Unlike all the ARM foundries all the FABS that are producing Medfield/Clovertrail 32nm wafers would have been paid for by 32nm Sandy Bridge. All the 22nm FABS that will porduce BayTrail would have been paid for by Ivy Bridhge and Haswell. I can't think of a more cheaper way to get a leading edge wafer and as ARM is at least one generation behind prcoess wise what Intel is using on its mobile prioducts is leading edge for ARM vendotrs ;-).
[and the following:
- Intel does not have to license from arm
- Intel does not have to license the graphics as they have their own (starting with Merrifield and baytrail)
- Intel does not have to license baseband from qualcomm as they have their own
- Intel manufacture their own chips instead of fabless companies like qcom and nvida who have to pay fabs like tsmc
- Intels products are better so they can price at premium]
Here's an entirely different take on margins from Luke Tomasello in comments section:
"I understand the reduced margins, but I'm not sure I understand how this is bad considering these sales will be in addition to their other products, i.e., their not cannibalizing other sales.
Example: Apple may have cannibalized some of it's own iPad margins by introducing the lower margin iPad Mini.. But Intel selling smartphone chips (at less margin) is not taking business away from any other Intel division.
So how is adding lower margin sales a bad thing when they are in addition to all other sales?"
[Makes sense; all the phones and tablets sold with Intel inside are like icing on the cake for Intel. Unless of course you are dumb enough to really believe that smartphones and tablets are cannibalizing the PCs]
Furthermore, analysts have expressed concern for Intel’s margins. As Raymond James analyst Hans Mosesmann stated in a recent research note seen by Forbes, the company could be heading for “a potential gross margin nightmare.” With the increasing popularity of lower-margin chips and rising competition from ARM in the microprocessor market, Intel could see its margins fall from the mid-60 percent range to the 50-percent range in the next two years.
But as the demand for PCs continues to decline, it is important for Intel to be able to tap into the market for mobile devices. Gartner, a leading information technology research company, announced in October that worldwide PC shipments totaled 87.5 million units in the third quarter, representing an 8.3 percent decline from a year earlier. IDC, another research firm, also showed that shipments fell more than 8 percent in the most recent three month period. It is the largest drop for the industry since at least 2001.
Taking into account the components of our CHEAT SHEET investing framework, Intel is a STAY AWAY until shares can stabilize and the company increases its exposure to non-PC products.
"“a potential gross margin nightmare.” With the increasing popularity of lower-margin chips and rising competition from ARM in the microprocessor market, Intel could see its margins fall from the mid-60 percent range to the 50-percent range in the next two years."
Intel worst case, bad, ugly margin NIGHTMARE would be if the Intel margin dropped to the same level that TSM has today. TSM margin is 48%. TSM PE ratio is 17.4. INTC PE ratio is 9.2.
I expect that Intel will at least cram the whole phone or whole tablet logic into 1 package and then sell that part for value and there will be a number of companies who will want a share of the smartphone and tablet markets. Intel will put them into the market and make them successful. They will sell the transistors and fill the fabs. There will 4 players: Samsung, Apple, QCOM and Intel.
The market is pricing in a collapse in Intel revenue and margins. If it happens, Intel will probably have to slow down the dividend increases. If it doesn't the analysts will eventually have to face their error.
Here's alexander#$%$ commentary:
We will find out more on Monday at 1pm PT @ CES2013 Intel event.
I think the event will be webcast. It was last year. That is one reason for flurry of bashing during the last couple days. The AnandTech article was eye-opening for many.
"They can't compete on pricing with x86 vs ARM. I'm in the business, I know."
If you are in the business then you know that your statement is not exactly accurate. Intel can compete on pricing. They chose to compete on value. There goal is to pile up the transistors on the chip and to create value for the customers. They will do that by putting more on the single chip, tightly control power, performance and then set pricing based on value.
"If they were strictly making chips, ..."
When the CPU, SoC or ... is in the final machine on the retailer's shelf, each CPU in that sysem has undergone ALL the same phases. People just confuse who the competitors are and fail to consider all the costs. They assign the cost of the ARM CPU out the foundry as the "cost of the CPU". There is additional engineering overhead that is part of that CPU beyond what the foundary charges. Intel customers will not have to staff those engineers (like QCOM, NVDA, ...) and can use the off the shelf parts. You have to be careful when assigning cost and price pressures.
"However, coupling the own design R&D expenses with own production, and their cost is higher. Products might be better, but cost is higher."
The Intel cost may or may not be higher. You do not know the Intel cost, nor do I suspect that you know the actual other company costs. Even if you are "part of the business" you can only estimate these things.
"You may think performance is important, but in reality it defines 1% of the decision of 1% of the market...."
1% of 1% is your guess. Right?
My guess on decision elements .... not in any particular order except #1:
1. vendor (is it an Apple, Samsung, ....) of product that can basically do the job.
2. price ("cost" to them but not the "cost" to build the device) ... most won't buy if they can't afford (except maybe the guys)
3. is the product in front of me available. They won't buy a product that is not available 8-) and many won't wait for the back order to be filled
4. battery life ... makes a difference if say 50% different
The reason that AMD is performing so poorly is that they misfired on several product designs. Big time.
Here's further commentary from marsavian:
These clowns obviously think Intel only sells one cpu sku ;-). Intel can and will easily add low power Ivy Bridge/Haswell Celerons that are 50-100% faster than A15 but still fit in nicely with their present cpu price structures because the top line Core skus are 400-500% faster than A15 so there is plenty of room to fit these cheap Celerons in without affecting the price of the higher performance skus. Anybody would think ARM was a serious threat lol but I have always told you bashers it was no contest on the performance front and performance has and will always determine cpu price.
and you clown still don't get it...?lol
Intel has to shift to $30 SoC from $200 CPU sales, what impact it will have on Revenue & margins?lol
netbooks are dead..
Laptops....AMD has enough competition with Intel
DESKTOP is shrinking too fast
MASS MARKET SHIFTS TO SMART PHONES/TABLETS and who rules that market?
QCOM/APPLE/SAMSUNG will never use INTEL inside....ARM camp solid now.........
Gadget Gorillas control........ the game now not Intel
Sentiment: Strong Sell
Here's why prices will rise, courtesy of wallisweaver and semi_equip_junkie
---- The economics of 20nm /14nm foundry process will be equally sobering... 1.7 increase of 20nm litho cost compared to 28nm means - that guy and other ARM suckers still believe ARM can manufacture "cheaply".... According to Barclay 80% of Intel's 22nm equipment is extendable to 14nm - TSMC will spend a huge chunk on capex ----
A key function of fab engineering teams is tool selection. Are you implying that TSMC and others are inferior to INTC in identifying toolsets that deliver on cost, performance, and lifecycle? What evidence do you have to support this claim?
[He's not implying anything of the sort. Can't you read? He's saying that Intel has a 22nm /14nm foundry process that allows them to use 80 percent of equipment from 22nm on the 14nm process and ARM doesn't have the same 20nm/14 process. Because they don't have any 20nm/14nm process. ARM has to spend huge Capex to get to Intel's level. And it will take at least 4 years...]