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

[Idontcare]

I wonder if you are just being stupid now or simply ignorant.

Let's be a little nicer to each other please.

Nokia's strategic partnership with MS is indeed bizarre. It resulted in nokia discontinuing every single non MS smartphone they had. Which included their best selling phones (IIRC Android).

This is the thing, all it takes to bring the end to a world-class seemingly indomitable business is to load it from the top with decision makers who are just daft or have serious conflicts of interest.

Intel's demise will not be from external competition, when it comes it will be purely from internal forces.

 

[videogames101]

It's hard to bet against Intel executing in the low-power space during the coming years.

 

[wlee15]

Nokia's strategic partnership with MS is indeed bizarre. It resulted in nokia discontinuing every single non MS smartphone they had. Which included their best selling phones (IIRC Android).

Nokia has never released a Android phone.

 

[taltamir]

Nokia has never released a Android phone.

You are right, a bit of research shows this was a symbian, an open source OS which was overtaken by android.

Anyways,
https://en.wikipedia.org/wiki/Nokia_N8
Ran symbian, was nokia's flagship phone, highest preorder amount in nokia's history, 4 million units sold. It was less than 5 months on the market when nokia announced their partnership with MS.

This is the thing, all it takes to bring the end to a world-class seemingly indomitable business is to load it from the top with decision makers who are just daft or have serious conflicts of interest.

Intel's demise will not be from external competition, when it comes it will be purely from internal forces.

Oh I absolutely agree there. That is true for any endeavor. It takes only a few incompetents in a position of power to bring down anything.
Companies, organizations, countries, militaries, forums... Nothing can survive top down incompetence for long.
Intel almost did it before with the P4... they corrected their course in time that time. But what of next time?

 

[BenchPress]

It's already a universal need. It's the reason Qualcomm makes so much money. But because of IPs, Qualcomm is currently the only maker of great LTE chips. So if you want very good LTE in an Intel SoC, guess who Intel is going to?

You should have a look at Intel's huge strides in digital radio technology.

You see, there's only so many tick-boxes. Once Intel has every feature for 90% of the mobile market, how can anyone offer a better value when Intel has this huge advantage in process and design technology?

 

[Idontcare]

Once Intel has every feature for 90% of the mobile market, how can anyone offer a better value when Intel has this huge advantage in process and design technology?

Better value will come in the form of lower prices at the expense of lower margins. Intel is not going to be interested in selling 1 billion handset chips per year at 35% gross margins.

If they can only sell 300million handset chips per year at 60% gross margins then that is where they will set their price, leaving the rest of the handset TAM for everyone else to fight over on the basis of accepting lower and lower margins.

When TI held the dominant position in the handset market, some 80% marketshare circa 2004, the gross margins for those chips was not all that stellar.

I keenly remember the quarterly internal satellite broadcasts held by Tom Engibous in which he would talk about how it was great to be uncontested in that marketspace but if we couldn't leverage our advantages into obtaining higher margin (higher prices to our customers) then we were in the wrong business.

Two years later we had a huge internal drive - we would either hit >50% gross margins on our handset portfolio or we'd raise prices and shrink our market share to whatever extent that the remaining revenues we received from the market resulted in >50% margins (sell fewer chips but make more money per chip, fire the people who then become redundant and unnecessary commensurate with the lower revenue).

That strategy set in motion what you see today - TI now has <5% marketshare from once having >80%. That is what maniacal focus on gross margins leads to, and Intel has shown this in the past with their decision to get out of HDTV and discrete GPU.

The only way Intel will dominate the handset marketspace is if they manage to innovate and create a must-have product (as they did in the PC and server spaces) for which nearly everyone will be willing to pay high enough prices that Intel is able to command a 60% gross margin.

If they can't then they will continue to look elsewhere in the industry for those margins. For Texas Instruments the answer there was to abandon CMOS (nearly entirely at this point) and pursue the analog market (>70% of TI's revenue is now all analog-based).

 

[ShintaiDK]

http://www.xbitlabs.com/news/cpu/di...hipmakers_Come_and_Go_the_Strongest_Wins.html

Seems Intel doesnt have a doubt in the area.

