you are viewing a single comment's thread.view the rest of the posts
At 20nm they will. Not everybody will move to 20nm due to the cost and complexity so there will be more available capacity.
Can Intel be successful in the foundry business? Of course they can but it will not happen anytime soon. It took Samsung 10+ years to get to the number 4 spot. Today Intel Foundry uses the ASIC business model like Samsung did in the early days. Customers throw RTL designs over the Intel wall for physical implementation. This helps Intel learn the SoC foundry business and it protects Intel process secrets. Moving forward Intel will have to develop a fabless semiconductor ecosystem (exposing process secrets) and forge EDA and IP partnerships with the likes of ARM.
Intel will also have to avoid the competing with customers conundrum. The Intel UltraBooks are a blatant copy of Macbooks. The Intel Atom will someday compete with ARM and don’t be surprised if Intel comes out with an SoC of their own. Sounds a bit like Samsung right? Deja vu all over again. TSMC on the other hand is a pure play foundry and does not compete with customers.
My bet is: moving forward Apple will use Samsung for 28nm (iPhone 5s) and TSMC for 20nm (iPhone 6). Intel certainly has a shot at 14nm and 10nm but never ever count out TSMC.
"Intel certainly has a shot at 14nm and 10nm"
Why would AAPL go with TSMC for 20nm and then consider INTC for 14nm months later? INTC has 14nm factories up and running now and can go into full production sometime in 2013.
From a manufacturing process point of view, INTC is the clear choice because they have the compacity now at 22nm and 14nm later in 2013. And INTC has clear superior manufacturing process (tri-gate) and a clear roadmap for 14nm and beyond.
From AAPL's product roadmap point of view, it's clear that they need something to combat against Windows convertibles and hybrids which will cannibalize their 9.5" iPad in the upcoming months. Yes, you heard me right...Windows convertibles and hybrids will CANNIBALIZE the iPad! Why buy an iPad when you can get a PC and tablet in one device! Apple will need to come up with a 'MacPad' device to combat this. The 'MacPad' will need to run Intel's x86 chip to continue the MacBooks' multiplatform capabilities (MacOS, Linux and Windows). The multiplaform capabilities (and superior design) are why the macbooks are so popular; without it, macbooks sales will flop.
I forgot to mention that Windows convertibles and hybrids will not only cannibalize the 9.5" iPad but also all existing Macbooks! It is paramount that AAPL come up with a 'MacPad' device and eventually incorporate touch in all the Macbook product lines.
Ultrabooks have Microsoft o.s. win8 and not the Apple OS. This fact alone shows it couldn't be nor want to be a macbook copy.
From experiece I prefer the Ultrabook better since it has Win8 which is superior to Apple OS
So Apple is going to pay TSMC the higher 20nm prices consume a bulk of the TSMC 20nm capacity at Apple vendor pricing ? Interesting. So TSMC is going to sell their highest margin process in bulk to Apple at Apple prices.
Intel started the FPGA model foundry experiment. Achronix seems to have a different view of what Intel is capabile of than you do.
Proving Intel as a foundry business to 14nm and beyond
April 26, 2012 // Nick Flaherty
Achronix has been used as a business process leader for Intel’s move into the foundry business in the same way that FPGAs are a process technology leader for foundries, says the company’s founder John Lofton Holt.
Using the highly regular structures and redundancy of an FPGA as a process prover hasn’t been necessary with Intel’s 22nm process as that has been fully tested out for the Ivy Bridge processors launched this week. But the deal with Achronix as a startup has been about proving that Intel could work as a foundry with smaller companies, and sourcing the right IP for the deal, he says, despite Intel pointing out problems withthe fabless semiconductor model.
“This was not negotiated at a foundry deal but as a long term partnership for 22nm and 14nm and beyond to jointly develop this FPGA take it to market,” said Holt, founder and chairman of Achronix. “They needed a lead customer that had a relatively complex design flow as a superset for other customers. They got to learn from us about how to turn a great fab into a great foundry.”
Proving the business model has been effective, says Holt. “Intel have a lot more customers than the three they have announced,” he said. “They are truly a foundry with access to large customers.”