Samsung is considering spinning off its foundry business, a move that has profound ramifications for semiconductor manufacturing all over the world.

The news was reported by Business Korea.

There are scores of semiconductor production facilities worldwide, but only four companies that have been both willing and able to invest the billions and billions of dollars required to erect successive generations of wafer fabs capable of building integrated circuits at the next most exacting dimensions. Samsung Corp. is one of those four.

There are no longer any semiconductor companies that can justify investing in these fabulously expensive leading edge facilities based on internal demand alone. Intel Corp. (Nasdaq: INTC) and Samsung have been able to keep building new leading-edge wafer fabs only by attracting supplemental foundry business from amongst the fabless design houses. The other option is being a pure foundry business, like Taiwan Semiconductor Manufacturing Co. (TSMC) (NYSE: TSM).

Samsung's problem is that it recently lost one of the biggest, most demanding foundry customers in the world, Apple Inc. (Nasdaq: AAPL), to TSMC. And it lost that business while it was building in South Korea a new $15 billion fab announced in 2015 and due to come online next year.

Samsung could spin off its manufacturing arm and go fabless, relying on foundries, just as countless other semiconductor companies have done.

Want to know more about communications ICs? Check out our comms chips channel
here on Light Reading.

The question is who would buy the spinoff? For Intel and TSMC, picking up Samsung's foundry business would mean adding an immense amount of manufacturing capacity they then would be responsible for exploiting to its fullest.

But if losing Apple was a big enough blow that Samsung considered a spinoff in the first place, how economically viable would the spinoff be? Assume it would have Samsung's business, at least, but what other customers?

There's a flip side to that question. If Samsung is losing foundry business because potential customers are reluctant to work with a fab so closely allied to a competitor (Apple and Samsung compete in smartphones, for example), spinning off the fab might be the way to attract more foundry business.

GlobalFoundries Inc. is the fourth of the leading-edge manufacturers. It is the smallest of the four, but this might be an opportunity to get larger.

The country of China might be interested. The nation's long-term economic growth plan includes devoting $150 billion on purchases of semiconductor resources to be able to supply 70% of its internal needs (from a current 9%). With varying levels of success, Chinese investors have in recent years targeted Anadigics, Atmel, Lattice Semiconductor, Marvell Technology, Micron and PMC-Sierra among others.

There are points arguing against China bidding for Samsung's foundry business. Most of Chinese investors' acquisition targets are fabless companies. Foundry is exceedingly hard; leading edge manufacturing requires another level of sophistication beyond that, and China does not have that yet. Instead of building its own leading-edge fabs, China has been actively encouraging semiconductor manufacturers to build their plants in China. AMD, GlobalFoundries, Intel, TSMC and Samsung all either have plants there or are in the process of building one.

— Brian Santo, Senior Editor, Components, T&M, Light Reading