The most worrying news is that OEMs' inventories have been growing. In past downturns inventories have turned slowdowns into train wrecks. When things are good, parts go on allocation, and the OEMs overorder to make sure of needed deliveries. When the OEMs see their own businesses slow down, they decide not to cancel parts orders, not wanting to harm their positions in the order queues. So parts inventory accumulates. As the slowdown in end-user demand increases, at some point the OEM pulls the plug on parts orders, reducing the run rate at the parts supplier far faster than it was reduced at the OEM. The folks who build chips, sitting at the head of the supply chain, feel the full force of this multiplier, and can see major-account business shrivel in weeks.

Ironically enough, the slowdown will hit hardest in the companies that had the best run rates: that is, those that were already serving a volume market that is slowing down. They will be the ones to see the largest changes in order volume. Companies that weren't in full production yet, or those that were serving OEMs in an emerging market rather than a mature one, may not see an immediate impact, although slowdowns have a way of getting around to everyone sooner or later. Venture companies developing for the next generation may feel no impact, except in tenser relations with their funders.

The key issue is the end-user markets. Clearly PCs are slow. Big surprise. But the data suggests that wireless telephony-both handsets and infrastructure-and the Internet infrastructure buildout are slowing as well, for rather different reasons. And except for Sony, which can't seem to build enough Playstation 2s and electronic puppies, consumer is in the doldrums.

This situation could persist until somebody offers end users compelling new capabilities, not just evolutionary new products. Early data on 3G wireless, Bluetooth and digital cameras are not encouraging for a jump-start. We could be in this for a while.