A couple billion here, a couple billion there-pretty soon, you're talking real money. Intel plans a $2 billion fab in Ireland. That investment is in addition to the $2 billion being spent on Fab 22 in Arizona and the $1.5 billion to build a 300-mm fab in Oregon.

Intel is also spending $1.5 billion to equip Fab 24, bought from Rockwell earlier this year, and $800 million to refit the old Digital Equipment fab in Massachusetts. These investments represent Intel's biggest fab capacity expansion since 1995.

What does the company plan to do with all this capacity? These new fabs won't help with the current shortage of PC processors, since the fabs won't begin production until 2001 or later.

Some capacity will be needed to shift production from Pentium III to the larger Willamette die, which will begin ramping next year. Indeed, the 1995 fab expansion enabled the transition from 486 to Pentium. But I expect the bulk of the new capacity will be used for non-PC devices.

Just a few years ago, Intel saw its flash-memory and embedded-processor businesses as mere PC infrastructure, supplying PC BIOS and storage controllers, respectively. Now, the company is actively pursuing new markets.

Intel's flash business is booming due to the huge popularity of cell phones, where its devices have become popular. But Intel wants more. The company is putting together all of the pieces-the StrongARM-2 processor, the DSP it is developing with Analog Devices and flash memory-to form the core of a third-generation (3G) phone.

Intel will offer these first as a chip set and ultimately as a single integrated device, reducing power, size and cost. With this technology portfolio and its existing relationship with cell phone makers, Intel is poised to be a major supplier to the 3G market.

Some analysts project that cell phones will be a billion-unit market within four years. Combine a big chunk of that market with opportunities in handheld and automotive computers, information appliances and network infrastructure, and you can easily fill a few new fabs.

We've heard Intel talk about new markets, but so far, they haven't contributed much to the bottom line. Yet this capacity buildup implies big plans for growth. And the price tag suggests that revenue expectations are also running high. Watch for Intel to get very aggressive once these fabs start to come online next year.

Linley Gwennap is the founder and principal analyst of the Linley Group (www.linleygroup.com), a technology analysis firm in Mountain View, Calif.