In a research note from Bernstein Research via Barron's, analysts say that worldwide smartphone growth is expected to "taper off quickly" during the next few years. Although the market is expected to deliver robust growth to the tune of 16%, Bernstein Research reportedly says that the growth rate will slow to 8% in 2016, and to 5% by 2018.
The analysts note that semiconductor manufacturing giant TSMC's (TSM +2.73%) growth has been largely fueled by the growth in smartphone unit shipments, as it is the main contract manufacturer for many vendors of smartphone-bound chips. And with Wall Street expecting TSMC to grow sales at a 10% compounded annual growth rate, the Bernstein analysts seem to think that this might not be plausible in light of the cited projected smartphone unit growth.
I'd like to offer some perspective as to how TSMC could potentially hit those 10% growth estimates even if smartphone unit growth slows to below that 10% mark. To be clear, I'm not taking a stance on whether TSMC will or will not hit these growth estimates; but I am, instead, trying to explore how TSMC could potentially hit those numbers.
Content per smartphone growth
Smartphones, generally speaking, are becoming more complex over time. This means that the chips that go inside of them become more complex, and there may simply be more chips per smartphone. In that case, if TSMC can find a way to grow its average selling prices per chip and/or grow the number of chips that it sells per smartphone, its revenue growth could outgrow smartphone unit growth.
An example of such content growth would be something along the lines of Apple's (AAPL 0.17%) Touch ID. It's said that TSMC manufactures Apple's Touch ID fingerprint sensor. When Apple started including Touch ID as part of its iPhones, TSMC presumably saw a boost in revenue per iPhone -- although certainly not as large as the boost that it saw when it won the applications processor manufacturing contract for the A8.
Furthermore, as many industry analysts are so keen to point out, wafer prices for the latest-generation manufacturing technologies continue to go up. Now, with newer technologies, the transistor area reductions usually mean that a greater number of chips can be packed into a given area, so wafer price increases might not necessarily translate into a corresponding amount of revenue growth.
Semiconductor analyst Handel Jones recently predicted that the cost-per-transistor at 20 nanometers and 14/16 nanometers went up from the 28-nanometer node, and will only approach 28-nanometer levels at the 10-nanometer generation.
Assuming that Jones' numbers are reasonably accurate, and assuming customers of the foundries pack in more transistors into their future chip designs, which I believe they will, then this could lead to increased revenue per chip for the foundries.
Other applications
The smartphone market is absolutely huge, and every smartphone sold has many chips, making the opportunity that much more valuable for companies like TSMC. However, there are applications beyond smartphones that TSMC could potentially see meaningful growth from over the long term.
For example, ARM Holdings (ARMH +0.00%) estimated in a roadshow presentation that the total addressable market for chips that enable "embedded intelligence" will grow from $15 billion in 2014 to $25 billion by 2020. That's not the only opportunity, either. ARM expects the market for chips that go into automotive applications to grow from $10 billion in 2014 to $15 billion by 2015, and it thinks that the market for chips that enable "embedded intelligence" goes from $15 billion in 2014 to $25 billion in 2020.
There are other segments that ARM calls out, too, but you get the point.
Although TSMC will have to compete with other foundries for this business, it does look like there are other relatively solidly growing markets within chips that could help prop up TSMC's growth rate over the long term.
