The Apple A9 variant manufactured by TSMC. Image source: Chipworks.
Later this month, Apple (AAPL 0.64%) is expected to launch a revamped 4-inch smartphone which 9to5Mac reports will be called the iPhone SE. The device is expected to be, for all intent and purposes, the guts of the iPhone 6s crammed into the body of an iPhone 5s (with a few alterations). This includes the same ultra-powerful A9 chip found inside of the iPhone 6s.
Indeed, I believe that the fact that Apple is using this chip inside of a phone that will likely sell at a "budget" price point -- at least as far as iPhones go -- the iDevice maker is planning a substantial boost in chip performance with the A10 that will presumably power the iPhone 7/7 Plus/Pro.
The chip isn't everything, but it's a pretty big deal
Come this fall, the iPhone SE will be Apple's "budget" phone in mature markets, with the iPhone 6s/6s Plus likely seeing price reductions as well (starting at $550 and $650, respectively). All of these devices will be powered by the A9 chip.
In order to make sure that customers with the cash to pay more actually go ahead and pay more, Apple will need to differentiate the iPhone 7/7 Plus with as many interesting features and performance enhancements as possible.
I believe that the thinner design, improved camera, and likely much better display will do a good job of providing such differentiation, but Apple has made it a point each year to emphasize the CPU/graphics performance gains generation-on-generation. To really seal the deal, the A10 chip inside of the iPhone is going to need to be quite a bit faster than even the A9 inside of the iPhone 6s/6s Plus.
Micro-architectural innovation coupled with neat packaging goodness
Over the last several generations of A-series processors, Apple has enjoyed the benefits of both a migration in chip manufacturing technologies (which give something of a free boost in performance and power) as well as new chip architectures (the design work that Apple does).
This time, Apple won't have the benefit of migrating to a new manufacturing technology, which means that the main benefits will have to come from innovations in the actual design of the chip. However, there are some nice bonuses that Apple will get with the A10 that it didn't get with the A9.
Firstly, Apple is expected to be among the first, if not the first, customers to use TSMC's (TSM 2.84%) InFO chip packaging technology. This, according to the Taiwan-based chip manufacturer, is expected to lead to gains in performance as well as reductions in power consumption compared to other packaging techniques.
Next, Apple is said to be sole-sourcing the A10 from TSMC, rather than splitting it between TSMC and Samsung (NASDAQOTH: SSNLF), the latter of which is said to have an inferior-performing chip manufacturing technology than TSMC at the 14/16-nanometer node.
Although I have seen some pessimism on the Web suggesting that the A10 might be a modest performance bump much as the A8 was over the A7, I'm optimistic that Apple will deliver something as impressive as the A9 and the A7 were when they first hit the market.
Indeed, given that Apple needs to give customers every possible reason to upgrade, the iDevice maker's chip teams practically have to deliver massive performance/feature jumps with each generation.