The world&#8217;s largest chipmaker seems confident in the outcome. Chief executive of the company compares ARM partners to companies like Transmeta (which designed ULP x86-compatible chips, but eventually went bankrupt) or Via Technologies (which offered low-power low-cost chips, but also failed to gain market share) and reckons that Intel's CPUs are the best and are getting better quicker than everything else.
&#8220;I happen to be around long enough to remember those guys. People come and go, and we have never had an exclusive, if you will. And, overall, the best chip has won,&#8221; stressed Mr. Otellini.

 

[Idontcare]

&#8220;I happen to be around long enough to remember those guys. People come and go, and we have never had an exclusive, if you will. And, overall, the best chip has won,&#8221; stressed Mr. Otellini.

Well, that is a bit of history revisionism in action there if you ask me because all those "people" that came and went found themselves going away because Intel was aggressively engaging in anti-competitive practices to ensure their chips won out at the end of the day.

It is only natural for Otellini to want to white-wash Intel's history of the sordid methods they employed to hasten the demise of their competitors. Intel's chips may be better now, but they have no excuse considering more R&D bucks are spent on them in the first place...but the reason Intel has more R&D bucks to spend on their chips today is because Intel did some unsavory anti-competitive stuff back when they didn't have more bucks to brute force their chips over the top of the competition.

So to claim that it was "the best chip" that won is less than fully ingenuous, it would be a little closer to the truth if Otellini had said "the best management team has won", a management team that engaged in anti-competitive practices and successfully undermined their competitors to the point where they became revenue starved and then couldn't keep up with the R&D budget of an ever larger Intel.

 

[ShintaiDK]

Same can be said about MS and alot of other companies

Business aint for the feinthearted.

 

[BenchPress]

Rumors are stating that Haswell ULVs will be even more expensive than Ivy ULVs.

There's competing on price and then there's feigning competition. Right now, at least as far as Ultrabooks are concerned, Intel is feigning competition and maintaining very high prices while OEMs are operating at sub-7% margins.

Same argument. Intel will lower the price when that will result in higher profit for them through higher volume. It's that simple. It's not like anyone's beating them to this new market. And they can't sell ultrabook CPUs for cheap without also bringing down the prices of laptop and desktop CPUs. Haswell GT3 will certainly be more expensive than Ivy Bridge GT2. It's a much better chip in every possible way. That's plain obvious so why would you expect them to introduce it at a lower price? Prices will come down eventually, but only gradually and only on Intel's terms.

If OEMs have to operate at sub-7% margins, maybe they should look at how Apple succeeds to sell the MacBook Air for a decent profit. It's not Intel's problem to figure out how to do that.

So I don't see what you mean by "feigning" competition. There is no competition. Hence prices are high. Once AMD or ARM start competing, prices will drop, but Intel will still dominate the ultrabook market.

 

[Idontcare]

Same can be said about MS and alot of other companies

Business aint for the feinthearted.

No question there. I'm pro-capitalism so it never personally bothered me that these companies get creative in their efforts to convince people to buy their products versus those of their competitors.

I draw the line at the mobs, physical violence and property damage versus extortion is just unacceptable. But if people are willingly enabling others to take their money through sexy marketing or the prospects of getting some coin in their own pocket (subsidized advertising, etc) then that is just business. Fools and their money are soon to part, etc.

Lord knows I've been the fool a time or two in other aspects of life and commerce, and my money and me most certainly parted ways on those occasions. I was never swindled though, I was a willing participant thinking I was the smarter party in the deal.

It would be poor sportsmanship on my part at this juncture to claim the game was played unfairly or that the other party cheated. You can't cheat an honest man, and every time I have had my financial ass whooped and handed back to me it was because I thought I was going to get the better end of the "bargain" that was being sold to me.

Intel didn't really do anything that businesses have done for time immemorial. At the same time though, tis a tad bit disingenuous of them to act like they never played the game that way as they are now attempting to have the history books reflect

But survival of the fittest does benefit the consumer. We would not have x86-64 and multicores if not for AMD convincing a paranoid Intel that they had to move towards that space and dominate it. That happened to our benefit, and Intel's, but it didn't really work out too well for AMD.

