Should Taiwan Semiconductor Worry About Increased Competition?

It's well known that Taiwan Semiconductor had a virtual monopoly on 28-nanometer capacity during 2011 and 2012. Samsung , which looks to be on track to be the second largest general purpose foundry, managed to get its own 28-nanometer process out in high volume in time for Apple's iPhone 5s ramp, but this was years after TSMC began shipping wafers to customers.

The question, however, is whether TSMC -- which enjoyed robust margin expansion as a result of this 28-nanometer lead -- will be able to continue this pseudo-monopoly going forward.

28-nanometer is getting crowded
With Samsung, Semiconductor Manufacturing International , Global Foundries, and United Microelectronics all ready (or readying) viable 28-nanometer manufacturing technologies, the 28-nanometer market is going to get pretty crowded.

In fact, Semiconductor Manufacturing International recently scored 28-nanometer orders for Qualcomm's Snapdragon processors (the world's most popular mobile chips), and is even readying specialty processes for things like RF front end in a bid to capture even more Qualcomm business. 

The good news for TSMC is that at this point, its cost structure is probably better as the 28-nanometer factories have been running for quite some time, which means that the ability to compete on price on the part of its competitors is probably limited.

Additionally, TSMC's yields are probably better as it has had more time to improve them (which again helps TSMC's cost structure/margins). The bad news is that with more competition, average revenue per wafer could come down a bit.

Beyond 28-nanometer? We'll see.
There have been a lot of claims from many of the other semiconductor foundries claiming that the move to 20-nanometer and the 14-nanometer FinFET node will be extremely swift, but this is simply hard to believe. Moving from one manufacturing node to another has been getting harder, not easier, both from the actual manufacturing side of things as well as from the design side of things.

It seems unlikely that TSMC -- which has been ahead at 28, and now at 20 -- will all of a sudden "lose" its lead at next-generation nodes in terms of when products ramp at good yields in volume to the market. If this thesis holds true, then TSMC should continue to have cost structure and pricing power advantages at next-generation manufacturing technologies relative to its general-purpose foundry peers.

Foolish bottom line
It does seem likely that competition at 28-nanometer will intensify, which could somewhat pressure TSMC's margins and market share. However, it looks like TSMC should have a non-trivial lead over its general-purpose foundry competitors at the leading edge. If this is the case, then TSMC should be able to continue its robust revenue growth (as wafer prices continue to rise) as well as its high-40% gross margin profile.