once they show up roughly a year from now, it might be really tough for Intel to get some high-volume design wins in the phone space, despite the transition to 14nm
"A large well funded company that is on a multibillion dollar quest for total market dominance that shows up tot he market with superior products may have a problem getting design wins."
Watch and learn, designers can't beat a path to Intel fast enough.
Someone else wrote an article in MF about the declining iPad sales. They now have a global market share under 30%. Looks like that A7 did wonders for them. I thought that corporate America was going to replace their aging XP computers with them? Here's what Apple needs to learn right now. For under $150 you can buy a toy, for over $400 you can buy a computer. Deal with it.
I used to pray for the day this would happen. Now, not so much. The A7, the pride and joy of the ARM community is being purchased for $18 each! These are going into tablets that are selling for over $500. They are not selling out like they once did. It seems that the A7's superior performance hasn't turned into consumer demand. I think Apple is going to turn this logic around around and say that hardware doesn't sell tablets, shiny Apple symbols do. I doubt Apple will be willing to pay a fair margin for Intel processors as they squeeze vendors to lower the BOM.
One factor that may affect this number may be roughly attributed to a shift in policy at TSMC. Up until the change you could set a contracted price per CPU. But TSMC had yield problems they said they had at 40nm and claimed they resolved and then a year later they said the same thing. When it came to 28nm they stated they were shifted to a pay per wafer model. If the wafer produce 1000 CPU's it was one price, but if it only produced 7 it was the exact same price. Some called it the "pay for many, get few" model.
One of the first 28nm parts to come off the TSMC fab was for a $600 NVidia video card. It took them about half a year to be able to make enough GPU's for these cards to satisfy demand. But I was thinking "how much demand is there for a $600 video card? 20K units perhaps? " I think Nvidia didn't care if it was paying $25 a GPU or $60, there was fat markup and backorders.
Flash forward a year and a half later at 28nm isn't the de facto node of TSMC. I think this 'pay per wafer' decision may have cost them a shot with Apple. The backup caused by this I also think can be traced to the ARM Android tablet market of today. The sub-$150 7" tablet market isn't night and day worse than the over $200. For more money you get a nicer screen, more RAM and a slightly faster processor.
If you look at many of the sub $160 7" tablets they have MediaTek SoC's which I believe are fabbed at UMC. By not pushing 28nm faster, harder, and out of their pockets TSMC left the door open for competition. I think there will be significantly more tablets sold this quarter with TSMC losing market share. to Samsung and UMC and of course the new player, Intel.
Step #5. When Intel has more business than it knows what to do they can kiss ARM goodbye.
By this time Intel will have a large enough lead that it doesn't have to worry about ARM anymore. It will be thrown under the bus.
BTW, bribes are something you usually don't disclose such as giving out $1 million in offshore funds to the first 1000 design wins for Intel Atom. These rebate programs are perfectly legal and effective. The reason why people get mad about the rebate programs is that they work a bit too well. Unlike the $1 billion Microsoft lost on the RT surface.
I'm long INTC for it will play out exactly the opposite of your claim..
BTW, I wrote about two years ago that Intel could be a FAB for ARM and Getanid61 called it "desperate" at the time. I admit that was a bit of a longshot.
The thinking of Daniel Nenni and other ARM fans is that Intel fabbing ARM is an admission of defeat. ARM can now boast about having the best fabs in the world supporting their model. The logic goes that once Intel makes fat ARM profits they will never stop fabbing for ARM. The ARM ecosystem will be supported by Intel on the high end on down. The ARM franchise will become even more valuable.
My belief is that Intel isn't interested in foundry fabbing profits as much as they are interested in market domination. ARM64 at a similar node to Intel could inflict some real damage to Intel in the long term.
Step #1- Defund ARM Foundries. The best way to do this is to take business at the high end that cuts demand at TSMC. This will #$%$ node development by cutting funding.
Step #2- Use fat ARM foundry profits to pay for bribes. The best part of the $1 billion "bribe" is that ARM is helping to fund it. Imagine if they did this for year after year until each year ARM has less and less business. The ARM profits also help pay for the Intel Capex.
Step #3-Find out where the bones are buried- Its easier to score Intel design wins when Intel is a foundry for ARM. They will have a working the designers and they can talk about the benefits of Atom. At first it will be the low priced alternative to ARM and yes at first it will be a harder sale.