 

[ShintaiDK]

There's competing on price and then there's feigning competition. Right now, at least as far as Ultrabooks are concerned, Intel is feigning competition and maintaining very high prices while OEMs are operating at sub-7% margins.

Whats the construction and R&D costs for OEMs again?

Margins doesnt mean pure profit you can pocket.

 

[Olikan]

http://www.xbitlabs.com/news/cpu/di...hipmakers_Come_and_Go_the_Strongest_Wins.html

Seems Intel doesnt have a doubt in the area.

like they would say: "nop, we can't compete" :awe:

 

[BenchPress]

Better value will come in the form of lower prices at the expense of lower margins. Intel is not going to be interested in selling 1 billion handset chips per year at 35% gross margins.

Intel is at least a process node ahead of everyone else (for mass production). That's how they manage to have higher margins than anyone else.

Margin is just one factor in determining total profit though. They'll use their precious fabs for whatever makes the most money. If that includes mobile chips, even at a lower margin, they won't ignore it. Looking at how the mobile market it pushing for ever higher performance and lower power consumption, that's going to be right up their alley in a few short years.

When TI held the dominant position in the handset market...

Two years later we had a huge internal drive...

TI now has <5% marketshare from once having >80%...

That is what maniacal focus on gross margins leads to, and Intel has shown this in the past with their decision to get out of HDTV and discrete GPU.

That's a cute story but Intel isn't TI. They have a massive process advantage, and they're very innovative in design as well. TI tried to increase profit by increasing margins, solely based on their market share and not technological advantage over the competition. That never works out.

Also, Intel didn't get out of HDTV. They simply expected Smart TVs to be the new hype two years ago. They entered it too soon. Smart TVs are definitely the future, but it takes a lot more than a faster CPU to make it happen. The software and content distribution are the big hurdles. Sooner or later Apple, Google or Microsoft will get the interface aspect right, and services like Netflix and GameTree TV can provide the content. These things take time to align, but the demand for better CPUs to power Smart TVs will no doubt increase. So Intel isn't out of this market at all. They simply merged the division with the mobile one and they're waiting for the right time to take on these markets.

Their discrete (GP)GPU is a different story, except that Intel is slowly conquering the graphics market from below and integrating GPU-like wider SIMD units into the CPU. But let's not derail the discussion too far. Suffice to say Intel is doing really well despite / thanks to a few lessons learned the hard way.

The only way Intel will dominate the handset marketspace is if they manage to innovate and create a must-have product (as they did in the PC and server spaces) for which nearly everyone will be willing to pay high enough prices that Intel is able to command a 60% gross margin.

Again, it's not about gross margin. It's about profit. If there's billions to be made, Intel will enter that market, regardless of margin. That said, they actually do achieve large margins (and hence billions in profit) thanks to low production cost, due to having a technical and volume advantage over everyone else.

 

[firewolfsm]

I bought into that stock too! I figured it's either them or intel, and if that's the case, it's got to be them.

 

[Khato]

The only way Intel will dominate the handset marketspace is if they manage to innovate and create a must-have product (as they did in the PC and server spaces) for which nearly everyone will be willing to pay high enough prices that Intel is able to command a 60% gross margin.

The real question here being where exactly prices need to be in order to command that 60% gross margin? Because according to TSMC's financials, they're frequently seeing gross margins in the 40-50% range just for the manufacturing side. In order for Intel to not do at least that well they'd have to be designing some chips that are way too large in die size for the intended market... but Medfield at least is right around where it should be in that regard.

It's unfortunate that sales breakdowns aren't available in the public releases, and a quick search of their financials doesn't yield a mention of what the average selling price currently is. But one figure I find elsewhere says Intel's ASP currently is around $94 per system. Which seems surprisingly low for both processor + chipset, but I guess that does equate to the list price of an H61 chipset with a low end Pentium processor. Regardless, when you consider the combined die size of processor and chipset (factoring in that the chipsets still always use an older process node) and reduce that to SoC sizes, well, then getting 60%+ margins on a $25 sale isn't quite so difficult to believe.

 

[Idontcare]

Also, Intel didn't get out of HDTV. They simply expected Smart TVs to be the new hype two years ago. They entered it too soon. Smart TVs are definitely the future, but it takes a lot more than a faster CPU to make it happen.

