ARM profits up 22%. Maybe the tech economy is doing well after all.

[Siberian]

http://www.engadget.com/2012/10/23/arm-q3-2012-financials/

 

[bononos]

How much of it is at the expense of other sectors like more traditional desktop/laptop market?

 

[dma0991]

The 'tech' market you're referring to is probably limited to portable and embedded systems only.

 

[ShintaiDK]

As they write its mainly due to other segments. Like your TV now got an ARM computer etc.

 

[Siberian]

How much of it is at the expense of other sectors like more traditional desktop/laptop market?

Apple showed a slide yesterday of the iPad outselling all other PC companies.

 

[BenchPress]

Apple showed a slide yesterday of the iPad outselling all other PC companies.

Note that Apple was comparing against each PC manufacturer individually, not all combined (like you're wording it). There are still more desktops/laptops being sold than tablets.

Also keep in mind that everyone already has a PC, and will only buy a new one if the current one is either broken or too slow. So it takes years between purchases. Tablets on the other hand are a relative novelty and I bet that the vast majority of sales are people buying their very first tablet.

It's also a massive hype. People buy one just for bragging rights or because their neighbor already has one. But it's a waste of money for those who thought it could replace their desktop or laptop.

Once the dust settles I don't think tablet sales will be as phenomenal. It's just one form factor among an entire range. Of course it benefits ARM right now, but keep in mind that manufacturers are seeking to eventually bring desktop applications to tablets and this requires desktop performance. ARM has trouble delivering that, while Intel is only a process shrink away from conquering the mobile market. Plus is has a huge lead in desktop application compatibility.

So ARM's growth might be short-lived. They don't have the R&D power that Intel has, nor do they innovate on process technology. Intel has a huge margin on every chip it sells, while ARM only gets a minor fraction. That's not going to change any time soon. If ARM tablet sales threaten Intel's markets, they'll strike back real hard.

 

[Arachnotronic]

Don't cream yourself just yet - ARMH made $108M in pre-tax profits this Q. Intel made $3B in net income.

I'd say one of these companies is still doing way better than the other.

 

[Arachnotronic]

Note that Apple was comparing against each PC manufacturer individually, not all combined (like you're wording it). There are still more desktops/laptops being sold than tablets.

Also keep in mind that everyone already has a PC, and will only buy a new one if the current one is either broken or too slow. So it takes years between purchases. Tablets on the other hand are a relative novelty and I bet that the vast majority of sales are people buying their very first tablet.

It's also a massive hype. People buy one just for bragging rights or because their neighbor already has one. But it's a waste of money for those who thought it could replace their desktop or laptop.

Once the dust settles I don't think tablet sales will be as phenomenal. It's just one form factor among an entire range. Of course it benefits ARM right now, but keep in mind that manufacturers are seeking to eventually bring desktop applications to tablets and this requires desktop performance. ARM has trouble delivering that, while Intel is only a process shrink away from conquering the mobile market. Plus is has a huge lead in desktop application compatibility.

So ARM's growth might be short-lived. They don't have the R&D power that Intel has, nor do they innovate on process technology. Intel has a huge margin on every chip it sells, while ARM only gets a minor fraction. That's not going to change any time soon. If ARM tablet sales threaten Intel's markets, they'll strike back real hard.

Why in the world are we not best friends? This is the most eloquent, logical, and level-headed expression of my view of the situation that I've ever seen put to pixels.

+1 to you, sir.

 

[mrmt]

So ARM's growth might be short-lived. They don't have the R&D power that Intel has, nor do they innovate on process technology. Intel has a huge margin on every chip it sells, while ARM only gets a minor fraction. That's not going to change any time soon. If ARM tablet sales threaten Intel's markets, they'll strike back real hard.

ARM develops the ISA and the vanilla core for the generation, it stops here. Process is developed elsewhere and any customization beyond the vanilla cores falls to others.

Intel R&D goes to designing, tweaking and developing the manufacturing process, the entire deal.

