Is IBM ready to let go of its semiconductor biz?

In a recent Financial Times report, IBM has retained Goldman Sachs "to sound out possible buyers for the business." However, IBM is still mum about the matter.

At every down cycle of the IC market, talks about the potential sale of IBM's semiconductor business have come up. In recent years, the speculation has intensified, identifying Globalfoundries as the potential buyer.

So, what's different this time around? Is it finally happening?

Here are some factors contributing to the impetus behind the talk. First, IBM's transformation from hardware supplier to software/service company is almost complete. IBM announced the sale of its low-end server business to Lenovo for $2.3 billion last month. Second, of IBM's declining hardware business, what's left of it, the microelectronics group is the biggest loser. In the latest quarter, its revenue was down by 33 percent. Third, it's been two years since Virginia Rometty took the helm at Big Blue. It's time for her to prove that she can make tough choices that her predecessors couldn't.

During the latest earnings call, Rometty noted, "Looking ahead, we continue to invest to deliver innovations for the enterprise in key areas such as big data, mobile solutions, social business and security, while expanding into new markets and reaching new clients." It's not clear how IBM's semiconductor business could possibly fit into the big picture she painted in her statement.

Before diving into the speculation, here are some basic facts we all need to know.

IBM still makes chips at East Fishkill, New York. However, IBM does not sell chips on the open market. IBM is big on semiconductor R&D, with stellar records. The company's contributions to the chip industry include a host of manufacturing processes and innovations include Silicon-germanium (SiGe) and Silicon on Insulator (SOI). IBM innovated, more than a decade ago, a new model of sharing R&D costs to develop chip technology. Called "the Common Platform," the alliance consists of IBM, Samsung and Globalfoundries.

The Common Platform is moving forward. The latest focus by the Common Platform is on the development of the 14nm bulk node which will introduce FinFET technology. Its mission is to offer the next generation of process scaling with low voltage operation, according to the alliance.

IBM has played well with public authorities and shown its industry leadership skills when it helped the state of New York revitalize the upstate New York area to create a semiconductor R&D hub. In the fall of 2011, the company promised to contribute $3.6 billion to the total of $4.4 billion five-year project, all dedicated to the future of chip making.

IBM chip business is keeping an eye on the emerging market. IBM is involved in one of the upcoming foundry projects in India, partnering with Israel's Tower Jazz and India's Jaiprakash Associates.

Nobody disputes an exemplary leadership role IBM has played in the global semiconductor industry. Even though the company continued to hack away at its workforce in its chip business, IBM, especially in its R&D activities, still commands respect and awe.

A big unanswered question, however, is what, then, IBM gets in return for all that money it is spending for research in manufacturing processes, extreme ultraviolet lithography and other advanced technologies. Assuming there is an ample evidence for fruits of the company's semiconductor R&D efforts, the next question is what plans IBM has to leverage it in its future business.

These are the questions Rometty needs to ask. If there are strong reasons to keep IBM Microelectronics division, she needs to be convinced of it.

As for a potential buyer of IBM's chip business, it's questionable if anyone would be happy to acquire IBM's fabs alone, sans access to IBM's semiconductor R&D team and IPs.

Certainly, if a buyer of IBM's chip division were to be Globalfoundries, already a partner with IBM in the Common Platform alliance, the Abu Dhabi-owned foundry is surely aware of the value of IBM researchers and patents, beyond IBM's 300mm wafer fab at East Fishkill.

As our former EE Times' colleague Peter Clarke once wrote, besides price, there are two factors that could kill a deal between IBM and any entity outside the United States (such as Globalfoundries). They are US national security and national pride.

You may call these two factors more of an emotional issue than anything else. But precisely because of that, it's hard to attach any value to the potential deal.

- Junko Yoshida
??EE Times