Uh...

Intel Exits LCOS Business - The Real Story

Gross margins on most of Intel's silicon products are expected to be 50% to 60%, but Intel needed to offer panels at competitive pricing in a market with fast-declining prices. Add to this environment the pressure of developing and commercializing a new technology - and doing it within a defined window of opportunity.

 

[Exophase]

Because it's not ARM who has done the stealing. If the iPad was ugly and had a cumbersome input method and shitty software, ARM wouldn't have saved it in any way.

And Intel isn't making tablets or tablet OSes but you're still attributing market success to them.

Here's the problem. If the mobile market is actually highly lucrative and it affects Intel's bottom line, they can and will adjust their business model to the market's needs. ARM on the other hand cannot ever get to the same level as Intel on the technological front.

Things are changing in Intel's favor on both aspects.

That's a nice theory, but until Intel shows actual willingness to license their IP it doesn't amount to much.

However, let's say they change to this. That means one of two things: either Intel is offering their low-power CPUs as IP that can be manufactured in another fab or they offer their fabs for outside utilization (perhaps with the stipulation that it must be built around an x86 core they're offering).

In the first scenario they not only lose their fab advantage but have to start over in optimizing their IP for other fabs (or need to get third party companies to do so, or the licensing parties need to do it themselves). This takes time.

In the second scenario everyone else has to port their stuff to Intel's fabs, and Intel probably has a lot of work ahead of them providing their tools and getting people to learn how to use them.

Either way is tricky for Intel because to make these solutions viable to the industry it has to cost less to license than to buy Intel's own chips. Therefore Intel is at the very least losing money vs selling chips and at worst creating competition against their chips. So long as there are markets where it's more profitable for them to sell instead of license they'll risk having to compromise.

You hint that SoC flexibility no longer matters because it's well established what a tablet/smartphone SoC contains and Intel is making just that. The problem is that Apple isn't content with such "standard" SoCs. Take for example A5X in iPad 3. It dedicates a huge amount of die space to delivering best of class graphics. Clovertrail, Intel's suitable tablet response, offers nowhere close to this capability despite being a substantially newer chip on a better process - Apple's 32nm A6X coming out pretty soon will greatly leap over it once again.

The reason why Apple can afford to make these big SoCs that they're only putting in tablets is because they have not just the tablet volume but the revenue per tablet. It's a much harder sale to negotiate a custom high-GPU performance design from Intel, specifically for Apple, because they'd make far less money from doing so.

So you can see how Apple benefits from being able to either license or design in-house everything that goes into their SoC.

So, coins instead of notes? :sneaky:

Seriously now, I don't think iPads will use Intel chips by then either. But I do expect the competition to deliver some worthy devices with Windows 8 capable of running a range of desktop applications. And that means ARM's superficial market lead starts to crumble. Apple is slowly but surely losing its dominant grip on the form factors they first introduced.

Okay, I guess let's look at the market in like a year and see how much market Windows 8 tablets have taken from Apple and other ARM-based tablets. I for one don't think desktop apps are going to be a big push at all.

I don't know about Apple losing their grip but so far there's zero indication that they're losing it to x86 devices, so it's all speculation on your part.

 

[taltamir]

Uh...

Intel Exits LCOS Business - The Real Story

That article is from 2004 and discusses their withdrawl from TV rear project LCD production.

2 years ago was 2010... I think he might be referring to (or mixing up references) with the intel notion of an atom chip in every TV for smart TVs.
Which I was not aware they actually backed out of. (but ARM processors are proving a more popular CPU for those)

 

[Idontcare]

That article is from 2004 and discusses their withdrawl from TV rear project LCD production.

2 years ago was 2010... I think he might be referring to (or mixing up references) with the intel notion of an atom chip in every TV for smart TVs.
Which I was not aware they actually backed out of. (but ARM processors are proving a more popular CPU for those)

That may be true, but I was not talking about Intel leaving the smart TV market for margins, I was talking about HDTV and them leaving it for sake of margins. Regadless which example is being discussed, benchpress does not believe Intel cares about gross margins...which I find to be kinda an odd position for him to take as it is common knowledge in the investment community.

 

[BenchPress]

Uh...