If you want to know how much R&D is being spent on ARM you must add TSMC, Samsung, Apple, Nvidia, TI, and Qualcomm.

 

[Idontcare]

ARM develops the ISA and the vanilla core for the generation, it stops here. Process is developed elsewhere and any customization beyond the vanilla cores falls to others.

Intel R&D goes to designing, tweaking and developing the manufacturing process, the entire deal.

If you want to know how much R&D is being spent on ARM you must add TSMC, Samsung, Apple, Nvidia, TI, and Qualcomm.

+1

The billions and billions of R&D dollars spent by businesses that are developing the ARM-based products is what makes ARM products competitive with x86 products.

 

[Arachnotronic]

The ARM guys' R&D isn't additive among competitors in the space.

 

[pelov]

The ARM guys' R&D isn't additive among competitors in the space.

If you're looking at x86 versus ARM as ISAs?

If AMD's big announcement is that they're willing to integrate ARM cores + x86 onto the same die for more than just security, this could really hurt Intel in the long run.

Not in the performance aspect, but rather in perception -

If the only other big x86 chip maker is willing to deviate from x86 and dive into ARM, do we really want to look toward x86 as a long term platform?

Where the performance isn't needed, that's going to suddenly be a tougher sell for Intel.

As I've said before, the reason I favor ARM over x86 is because of the fierce competition that x86 has lacked for many years now. They've got more incentive to produce larger gains and provide more than an AMD or Intel has in the traditional PC space. Ultimately, that competition also means you wind up with some very good products (see A6 and Snapdragon S4 SoCs)

Adding another 2-3 x86 licenses would actually favor Intel :hmm:

 

[BenchPress]

If you want to know how much R&D is being spent on ARM you must add TSMC, Samsung, Apple, Nvidia, TI, and Qualcomm.

Sure, but each is investing that R&D to compete with each other, not with Intel. None of these can decide on their own to go head-to-head with Intel in the desktop market. Likewise, once Intel has it's mind set on the lucrative mobile market, who's going to stop them?

ARM (the company) is a happy little fish right now, but still a little fish. It has no chance of competing with Intel several years from now. So ARM (the architecture) isn't going to displace x86 any time soon. Quite the opposite is going to happen, pushing ARM (the company) back into low profit markets.

 

[pelov]

So ARM (the architecture) isn't going to displace x86 any time soon.

It's already replaced it. You're still looking at server and PC. In the very lucrative mobile market, ARM isn't the little fish but the big one. So in a time where AMD and Intel are cutting margins by close to double digits (or double digits in AMD's case) and dropping their quarterly estimates, the ARM space -- ARM and its licensees -- are seeing increasing revenue.

Intel isn't selling its own chips against Samsung's Exynos, it's also selling x86. Your analogy would only work if Intel were also creating ARM processors, but they're not and won't ever be. Thus it isn't about Samsung versus Intel in the ISA race but rather x86 against ARM, meaning AMD+Intel against ARM+all of the others.

You are right that ARM won't catch up in the PC and server space for years to come. They've still got quite a way to go.

 

[Lonyo]

ARM hasn't replaced x86, and probably never will unless Intel collapses.
They will co-exist and be somewhat irrelevant for 90% of people, as the actual ISA of the processor becomes something users don't need to be concerned about.

It's already happening with Android and Windows.

 

[ShintaiDK]

I think a few things gets wastly overestimated.

ARMs main profit expansion happens due to the introduction of smart TVs and the like. Plus the continual expansion of smartphones vs regular phones. ARM earns money per core. So ARM makes twice the money if you go from 1 core to 2 cores.

ARM is gaining nothing on x86, its x86 thats gaining on ARM.

Not to mention there are plenty of ARM supporting companies thats going very badly. Nokia is one to mention for example. With x86 its much easier to measure due to AMD, VIA and Intel. But people forget how many companies loses on ARM.

