One of the big concerns in the semiconductor industry has been that the economics of next-generation process nodes beginning with the 20-nanometer node will lead to stagnation in next-generation chips. Since the dawn of the semiconductor industry, each new manufacturing technology actually did something that seemed too good to be true. These new technologies brought more performance, lower power consumption, and less cost per transistor. However, as the cost per wafer outpaces transistor scaling, the proverbial party is set to come to an end. The question worth investigating is: How does this impact Apple (AAPL 1.07%)?
Apple's A8 -- what's the damage?
According to IHS iSuppli, Apple pays Samsung (NASDAQOTH: SSNLF), its semiconductor manufacturing partner, about $17.50 for each A7 chip on average -- this is making the bold assumption that the cheap M7 micro-controller costs about $1.50-$2. This chip is built on Samsung's 28-nanometer, high-K metal-gate process and measures at 102mm squared, with 1 billion transistors on board. Assuming that the next-generation Apple A8 packs between 1.5 billion and 2 billion transistors, and that the cost per transistor stays roughly constant between 28 and 20, this would imply an increase in manufacturing cost on the order of 50%-100%.
Assuming that Samsung's gross margin is about 40%, implying a manufacturing cost of about $10.50, this would further imply a manufacturing cost at either TSMC or Samsung of about $15-$20 -- with TSMC likely cheaper due to its scale/better yields. Assuming 40% gross margin on this part for the foundries leads us to believe that Apple could wind up paying $25-$28 per chip. This may not seem like much, but this could lead to a raw gross-margin dollar impact of nearly $1 billion, assuming 150 million units sold in a year. Note that in a next-generation iPhone, Apple will likely use more expensive components across the board.
Something's got to give
This analysis leads to a pretty uncomfortable conclusion for Apple: Something has to give. For a next-generation iPhone, Apple will likely use a higher resolution, higher-DPI, and a larger panel, all of which will drive up cost. Apple will probably also add an additional gigabyte of DRAM in its high-end devices, which again drives up cost. A larger device also requires that aluminum for the case is used, another cost addition. Apple is also likely to upgrade the connectivity combo chip, cellular modem, RF front end, and just about everything else.
This isn't going to be cheap. An Apple optimist may claim that Apple will simply squeeze its component suppliers -- all of which are likely enthusiastic at the chance to supply Apple -- and maintain roughly the same bill-of-materials cost even in a more sophisticated phone. Others could argue that such a larger phone is likely to be sold at a premium at which point Apple would be passing on the increases to the consumer. This strongly suggests that Apple will simply add a higher tier to its lineup rather than replace the current iPhone. Or, as Apple bears often suggest, Apple will simply have to take it on the chin in its gross margin and hope that the incremental sales drive profit growth.
Foolish bottom line
While the next-generation A8 processor is likely to be bigger, bolder, and higher-performing than its predecessor, the A7, Apple is going to have to pay the price of a more advanced chip. However, this is just one of the many incremental cost adders Apple is going to have to face in its next-generation iPhone, and frankly, it's plain as day why Apple has tended to stick to smaller phones even as the rest of the world has moved to larger ones: cost. Will consumers pay extra for a larger iPhone? It'll be interesting to see how the iPhone launch goes -- presumably in September -- and who ultimately ends up footing the bill for A8.