Step #4- Give the goodies to Atom, not ARM. Right now Intel has a compelling CPU to ARM64, but not a hands down winner. They also can't fill in the rest of the SoC by themselves. Too many essential parts are missing. When Intel makes a better GPU, and has better wifi and bluetooth connection along with better LTE it will get people excited. That will go into Atom not ARM processor
Hey for an extra $100 or more you can have an iPad Air with an A7. You know the benefit of having an A7? For no additional charge Daniel Nenni will call you up and tell you how wonderful it is.
This holiday season we've got an abundance of 7" tablets for $150 or less, some below $100 if you're not afraid of no name brands or refurbs. Dell, Asus and other worthy tier 2 brands in the tablet market are making a name for themselves.
The problem is that from last year to this year with the exception of Apple and Samsung ARM hasn't moved that far along the spectrum. From what I can tell the new selling points are a bit faster processor and better screens running at higher resolutions, more ram and a tad bit lighter. The problem is that for what the 2013 version of the Nexus 7 costs (close to $300) you can have two and a half 2012 versions (refurbished).
Go look for yourself, there are compelling ARM Android tablets for under $150 that aren't state of the art but will more than take care of your tablet needs. Even Intel has a dog in the fight with the Dell Venue 7. This sub $150 market will grow the tablet market and make Apple's prices for the iPad Mini seem excessive at $300+. The tablet market will grow with Apple getting less and less market share. And when all those old $500 iPads and iPad2s expire what will people replace them with? Apple? Don't be so certain.
The tablet market isn't a computer market. People are interested in 'bang for the buck' as people think of tablets as simply consumer electronic devices. 1 Billion dollars and an aggressive advertising program to differentiate Intel's PC tablet offerings will stir up the market. It will go a long way to help Intel stand out in a consumer electronics market clogging with commodity ARM technology.
The new A7 processor is lost on the masses. the iPad2 was out of stock for months at a time. Now it's readily available on Amazon with 4 out of 5 star rating. There are only 40 reviews. Oh no Apple fans, looks like something is wrong!
We're going to see a little experiment next year. The experiment is what is a better $1 billion investment? Intel's investment into the tablet market or Microsoft's same investment into bringing ARM into it's market. Part one of the test is over, part two is about to start. We'll see how many Intel tablet units get sold next year. Nenni get your spin-cycle ready.
Yeah isn't it wonderful? it would be like the United Sates over throwing China using Chinese guns and money. But since Apple and Samsung didn't decide to share their ARM64 efforts with the rest of the ARM would in the true spirit of communism I mean the Fabless Model can you really point the finger at Intel for destruction of the model?
This was one of the three big talking points of ARM. Windows RT to destroy the Wintel franchise, ARM64 and Microsevers. You cannot for a moment say that Microsoft didn't take Windows RT seriously. They played commercials for it 24/7 and lost $1 Billion on it. This only proved what we've said all along, ARM processors by themselves aren't very special and don't provide any extra value. Apple not ARM brought value to iOS and Google did the same with Android.
When Microsoft and Intel decided to find new partners in their now open marriage Microsoft's partner ARMH fizzed out while Intel's partner Google seemed to get establish a new long-term relationship.
ARM64 currently hasn't made it to desktop with the A7 burning up to 8 watts of power at peak performance. Yowza! Everybody else other than Samsung seems to be a year away from ARM64. Microservers are about 10% of the server market and now AMD also wants to get into the dogfight. But the ARMH stock price is doing well which only proves ignorance is bliss.
When Windows RT gets pulled from Microsoft next year I feel that ARMH stock price should pay a penalty. But it probably won't.
TSMC has been an Intel wannabee for many years now. Intel has already surpassed TSMC in manufacturing. You find outsourcing at most manufactures to some extent. Intel has outsourced some of it's design work to ARMH which makes Intel the largest customer of ARMH.
Will Intel contract manufacture ARM processors? They have agreed to for Altera. This will be the first of many. As much as I love Intel and I am beyond impressed how quickly it can go from thinking about entering the low power processor market to actually being considered a major player as Ashraf mentioned a few times Intel still isn't in the driver seat in the market. But what I do know about Intel is that they from time to time do less than stellar efforts to enter certain markets or they get half way in and find out that their plans aren't working out and give up. But when Intel sets its sights on something they never give up. Intel's commitment to mobile is about as furious as they are with servers.
