Recent reports have Intel displacing Qualcomm as the modem supplier and TSMC as the foundry for the next Apple A10 SoC. That is if you call this a credible report:
“Apple will possibly switch over from Qualcomm to Intel for modem chips as well as a new Apple system on a chip…”
I beg to differ of course for many reasons but for this blog let’s focus on the recent TSMC investor call and the information a seasoned semiconductor professional can extract from it.
A little background: The Intel – Apple rumors have been flying around since the iPhone first came to market. It is now known that, according to ex-CEO Paul Otellini, Intel had been in discussions about a mobile chip for Apple before the original iPhone design but they were skeptical of Apple’s volume projections and passed on the deal.
“There was a chip [Apple was] interested in that they wanted to pay a certain price for and not a nickel more, and that price was below our forecasted cost. I couldn’t see it,” said Otellini.
In 2010 Intel became a foundry and took another try at the Apple SoC manufacturing business with their 22nm and 14nm processes but failed to yield a contract. TSMC of course took the A8 (20nm) Apple business (iPhone 6) and today shares the A9 (14nm/16nm) business (iPhone 6s) with Samsung. The multi-billion dollar question is: Who will get the Apple iPhone 7 A10 business?
If you take a look at the TSMC earnings calltranscript there are three points to consider:
Hopefully we can all agree with point number one. If not, you can make your argument in the comment section but be sure and wear a helmet to prevent blunt force trauma.
Point number two depends on how you define “shrink.” In the traditional semiconductor definition it is a process level shrink but we know that 16FFC uses the same design rules as 16FF+ so it is not a process shrink. 16FFC does however offer smaller libraries (7 track versus 9 track) so you can call that a shrink. The smaller libraries will result in a cost/power reduction which is great for IoT but not great for high performance applications. The non-shrink version of 16FFC (which had never been publicly disclosed) will be optimized for SoC performance and power which to me is all about Apple, absolutely.
The most interesting point is the TSMC InFO packaging technology. Tom Dillinger did a nice write-up on InFO HERE. Here’s the thing about wafer-level chip-scale packaging technology, it is perfect for an SoC. Not only can you get an additional 20% in performance and 10% better power dissipation, you can also add more features (die) to the SoC, such as a modem, right?
The clincher here is that TSMC is predicting that InFO will contribute more than $100 million in revenue by Q4 2016. If you consider packaging is $2 or so per chip in revenue contribution that is a SIGNIFICANT amount of chip volume which again points to Apple using TSMC for the A10 SoC. Sound reasonable?
Also read:BATTERYGATE: Is Apple’s Samsung made iPhone 6S Core Rotten?