamicro. They actually bought Seeamicro due to their lack of diversity and losing in the 2 segments they was in.

 

[sushiwarrior]

Not sure if you are serious. Because your post is simply embarrassing.

Intel sells, among other things:
Desktop and server motherboards, SSDs, network chips and cards, Wifi chips and cards, ethernet switch controllers, server chassis, NUC systems, RAID controllers, I/O expansion modules, different types of software, modem, tuner and demodulation chips and so on.

Intel is done selling consumer motherboards. You are right about network/wifi equipment (even though you just said it in 6 different ways). An NUC is a processor and a board, so it's no different than any other mini-ITX system?

Its amazing you can find all those products for AMD. Yet none for Intel besides CPUs and SSDs....

I looked at the "products" tab of each companies website to give an even consideration Hence I missed most of Intel's enterprise products, which aren't listed there.

So last I checked it still stands that while Intel has the advantage in networking, storage, and server racks, AMD has the advantage in consoles, GPUs (mobile, desktop, workstation), memory, TV, and semi-custom/embedded. Intel is not more diverse than AMD.

AMD doesnt make desktop memory. They rebrand it. And almost all the rest you list is their CPU or GPUs. Amd has almost zero diversity. They got 3 things more or less. CPUs, GPUs and Seamicro. They actually bought Seeamicro due to their lack of diversity and losing in the 2 segments they was in.

AMD sells desktop memory, yes it is clearly rebranded. If they have "only CPUs, GPUs, and Seamicro" then Intel has only CPUs, CPUs, and CPUs (which happen to all be essentially the exact same kind of CPU). Intel doesn't even have a more diverse CPU lineup than AMD.

Intel has Haswell and Bay Trail essentially. They sell 2 architectures, one of which "wasn't even for real" as Intel fans have pointed out. AMD has PileDriver, Jaguar, and now ARM.

Anyways, I've gone and turned this whole thing into an Intel vs. AMD sh**fest, which I regret. My original point is that Intel has stopped diversifying, outside of Bay Trail. Their finances in the recent times are down. AMD has diversified, and their finances are better than they were before. I think Intel needs to have some kind of change, the downturns have the possibility to get worse.

 

[SOFTengCOMPelec]

Not sure if you are serious. Because your post is simply embarrassing.

Intel sells, among other things:
Desktop and server motherboards, SSDs, network chips and cards, Wifi chips and cards, ethernet switch controllers, server chassis, NUC systems, RAID controllers, I/O expansion modules, different types of software, modem, tuner and demodulation chips and so on.

I meant in the significant market share sense, and percentage of their overall income.

I had missed SSD, which are very big with Intel.
I heard they were pulling out of the motherboard market.
I knew about NUC but considered it small fry.

I was serious, but agree with your post, that they do make a wider range of stuff, than I realized.

 

[ShintaiDK]

Intels MPU revenue account for roughly 70% of all Intels revenue. So there is still 30% or around 15B$ that is not from CPUs.

 

[SOFTengCOMPelec]

Intels MPU revenue account for roughly 70% of all Intels revenue. So there is still 30% or around 15B$ that is not from CPUs.

Wow.
Off the top of my head, I would have said, 90..99% of revenue (NOT including letting others use their wafer plants, as I don't know the figures for it), comes from cpus, as regards Intel.

BUT, some of that 30% would dry up, if Intel completely lost their cpu market share.
E.g. Chipsets related to their cpus (not sure if Intel do chipsets now), Intel compiler would not be of much use, etc etc.

EDIT: Just incase there is any ambiguity. I am saying I agree with ShintaiDK, and I was just saying, what I WRONGLY thought the %'s were.

 

[sushiwarrior]

Intels MPU revenue account for roughly 70% of all Intels revenue. So there is still 30% or around 15B$ that is not from CPUs.

According to Intel's Q2 finances, it looks like 63% was PC CPU's, 21% was servers (part of which is CPU income), 7% was the Other Intel Architecture Group (CPUs), 8% was other/software (non-CPU). What that adds up to is a worst case of 92% CPU revenue, or a BEST case of 71% CPU revenue. Assuming that Intel sold ANY server processors whatsoever, it would appear between 75-85% of their revenue is CPU based.

EDIT: Correction, the 63% would include chipset and PC networking sales as well. Well, not quite clear on how to ballpark that figure then.

 

[ShintaiDK]

According to Intel's Q2 finances, it looks like 63% was PC CPU's, 21% was servers (part of which is CPU income), 7% was the Other Intel Architecture Group (CPUs), 8% was other/software (non-CPU). What that adds up to is a worst case of 92% CPU revenue, or a BEST case of 71% CPU revenue. Assuming that Intel sold ANY server processors whatsoever, it would appear between 75-85% of their revenue is CPU based.

EDIT: Correction, the 63% would include chipset and PC networking sales as well. Well, not quite clear on how to ballpark that figure then.

Its not exactly news. Take Intels revenue for 2012. And then look at this chart:

Then you also avoid random speculation on what is what in the different groups.

Intels 2012 revenue was 53.34B$. MPU sales accounted for 36.89B$, meaning that 16.45B$ revenue comes from non MPU products.

 

[sushiwarrior]

Its not exactly news. Take Intels revenue for 2012. And then look at this chart:

Then you also avoid random speculation on what is what in the different groups.

Intels 2012 revenue was 53.34B$. MPU sales accounted for 36.89B$, meaning that 16.45B$ revenue comes from non MPU products.

Still seems very likely that an extremely significant portion of the other $16.45B came from products which support their x86 processors. Remove chipsets and wireless hardware for laptops and I'm sure that figure isn't very significant anymore. Which follows the original point - Intel is absolutely invested in a single market segment, with no real backup plan.

 

[Arachnotronic]

Still seems very likely that an extremely significant portion of the other $16.45B came from products which support their x86 processors. Remove chipsets and wireless hardware for laptops and I'm sure that figure isn't very significant anymore. Which follows the original point - Intel is absolutely invested in a single market segment, with no real backup plan.

Sounds like Qualcomm. Without cell phones, they'd be nothing. What's your point?

 

[sushiwarrior]

Sounds like Qualcomm. Without cell phones, they'd be nothing. What's your point?

Cell phone sales are up, laptop and desktop sales are down.

 

[Arachnotronic]

Cell phone sales are up, laptop and desktop sales are down.

Yeah, and what happens to Qualcomm when everybody else has an LTE modem and a decent apps processor?