By David Wolf
Apple's recent decision to drop IBM's PowerPC platform and shift to an Intel-based architecture provoked the expected emotional furor. Many Mac users saw this as nothing short of selling out to the devil. Professional Apple-watchers saw this as another example of how savvy Steve Jobs is. And CNET, normally a reputable source of information, hurled itself into a paranoid lather about how this is all a Jobsian plot to use Intel's content management to suck-up to Hollywood at the expense of users.

Now that we are a week past the announcement and the noise and emotion have died a bit, it's time to take a cold look at what this change tells us about the changing nature of the microprocessor business, and indeed whether or not IBM belongs in it at all.

Execution: Good Enough No Longer Is

The advances in process and materials technologies that have allowed Moore's Law to reach its 40th birthday have also raised exponentially the complexity and cost of microprocessor design, development, and manufacturing. As chip sizes have shrunk and wafer sizes have grown, quality control has become more demanding, to the point where the quality of a manufacturer's process arguably have become more important than that of its products.

Even for those companies that are initially successful, sustaining the business long enough to build credibility and a compelling product roadmap is a challenge that has emasculated otherwise promising processor manufacturers. Transmeta, after launching it's highly efficient Crusoe architecture, could not sustain its initial innovation and has now sold its microprocessor business and survives – barely – as a licensor of technology. BLX, China's own national champion in the effort to create a viable, locally-sourced microprocessor, introduced its Godson-1 microprocessor in 2002, followed up with a 64-bit CPU this year. But the processor is generations behind current technology, and the effort has drained the company of cash and executive talent to the point where its future is uncertain.

For the few companies that pass these tests, the final is the cruelest: success. Technology leadership, cool products, outstanding processes and brilliant product roadmaps stand for little if you can't scale, and there is more to scaling microprocessor production than adding another line or another fab. New workers must be hired and trained up, infrastructure must be created, and facilities built, all of which assumes very long lead-times and deep pockets.

IBM has met each of the above tests with varying degrees of success, but it remains vexed by the challenge of scaling, a problem that would grow more acute as Macintosh sales continue to rise, and as the Microsoft X-Box and Nintendo Revolution begin adding their demands to IBM's production schedule.

Semiconductors in a Service Business

As IBM's engineers have worked to address their microprocessor production and development issues, chief executive Sam Palmisano is changing the business out from under them. According to Business Week on April 18, Big Blue's boss has decided that IBM needs to be liberated "from the confines of the computer industry" and instead put its resources to use to help companies "rethink, remake, and even run their businesses – everything from accounting and customer service to human resources and procurement."

This is an abrupt shift for IBM, moving the company's core even further than before from manufacturing semiconductors generally, much less the more technically demanding microprocessors. In this, Big Blue is in good company: there is a growing list of strong technology brands who have reached the conclusion that the semiconductor industry has evolved beyond the point where chips can be a viable side-business. Motorola (with ON and Freescale), Hyundai (with Hynix) and Siemens (with Infineon) have all chosen to get out of the semiconductor sector via spin offs when they each decided that such a business demands focus and independence to survive.

Big Blue's Disposable Divisions

So as the challenges to IBM's microprocessor business grow in complexity and cost, and as IBM itself grows in a very different direction, it becomes increasingly likely that Palmisano and company will decide that IBM does not belong in the semiconductor industry, and that the decision would be reached to sell or spin out the microprocessor business.

Longtime IBM watchers I have spoken to are of the opinion that semiconductors lie far too close to the heart of IBM to be discarded, and that the volume that will be coming from the game machines will make this an important division. Yet recent events argue otherwise. In flogging the company's storied PC division to Lenovo, IBM's leadership have demonstrated that they are prepared to slaughter the company's sacred cows if they deemed a poor fit with the company's long-term vision. And frankly, what better time to make such a choice – and gain a premium price for the asset – than when the business prospects of the division look so good?

Apple: Unwilling to Marry an Orphan

Which brings us back to the Apple question. IBM's decision to orphan it's semiconductor division – or the perception that such a change may lie somewhere on or beyond the horizon – would be enough to cause Steve Jobs to look elsewhere for a processor supply. A move like this on IBM's part would add a level of uncertainty to the Mac's future that Jobs can ill afford. Add to this IBM's technical and production challenges in the face of growing demand, and Apple has all of the motivation it needs to work with Intel developing a viable alternative.

Apple's move will not be without difficulty or challenge. But as Apple demonstrated first with its move from the 68XXX to the PowerPC platform, and then in its migration to OS X, the company understands how to manage such leaps as well or better than any company in the business.

About the author:
Silicon Hutong is an ongoing series of thoughts and commentaries by David Wolf, President and CEO of Wolf Group Asia'a management advisory firm providing strategic communications counsel to technology, media, entertainment, and telecommunications companies in Greater China and the Asia-Pacific region. David's opinions are his own and do not reflect those of either WGA or it's clients. Past articles can be found at www.chinatechnews.com, the Silicon Hutong Blog can be found here and David himself can be contacted at [email protected].