Intel Exits LCOS Business - The Real Story

My bad. I didn't know Intel ever was in the LCOS business. Anyway, they don't have any technological advantage there. They can make more net profit in other markets so it makes sense they left it. I mistakenly though you were referring to the Intel CE processors for the Smart TV market, which they didn't leave. Some press wrongfully interpreted it that way when they merged that division with the mobile one, but it's still alive and gaining ground.

Regadless which example is being discussed, benchpress does not believe Intel cares about gross margins...which I find to be kinda an odd position for him to take as it is common knowledge in the investment community.

I didn't say I don't believe Intel cares about gross margins. My point was that margins are just one factor in their profit equations. They don't necessarily need a 60% margin to enter a market, as long as it's big enough. I'm sure they can settle for less if that allows them to gain strategic market share. They're pushing really hard to use the 14 nm node for the mobile market. This isn't going to be a cheap process to use, initially, but if they sell these chips at the same margin as everyone else they'll still be able to get higher performance, lower power consumption and more features out of it. Higher margins can follow later.

 

[mrmt]

That may be true, but I was not talking about Intel leaving the smart TV market for margins, I was talking about HDTV and them leaving it for sake of margins. Regadless which example is being discussed, benchpress does not believe Intel cares about gross margins...which I find to be kinda an odd position for him to take as it is common knowledge in the investment community.

IDC, I think this point about gross margins needs further explanation.

Gross margin is just a measure of performance, but is it absolute? Not at all.

By the end of the day all business are evaluated by how much capital they employed vs how much return did they bring. I'll give you an example usins extreme cases, don't take them as real cases but as an example that black and white are far extreme and you will always stay somewhere in the grey zones. For the sake of simplicity, let's use operating profits and not operating cash flows, that CAPEX is at 10% of the revenues for both companies and that they are both zeroed on debt.

Let's think of a business where you need some 100 capital at it to keep it running, and that it generates 50 in operating profits but 100 revenues. COGS stays around 25 so you have a whopping 75% in gross margins.

But let's think of another business that generates 500 in revenues and 200, with 250 of gross profits. That's 50% in gross margins. Where would you like to be as a stakeholder?

Well, the first business is probably a fast growing one but it is very small. The second business should be a mature one, where margins aren't that big but the company carries a bigger money stick to hit where it wants. It also generates better margins in free cash flow than the first business, so it is a more profitable business despite the smaller gross margins. To add insult to the injury, the second business is prepared to hit the first when they clash. More R&D, cratering margins, you chose the weapon.

Gross margin is one of many measures of business performance and it should not be analysed alone. Every time a financial analyst writes about good or bad gross margins for a given scenario, he is assuming that all other relevant premises, such as revenues, opex and capex are going to stay constant, and more important, the reader know that.

So, what happened to Intel on the TV business? Intel simply thinks that invest on low-margins chips now isn't the best thing they can do with the money. They are not leaving for the sake of margins, but because they think that improve Atom or Core will generate more cash than sell chips to TV. If they run out of ideas on what to do with they cash, they will either return it to investors, or will go into lower margin sectors without blinking.

 

[BenchPress]

And Intel isn't making tablets or tablet OSes but you're still attributing market success to them.

Only future success, when the CPU actually becomes the discerning factor between competing tablets. Apple was years ahead of everyone else in terms of software and design. Those factors are slowly but surely levelling out, and what the device can do will mostly be determined by the SoC. This is when Intel can start to play out its advantages.

That's a nice theory, but until Intel shows actual willingness to license their IP it doesn't amount to much.

Why would they have to license their IP? In a couple years from now they'll be able to deliver SoCs that are suitable for 90% of all mobile devices. Just like in the desktop/laptop markets, there's no need for others to license IP. They buy the whole chip.

You hint that SoC flexibility no longer matters because it's well established what a tablet/smartphone SoC contains and Intel is making just that. The problem is that Apple isn't content with such "standard" SoCs. Take for example A5X in iPad 3. It dedicates a huge amount of die space to delivering best of class graphics. Clovertrail, Intel's suitable tablet response, offers nowhere close to this capability despite being a substantially newer chip on a better process - Apple's 32nm A6X coming out pretty soon will greatly leap over it once again.

