Seasoned managers know not to go overboard when times are flush. When the kids with the tans and the MBAs and the spreadsheets tell the guy whose company makes gaskets that demand is running hot and he should break ground on a new plant, he reminds them about the business cycle.

Sure things are good now, he will say, but what happens if they build that new plant and then business slows? They'll either have to cut back on production, idling that fancy new equipment (for which they'd probably still be paying), or they'll keep production going full tilt and have to cut prices to get it out the door. Neither is good for the bottom line. This doesn't mean that the gasket-maker won't go ahead and build that new plant anyway, but the memory of past slowdowns is going make him careful.

That technology companies have run into a capacity problem is nearly inevitable, notes

Banc of America Securities

equity portfolio strategist Tom McManus, given the money that was being made.

"If returns are high enough, there's a lot of smart people out there and they're going to write the business plans, seek the capital, hire the employees and set up the production lines to compete," he says. "The higher the returns are, the more people get drawn in."

What to Do?

In such a situation it's very hard to know when a capacity problem is going to hit -- although obviously you know it's going to hit eventually. People talked for years about how the personal computer would eventually turn into a commodity item, but that day was always seen as being at some point far in the future. So it came as something of a surprise when demand for PCs started to fall off -- not just for analysts and investors, but companies themselves. Thus

Dell

(DELL) - Get Report

decided it had to trim its workforce by 4% last week, and

Intel

(INTC) - Get Report

told employees Tuesday that it was implementing severe cost-cutting measures.

There are two ways that companies deal with capacity problems. One is to simply shutter operations, which is what a copper mine might do if copper prices go too low. When prices rise anew, it can take up its picks and shovels and get back to work. In the case of technology, however, the picks and shovels don't have a very long shelf life -- the semiconductor-making equipment of five years ago would not be of much use to Intel today. Because the investment in that equipment needs to pay off (and in many cases be paid off), companies have an incentive to try to keep up demand by cutting costs. A good example of this are DRAM prices, which have fallen steadily over the past half year as chipmakers' capacity utilization has dropped off.

Such capacity excesses will be a problem for technology companies for the next year or so, says

J.P. Morgan

equity strategist Doug Cliggott, but beyond that things should get a little rosier. Here is where the short shelf-life of the picks and shovels may actually be a help.

"The one big difference between the technology sector and the auto sector is product changes aren't incremental, they're revolutionary," he says. "Real, effective capacity gets destroyed very quickly."