A report from Business Korea says that Samsung (NASDAQOTH: SSNLF) is considering spinning off its contract chip manufacturing business. The report goes on to say that Samsung began mulling this over "after the company lost Apple (AAPL 1.29%), the largest customer in the sector of application processors" to Taiwan Semiconductor Mfg. Co. Ltd (TSM +5.86%), or TSMC for short.
Image source: Intel.
It's hard to know if this will ultimately happen, but if it does, it could have some interesting implications for TSMC.
Without Samsung Electronics, Samsung Foundry might fall behind
Contract chip manufacturers like TSMC generally don't just build chips on the latest whiz-bang manufacturing technology. They build chips on a broad spectrum of technologies, ranging from brand-new technologies all the way to technologies that first began production a decade or more ago.
Very few chip manufacturers have been able to deliver those leading-edge technologies at a rapid cadence, with just Samsung, TSMC, and Intel (INTC +19.63%) at the forefront (with Global Foundries saying that it will be, too). Other chip-makers, like United Microelectronics Corporation (UMC +2.00%), or UMC, and Semiconductor Manufacturing International (SMI +0.00%), tend to bring technologies to market years after the top players do.
There are many chips to be built on older manufacturing technologies, so these smaller, weaker players can still go after a significant portion of the semiconductor market. However, on older generation technologies, a technology leader like TSMC has significant scale and potentially yields advantages over the smaller technology laggards (since its "mature" technologies compete with their "leading edge" technologies).
Nevertheless, I think that if Samsung were to spin off its contract chip manufacturer, the stand-alone Samsung Foundry wouldn't be able to afford the substantial R&D and capital equipment investments to stay at the leading edge of technology -- especially if it could no longer count on the captive demand from Samsung's chip design group (which, itself, has plenty of captive demand from Samsung Electronics).
In short, I think it would become another UMC or SMIC -- effectively eliminating the foundry competition that TSMC currently has in building leading-edge chips.
TSMC already captures much of the leading-edge foundry business out there, but additional share here certainly couldn't hurt.
Two potential challengers
If Samsung were to effectively be out of the leading-edge technology race, there's still Global Foundries and Intel (INTC +19.63%) that could potentially step in to take its place as the No. 2 leading-edge contract chip manufacturer.
Global Foundries talks quite a big game and has for years suggested that it would deliver leadership technologies alongside the major players, but thus far reality hasn't been so kind to the chipmaker. It couldn't develop a viable 14/16-nanometer generation technology on its own, which forced it to license Samsung's 14-nanometer LPE/LPP technologies, for example.
The chip-maker indicates that it's going to be in production on a competitive 7-nanometer technology by late 2018, but I wouldn't bet money on the company delivering on that promise.
Intel has the technology to compete effectively with TSMC and it has been making plenty of noise to suggest that it will eventually become a viable leading-edge contract chip manufacturer (right now, Intel primarily builds chips on leading-edge technologies for itself).
However, just as Global Foundries has a technology credibility problem, Intel has a contract chip manufacturing credibility problem. Even though Intel Custom Foundry has been around for more than half a decade, it's still not generating revenues that would be considered significant by any stretch of the imagination.
