In September of next year, Apple (AAPL -0.57%) is expected to launch new iPhone models. Inside of those new smartphones will be a new, Apple-designed applications processor likely to be branded the A11 Fusion processor.
Image source: Apple.
Current expectations are that the A11 Fusion will be manufactured in Taiwan Semiconductor Manufacturing Company's (TSM -4.86%) new 10-nanometer technology, which should allow Apple to cram a lot more functionality into the chip than it did with its prior generation A10 Fusion chip, manufactured in a 16-nanometer technology.
After the A11 Fusion, Apple will presumably release the A12 Fusion chip for the 2018 iPhone models and then the A13 Fusion chip for the 2019 iPhone models. I believe both processors will be manufactured in TSMC's 7-nanometer technology, which is expected to enable further area, performance, and power improvements over TSMC's 10-nanometer technology.
Although both the A12 Fusion and A13 Fusion are likely to be built on TSMC's 7-nanometer technology, a recent comment from TSMC co-CEO Mark Liu hinted that the technology that the A13 Fusion could use won't quite be the same as the technology that may be used to build the A12 Fusion.
TSMC to build performance-enhanced 7-nanometer tech
On TSMC's most recent earnings call, co-CEO Mark Liu said that the company will "complete developing 7-nanometer in 2017, upgrade 7-nanometer in 2018, and complete developing 5-nanometer in 2019."
What this suggests is that flagship mobile applications processors launched in the 2018 time frame will use TSMC's first-generation 7-nanometer technology. TSMC recently gave some details about its 7-nanometer technology in a paper, claiming the following improvements over its 16-nanometer technology:
- About a 40% speed gain relative to 16-nanometer tech at the same power.
- More than a 65% reduction in power compared to 16-nanometer tech at the same performance.
- 0.43X chip area scaling compared to its 16-nanometer technology.
TSMC's next generation 7-nanometer technology, which will likely be used to build mobile applications processors in the 2019 timeframe, should keep chip area unchanged (assuming identical chips; Apple will likely add more features into the A13 Fusion relative to the A12 Fusion) but deliver improvements in both power consumption and speed.
Apple has a nice path
Manufacturing technology is only one part of the performance/power equation -- the underlying chip designs are critical to defining performance and power consumption. That's why something like Apple's A-series is far more powerful than, say, chips from MediaTek, even though both have access to the same manufacturing technology.
It seems like TSMC is giving Apple (and, indeed, the rest of the mobile processor vendors) everything that it needs to keep cranking out faster, more feature-packed processors on an annual basis.
Apple's challenge from a business point of view, though, will be to sell this improved performance and increased functionality to customers. Given the lukewarm commercial performance of the iPhone 6s series of smartphones, which delivered large increases in performance from the iPhone 6 series, it's clear that customers don't buy new, higher-performance devices for performance's sake.
It's not clear how Apple can motivate developers to fully take advantage of the performance that it's delivering with these new smartphones. Developers need to design their applications to work on old hardware, which limits their ability to fully utilize all the performance that Apple delivers with its chips.
Furthermore, app developers tend to publish their apps for both iOS and Android, so even though the average iPhone is probably much more powerful than the average Android smartphone, apps' core functionality needs to be geared toward low-performance Android devices.
It'll be interesting to see what actions Apple can, and ultimately does, take to try to get users excited about increased performance.