The conclusions that the analyst make about Intel and the Monday morning quarterbacking they give with current data is such old news to Intel. Intel is planning five years out and then some. The CAPEX spending and the 10 billion dollar investment into R&D means that they liked something they saw in the crystal ball. What I can see is that this isn't a reaction to market conditions but business opportunities down the road.
Yes Daniel Nenni may be trying to set me up to agree with him that Intel will be just another contract manufacturing FAB. No, Intel is so far beyond this. Intel will be a much larger and much more profitable company in the future. ARMH makes itself vulnerable to be played and Intel is in the position to do just that. Intel will grow in mobile with the help and at the expense of ARMH.
I think Chrome OS is going to find a sudden death at the expense of Android 64. I like the sound of a Nexus 15 better with 8GB of RAM running Windows in Virtual Box. If Android 64 were to become a reality software companies like Adobe may start developing for it as a third platform. It is afterall a modified version of Linux.
You know what? For once I actually agree with Daniel Nenni. But the point of breaking even isn't exactly what I see. Intel tipped it's hand a few years ago when they nearly let TSMC produce the Atom processors. I think that Intel will have so much success with the foundry business that it's going to drive up the internal rate of return. The faulty logic up til now is that Intel will make lower margins on a lower cost per unit sale. My thinking is that Intel could easily make a $35 ARM 64 SoC right now and pull in 60% GTM. But what if they charged $75 for it? They could see fat, sick profits and payback the CAPEX much faster. They could also price the Bay Trail x86 at lower price points to make it more attractive. Those who need the cutting edge technology would pay dearly for it.
Intel would establish all types of relationships much faster than trying to push X86 to replace ARM. And once the relationships are established they can always talk up the cost savings and long term advantages of X86. Intels fabs will be going night and day.
Of course this cuts into TSMC. It's had to get to the next node when many of their clients are already paying for Intel's Capex. There can be volume discounts and long term contracts established. Heck, when you have the goods you can call the shots. But Intel can look at it's aging lines like 32nm and simply decide it's more profitable to let TSMC be the foundry. It may also be to Intel's benefit to keep TSMC close to keep an eye on them. Hold your friends close and your enemies closer.
Everybody talks about Apple and how wonderful it would be for Intel to fab for the A series for Apple. They have squeezed Samsung down to $18 for an A7. So much for fat profits. Intel has gone far but still needs further to go to really dominate over ARM in mobile. A better GPU and more of QCOM's forte will be needed. Fabbing will buy time and expertise. And most importantly pay for the CAPEX investments.
Did you notice the "Intel Inside" with the familiar tones at the end of the commercial? That means Intel paid for half the commercial. The advertising payments was something I called six months ago if not longer. Looks like I'm right again!
I share Wallis's view to some extent. To me, there is a difference between having an unpopular of opinion and just being a jerk. I don't mind if people come here and share what we may feel is bad news, as long as it's relevant. Remember that kid a few years ago who came on here and told us that Intel is a dinosaur? He came on, blew his vile and left. Someone wrote back "I like dinosaurs."
The best I can figure from Nuff is that he gets paid to be annoying or he's got a fat wife with emasculates him. Daniel Nenni is a rare breed. He's a regular Kenny Powers, twice the size and half the personality. There is a profit motive to make himself look like the go-to guy for semiconductor info. But week after week he blows his credibility on here, for what reason I don't know. I have talked to semi valley IT folks who know him and they really can't stand him either. Semi Wicked has done more to turn them off because it's biased.
I have become a huge fan of Daniel's sidekick Wacky Ed. On Seeking Alpha they have never seen anything like him, and didn't know what to make of him. But now with his unfounded claims and deep, deep left field posts I think it's best to know that this is the writing talents that you'll find on Daniel Nenni's website. BTW, Daniel, people will get tired of unsubstantiated fairy tales. We got his number, which happens to also be your number.
Please find his latest post on SA.
"So now Global will be drawn into Samsung's orbit and perhaps become an acquisition target as 14nm moves to 10nm and the cost to stay in the game overwhelms all but the most endowed, such as TSMC, Intel and Samsung. All of this game changing though comes at the direction of Apple who is out in front with the most profitable mobile model and more upside to come. With its 64 bit ARM roadmap about to expand, Apple has the luxury of building larger die that stretch the performance curve upward and in synchronicity with process technology improvements"
This is investment advice or science fiction? Please do me a favor and find the article and leave some comments about it's lack of reality.