SAN JOSE, Calif. – Rumors are running rampant that Apple Inc. and Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC) are expanding their foundry ties-a possible blow for Samsung Electronics Co. Ltd.
With little or no fanfare, Apple and TSMC have recently entered into a foundry relationship, sources said. As reported, TSMC will make the A5 dual-core processor on a foundry basis for Apple’s iPad 2. That can be read here and here.
Apple will use TSMC’s 40-nm process for the A5, according to a source. ‘’Apple will also work with TSMC on 28-nm’’ processes, according to a source.
This could be seen as a setback for Samsung. Samsung is making the A4 processor on a foundry basis for Apple’s original iPad. Samsung is also making the processor for the iPhone. It’s unclear if Samsung will make the A5 for Apple.
Apple itself designed the A4, A5 and processor for the iPhone. The processors are based on ARM's technology.
Apple, according to the source, will use TSMC for three reasons: 1. Samsung competes with the iPhone and iPad; 2. TSMC has the highest yielding 40-nm process in the foundry world; and 3. TSMC has the most 40-nm capacity.
Samsung, Motorola, RIM and a plethora of others are also selling tablet PCs. ''Overall inventories are healthy at the moment, but there are concerns of component overbuild in tablet space,'' according to a new report from VLSI Research.
''This is driven in part by overly optimistic unit growth, especially for non-Apple tablets where there’s a plethora of new devices (we’re tracking more than 200 SKUs). Moreover, as the tablet war shifts into the pricing front, many companies will find it difficult to compete with Apple, which has a significant cost advantage over the competition,'' according to the report. ''While some of these tables will likely gain traction in the market place, many others will not. If the tablet build up turns into a glut late in the year, foundries are likely to suffer, given their aggressive capital expansion plans.''
Still, ''iPad 2 production ramping faster than expected,'' according to a report from FBR. ''Our contacts now see 1Q11 iPad production at 5.5 million units, up from 5.1 million units previously. Importantly, iPad 2 production is now set at about 2 million units, up from about 300,000 units previously, as some key bottlenecks (likely touch panel availability) were addressed in time to meaningfully ramp
For 2Q11, ''our contacts now see 7.2 million units of iPad production, up from our prior forecast of 7.0 million units. Importantly, almost all of the 2Q11 production is of iPad 2 devices, meaningfully ahead of the prior iPad 2 production ramp plans,'' according to the report.
''Our contacts expect 45 million iPads produced in 2011. Our contacts continue to expect 45 million iPad builds in calendar 2011, meaningfully ahead of the Street. When looking at the linearity of 1H11 (13 million units) and 2H11 (32 million units) iPad production, it seems Apple is planning on some very large iPad 2 sales volumes in 2H11 in order to achieve its annual target, with even cheaper price points likely necessary in order to achieve these 2H11 volumes,'' according to the report.
Hardly surprising. Samsung is a competitor of Apple in the phone and tablet segment, so it makes sense for Apple to fab their processors elsewhere. If that's a sign of things to come, Samsung will perhaps spin-off its foundry business.
Thanks for the reply/comment @PHW_#1, it looks like it could be good information and a good qualification well worth noting. Do you happen to have knowledge if Apple's technology concerns went beyond cost? Besides the "in competition" aspect, the article makes it sound as if cost is THE motivating factor for them.
Paul's comment only right about global foundry camp including Samsung. Samsung has enough market to support their own fab's capacity. TSMC defines the specs for all 45/40/28 IDM/design houses. Global foundries "matched" to TSMC's specs in 45/40. TSMC and global foundries diverted a lot in 28nm. Apple has no choice in 28nm,just like Qualcom/TI/nVidia/AMD... Majority design house can't trust global foundries for manufacturing from past experience. You need to pay entrance fee for doing 28nm desgin in TSMC. It is not like you have alternatives in advanced technology. TSMC makes a lot of money in leading edge, their gross margin is no less than Intel and a lot of design houses are paying their future RD development fee. Too bad, no real players right now can compete with TSMC.
Well I am thinking that Samsung is seriously disappointed. They seem sincere that they want to succeed in the foundry business but if every potential customer can't get past Samsung being a [potential] competitor then it's going to be a tough road ahead since Samsung competes in a LOT of areas where semiconductors are deployed. Furthermore, reported within these EETimes pages in February Otellini pointed out that the foundry biz could be a rough one over the next few years due to oversupply... and if true then this announcement has got to represent a real bad way to start 2011. I also think it's hilarious that reason #2 for Apple to change horses is 40nm production; in March, 2010 Samsung made its case for better foundry outlook based on its 4xnm manufacturing, and in early 2009 reportedly one over Xilinx also based on its 4xnm roadmap. Possibly all of this may be a manifestation of what Ottellini stated in the February 18 article... that "Foundries make money not on the leading-edge but on the trailing edge, with long running products." Samsung was the right choice, competition or no, to boot-strap a lively new compute platform while the run-on products will do better (economically speaking) with a foundry that keeps cost LOW and does a fine job with trailing edge technology.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.