 

[dma0991]

As I've said before, the reason I favor ARM over x86 is because of the fierce competition that x86 has lacked for many years now. They've got more incentive to produce larger gains and provide more than an AMD or Intel has in the traditional PC space. Ultimately, that competition also means you wind up with some very good products (see A6 and Snapdragon S4 SoCs)

Give ARM dominance over a decade or 2 and it will still meet the same fate as x86. Bigger ARM manufacturers will devour smaller ones till we are left with 1-2 big players and monopoly happens all over again, just like WD and Seagate.

There will always be a need for more performance for users who actually use their computers/devices to do work. Till ARM actually becomes practical enough to provide similar performance to x86, Intel will still exist.

 

[BenchPress]

If AMD's big announcement is that they're willing to integrate ARM cores + x86 onto the same die for more than just security, this could really hurt Intel in the long run.

In what universe? They're already sacrificing CPU performance and GPU performance for their HSA delusion. You think sharing the die space with ARM cores as well is going to offer any net advantage whatsoever?

Not in the performance aspect, but rather in perception -

If the only other big x86 chip maker is willing to deviate from x86 and dive into ARM, do we really want to look toward x86 as a long term platform?

AMD is on the verge of being irrelevant. People stopped paying attention to what "the only other big x86 chip maker" is doing. If they add ARM cores, that's going to be perceived as something utterly desperate.

Where the performance isn't needed, that's going to suddenly be a tougher sell for Intel.

Why? They'll simply sell a smaller chip and still get a larger margin. And besides, the competition for higher performance in the mobile market is very fierce. Intel knows how to make fast chips, they just need one more process shrink to be highly competitive at power consumption and price.

As I've said before, the reason I favor ARM over x86 is because of the fierce competition that x86 has lacked for many years now. They've got more incentive to produce larger gains and provide more than an AMD or Intel has in the traditional PC space. Ultimately, that competition also means you wind up with some very good products (see A6 and Snapdragon S4 SoCs)

Intel is its own competition. They have to stay innovative to have developers create more powerful software and get people to upgrade at a regular pace. Looking at the revolutionary features of AVX2 and TSX, I think they've got their future secured.

 

[pelov]

Give ARM dominance over a decade or 2 and it will still meet the same fate as x86. Bigger ARM manufacturers will devour smaller ones till we are left with 1-2 big players and monopoly happens all over again, just like WD and Seagate.

There will always be a need for more performance for users who actually use their computers/devices to do work. Till ARM actually becomes practical enough to provide similar performance to x86, Intel will still exist.

I agree 100% with everything you've written here.

Intel still dominates in the x86 space, server and PC. They're making a pretty penny here, but even still they're not meeting their quarterly estimates and as a result the stock has been steadily dipping over the past three quarters. Whatever stranglehold Intel has here isn't enough to please neither the investors nor Intel. Clearly they need to make headway into newer markets.

Consolidation will happen on ARM just as it hsa on x86, but there's a significant difference here:

It's not limited to whoever entered it first. What I mean is, if me or you wanted to get a startup going that would use the x86 ISA and its derivatives, it would be quite literally impossible. Nobody is allowed to enter; no ifs ands or buts.

With ARM, they're allowing third parties the ability to tweak and play with the architecture as they see fit and others the full design + GPU if they need an entire product and sell it as their own SoC.

Can you buy an Intel x86 designed core and fab it yourself? Make any changes to the architecture at all?

The difference is that the x86 approach limits the competition that's allowed to enter the space. Imo as a consumer, it's ridiculous. Ultimately I'm the one who's shafted.

So while AMD and Intel were charging $1000 for CPUs, had there been another 3-4 competitors it would've been significantly cheaper.

In the end, all I want is a cheaper and better product. That's spurred on by competition and that happens to be something that the x86 space has been lacking for years now. We're looking at 15% performance bumps and being wowed... really?

 

[BenchPress]

It's already replaced it. You're still looking at server and PC. In the very lucrative mobile market, ARM isn't the little fish but the big one.

