Information Age: News, analysis & insight for IT & business leaders

Way back in the days when personal computers were still young - sometime around 1983 - I remember pulling into the sweeping drive that led to the worldwide headquarters of Zilog, the Silicon Valley company that, back then, built the chips powering most of the world's PCs.

The vice president of marketing greeted me with what seemed a forced smile, and within minutes he had launched into an energetic discussion of his company's plans. But as he spoke, his tone flattened; it soon became clear even to my inexperienced ears that the plans didn't amount to much. His company was being beaten, thrashed in fact, by an emerging super-giant down the road - Intel.

Not far away, there was a similar mood at Digital Research, led by the maverick Gary Kildall. His company had invented the first widely used desktop operating system (CP/M). But Kildall had fumbled the ball in 1981 when IBM came shopping for a PC operating system. In stepped a young Bill Gates, whose tiny company Microsoft cobbled together a product good enough to satisfy an impatient IBM, and there began Kildall's decline into alcoholism and early death.

Such stories are part of the folklore of the modern IT industry, and they help to explain why IBM's decision to sell off its the PC business, announced in December 2004, attracted such global interest. In today's market, IBM's withdrawal doesn't mean much: Only one in 20 PCs sold carries the IBM badge, and all of these are in fact built by its contract manufacturer Sanmina-SCI. Moreover, no one would argue that IBM was any kind of pacesetter in PCs, either in terms of price, innovation or business model.

But the symbolic significance of the move is huge. The shape of the modern IT world can, in large part, be traced back to IBM and its handling - and mishandling - of its PC business. It is not just that Intel and Microsoft's enormous success stems from IBM's choice of processor and operating system; it might equally be said that the glory years of Silicon Valley venture capitalism, which provided the fuel for the IT industry's extraordinary expansion, can be mapped onto the period from 1982 (the year the first IBM PC was launched) to its decision to exit first manufacturing PCs and then the market itself.

Most analysts agree that the pivotal moment occurred when IBM, panicked by the emergence of young companies building home computers that could support key office tasks such as spreadsheet manipulation, rushed out in 1981 to acquire the technologies it needed to build a PC. For reasons never satisfactorily explained, but which probably had much to do with arrogance, IBM decided it would pay an upfront fee for what became the PC-DOS operating system, and not to pay royalties. That meant Microsoft was free to license MS-DOS to others.

Initially, for IBM, and for its customers, the PC was a huge success. IBM's brand, and the flexibility and openness of the PC's design, encouraged the technologically cautious to buy computers in their millions. The IT department struggled to maintain control, as machines and suppliers proliferated.

The possibilities encouraged one venture capitalist, Ben Rosen, to invest in a rival to IBM - the Houston-based company Compaq. It became one of the most successful investments in the history of venture capital. Suddenly, Silicon Valley was full of cash. Even those that had little to do with PCs - such as Oracle, or Cisco - benefited from the investment machine.

For IBM itself, the story was more complicated. Until the early 1990s, IBM made money from PCs, although rarely as much as it did from other parts of its business. But after that, it losses mounted year after year. First Compaq, then Dell, eclipsed the company in innovation, logistics and profitability. IBM tried everything: highly automated factories, slimmed down just-in-time methods, devolved decision making, centralised decision making, highly integrated supply chains, and eventually, outsourcing almost everything. For several years in the late 1980s and early 1990s, IBM even tried to beat Microsoft at its own game with its alternative to Windows, OS/2.

For all its difficulties, three generations of management at IBM stuck with the PC, building a business with sales that reached nearly $10 billion in 2004. Now, CEO Sam Palmisano has pulled the company out in return for a payment that is a fraction of those annual sales. The significance of the move will not be lost on corporate customers or analysts: Nearly 25 years after the biggest and arguably most technologically advanced IT company in the world helped to democratise computing, it now thinks the desktop is not somewhere where it can add value.