They're still running out of tick-boxes. There will be OpenGL ES 3.0 / Direct3D 11 support next, but it takes time for mobile applications to make use of it, and there's no point in making everything run at over 60 FPS. So things start to stall, and the competition catches up.

The same thing happened in the desktop graphics market. Intel was well behind on everyone else. But year after year they implemented better API support and performance became adequate for the masses. It doesn't really matter that NVIDIA has a discrete GPU that's ten times faster, Intel is still gaining market share.

Okay, I guess let's look at the market in like a year and see how much market Windows 8 tablets have taken from Apple and other ARM-based tablets. I for one don't think desktop apps are going to be a big push at all.

It's actually one of the few remaining tick-boxes...

I don't know about Apple losing their grip but so far there's zero indication that they're losing it to x86 devices, so it's all speculation on your part.

It's just as much speculation on your part that they won't lose their grip. The direct competition between ARM and x86 has not started yet. It's like trying to predict which country will win a war over a colony. Looking at who got their first isn't a good indication.

 

[Exophase]

Only future success, when the CPU actually becomes the discerning factor between competing tablets. Apple was years ahead of everyone else in terms of software and design. Those factors are slowly but surely levelling out, and what the device can do will mostly be determined by the SoC. This is when Intel can start to play out its advantages.

It'll take years before Intel's Haswell-level designs are low power enough to even be a consideration for mainstream tablets. There's a reason Intel is still aggressively pushing Atom here.

But I have no reason to believe Atom is going to leapfrog ahead of ARM in performance. Rumor is that Silvermont will offer 20-30% higher IPC. They'll need to clock higher than Cortex-A15 to compete with it, let alone beat it. I really can't fathom them exacting a huge lead, and I'm sure ARM and other ARM core designers aren't sitting still in the mean time.

It doesn't help that a lot of Atom's perceived performance comes from Javascript benchmarks today, where they gain a boost from better code generation - that advantage will go down the more time passes. They also get a boost from Intel's SPEC2k scores that are pretty skewed from what you'll see in tablet software since no one uses ICC (which delivers a 25% performance increase over GCC in SPEC2k - but not nearly as much in most other programs).. we'll have to see how much people care about stuff like that in the future.

Anyway, if I'm reading what you're saying correctly, it's that Apple could outsell every other tablets in spite of having much less CPU power (sometimes a good 50% less MHz for the same Cortex-A9) because their software was faster. I don't see a reason to believe this will suddenly change. Today MacOS X still gets much better battery life on laptops than Windows, for instance.

And you're still ignoring the differentiating factor that GPUs can provide in SoCs.

Why would they have to license their IP? In a couple years from now they'll be able to deliver SoCs that are suitable for 90% of all mobile devices. Just like in the desktop/laptop markets, there's no need for others to license IP. They buy the whole chip.

Uh, you said they would change their business model if they had to. What did you think I was talking about when I said business model?

I already explained why Intel can't/won't meet Apple's SoC design wishes that they are uniquely allowed by their position in the market. I'm not explaining it again. If you think this so strongly then you tell me when you spot an Atom SoC that has a die that's huge compared to all the generally available ARM SoCs, so it can offer 2x as much graphics power. Because that sure as hell isn't what's happening with Clovertrail.

If you don't understand that Apple is interested in providing much more than what everyone else is when it comes to GPU then I don't really have anything else to say to you.

It's just as much speculation on your part that they won't lose their grip. The direct competition between ARM and x86 has not started yet. It's like trying to predict which country will win a war over a colony. Looking at who got their first isn't a good indication.

Did I ever actually say ARM won't? Nope. You think that just because I'm questioning your reasoning it means I'm supporting an opposite view. The reality is neither of us really know.

 

[anongineer]

Intel did make Atom available to TSMC. It failed miserably.

They have also made tentative foundry-ish moves: Atom + Altera FPGA, manufacturing for Netronome, Tabula, and Achronix. If Tabula or Achronix proves particularly successful, I could see Intel making a move to acquire it/them and expand its x86 + FPGA offerings.

Going forward, a farfetched idea is that Intel opens up its fabs to companies that don't compete directly, to boost fab utilization. Those that are successful and have potentially complementary IP end being acquired and assimilated.