That's a premature thing to say. ARM hasn't conquered the new mobile market from x86, they just rolled into it. The big clash still has to happen, and ARM's recent success at invading an empty market is no indication of how that battle will be concluded.

So in a time where AMD and Intel are cutting margins by close to double digits (or double digits in AMD's case) and dropping their quarterly estimates, the ARM space -- ARM and its licensees -- are seeing increasing revenue.

Again that's really irrelevant in the bigger picture. Intel is suffering a tiny bit from the economic crisis, while ARM is still growing strong thank's to little competition (for now). It in no way means ARM is actually winning over Intel.

 

[pelov]

That's a premature thing to say. ARM hasn't conquered the new mobile market from x86, they just rolled into it. The big clash still has to happen, and ARM's recent success at invading an empty market is no indication of how that battle will be concluded.

Again that's really irrelevant in the bigger picture. Intel is suffering a tiny bit from the economic crisis, while ARM is still growing strong thank's to little competition (for now). It in no way means ARM is actually winning over Intel.

Intel's only going to dive in feet first at the end of Q4 2013 or early Q1 2014, though. So they still have a whole year of having it all to themselves.

The issue Intel is facing is that the PC's role is diminishing, so their Medfield and successor products have to be top notch and go into top notch designs, otherwise this downward trend is only going to steadily go downward.

i definitely think they have the capacity for it, I just think they need to realize that leaning on that Microsoft crutch is only going to get them burned in the end.

 

[BenchPress]

With ARM, they're allowing third parties the ability to tweak and play with the architecture as they see fit and others the full design + GPU if they need an entire product and sell it as their own SoC.

Can you buy an Intel x86 designed core and fab it yourself? Make any changes to the architecture at all?

Custom SoCs are overrated. Remember when processors were a bunch of individual chips on a PCB? This allowed for customization, but then Intel offered a single integrated chip which could perform all those tasks combined.

Of course SoCs are integrated, but it's simply cheaper to get an existing design with more functionality than what you strictly need, than to develop a custom one. The market is still in a state of flux since it's trying to determine the functionality that consumers are looking for, but sooner or later there will be one design which can cover the majority of the mobile market. Who's going to get there first? Not ARM or any of its licensees.

 

[taltamir]

I agree 100% with everything you've written here.

Intel still dominates in the x86 space, server and PC. They're making a pretty penny here, but even still they're not meeting their quarterly estimates and as a result the stock has been steadily dipping over the past three quarters. Whatever stranglehold Intel has here isn't enough to please neither the investors nor Intel. Clearly they need to make headway into newer markets.

Consolidation will happen on ARM just as it hsa on x86, but there's a significant difference here:

It's not limited to whoever entered it first. What I mean is, if me or you wanted to get a startup going that would use the x86 ISA and its derivatives, it would be quite literally impossible. Nobody is allowed to enter; no ifs ands or buts.

Good points. I am not even so sure ARM will fully consolidate. There have been attempts or at least desire to enter the x86 arena (Transmeta, nVidia, and the chinese government).
nVidia was blocked from doing so by the fact intel/AMD have a MANDATED duopoly due to patent laws (and international treaties enforcing said laws across borders). Transmeta and the chinese government's attempt were both severely hampered by said patents (they didn't violate the patents by using a software emulation layer rather then a proper hardware interpreter which severely hampers performance).

Unlike transmeta which was crushed due to the crippling effects of patent mandated monopolies, the chinese government does not need to turn a profit nor is it as strongly bound by said laws (since for china its only a matter of adhering to treaties).

https://en.wikipedia.org/wiki/Loongson
https://en.wikipedia.org/wiki/Transmeta
http://www.theinquirer.net/inquirer/news/1050874/nvidia-trying-x86-chip

PS. for the curious. the reason its mandated duopoly... AMD got their x86 license when they won a lawsuit against intel. Intel tried to use them as a second manufacturer for a deal with a major seller (which required there be at least 2 suppliers of each part). But then tried to shaft AMD so AMD sued and won and was awarded a non revokable x86 license by the court.
Intel Itanium debacle was an attempt to move away from x86 to become once again a mandated monopoly, but that was a dismal failure because AMD went and extended x86 to 64bit. (x86 was originally 8bit, which was extended to 16 bit and later 32bit by intel; finally to 64bit by AMD). Such an extension allows the addition of functionally while being 100% compatible with all existing software (a 64bit x86 processor can be run in 32bit mode and nothing can tell the difference. Hence 32bit OS running on 64bit processor; any compatibility issues are results of the OS not the CPU; as running 32bit windows on 64bit processor has never had a single compatibility issue).

WindowsRT is a huge blow to x86 dominance.

 

[BenchPress]

Intel's only going to dive in feet first at the end of Q4 2013 or early Q1 2014, though. So they still have a whole year of having it all to themselves.

Oh goody. A whole year.

Time isn't in ARM's favor. The mobile market has a strong desire for higher performance, and this isn't ARM's comfort zone. It's Intel's territory. ARM would rather not go there if they can help it. Meanwhile Intel is very successfully lowering the power consumption of its designs. So ARM is being forced to fight an uphill battle, only to find a giant who's running down that same hill to destroy them.

The issue Intel is facing is that the PC's role is diminishing...

Nonsense. Tablets don't replace a desktop or laptop. Intel's margins are affected by other factors. ARM hasn't stolen any market share from Intel yet, but Intel is about to steal market share from ARM.

 

[dma0991]

Intel still dominates in the x86 space, server and PC. They're making a pretty penny here, but even still they're not meeting their quarterly estimates and as a result the stock has been steadily dipping over the past three quarters. Whatever stranglehold Intel has here isn't enough to please neither the investors nor Intel. Clearly they need to make headway into newer markets.

Which is why Intel is diversifying into making lower powered processors that would meet the requirements that smartphones and tablet require. Intel isn't idling while their piece of the pie is taken away.

Consolidation will happen on ARM just as it hsa on x86, but there's a significant difference here:

It's not limited to whoever entered it first. What I mean is, if me or you wanted to get a startup going that would use the x86 ISA and its derivatives, it would be quite literally impossible. Nobody is allowed to enter; no ifs ands or buts.

With ARM, they're allowing third parties the ability to tweak and play with the architecture as they see fit and others the full design + GPU if they need an entire product and sell it as their own SoC.

Can you buy an Intel x86 designed core and fab it yourself? Make any changes to the architecture at all?

The difference is that the x86 approach limits the competition that's allowed to enter the space. Imo as a consumer, it's ridiculous. Ultimately I'm the one who's shafted.

So while AMD and Intel were charging $1000 for CPUs, had there been another 3-4 competitors it would've been significantly cheaper.

In the end, all I want is a cheaper and better product. That's spurred on by competition and that happens to be something that the x86 space has been lacking for years now. We're looking at 15% performance bumps and being wowed... really?

By that virtue, AMD would be doing quite well since they do make x86 processors too but that isn't the case since it is far more complex for a small company to undertake. Intel wants to make the most of their customers, it is obvious that they wouldn't want competition against their product.

Assuming that Intel does decide to give out more x86 licenses, to a startup company without a huge wallet, R&D expertise and fab advantage like Intel does; it will just go bankrupt in a few years just from trying to get something that is decent into the market. It might not even be able compete against Intel's offerings.

Is it ARM based offerings cheaper? No. Sure, ARM SoCs are indeed cheaper than Intel/AMDs CPUs but the cost for a tablet as a whole is approaching the price range of most low end laptops. Is it better in terms of performance? No. Most low end laptops have sufficient performance that runs circles around any ARM based tablet at the same price range.

15% performance bump is plenty generous. The performance increase we're seeing on tablets/smartphones right now is just to make up for what they actually lack, performance. What tricks does tablets/smartphones have? Bigger screen, higher resolution, more cores, slimmer, etc. A boring progression in terms of development of a newer product, and it is only a matter of time that these tricks run out of steam.