It isn't often one gets to use a climatic metaphor in this business. But here it is January. The days are actually-but imperceptibly-growing longer. The shortest day of the winter has passed. But the nasty weather is just starting.

Economically, as well, it appears that the sharp contractions that tore the industry in the wake of the Internet bubble's rupture have passed, and a period of solid but nearly imperceptible growth has set in. Underlying indicators say that the industrial economy is in recovery. But you sure couldn't tell it from Silicon Valley, where some streets have no occupied office space at all, and everyone knows several good people who are still unemployed.

For companies that are holding on, cutting back staff and waiting for the good times to roll again, it's the beginning of a long, hard winter. Staff cuts will continue, business will grow slowly and spring, when it arrives, may be quite different than the one old-line companies were expecting.

The chip business has run to the mantra of "denser, faster, more integrated." The challenges have always been about achieving the next level of performance or integration or energy efficiency. How to pay for the design, or where to find a market, have never been issues.

But they sure are now. When we talk to design teams about their problems, be they startups or teams within large companies, we hear about lack of staff, lack of budget, inability to write a check for tapeout or inability to find a customer who can write a check for anything. We can do denser, faster and more integrated-we just can't pay for it.

This is creating a whole new kind of information gap in the engineering community. While conferences, publications and promotional activities continue to focus on the mantra-tool flows and foundries for 130 nm, new challenges for gigahertz design and systems-on-chip with 10 million gates-engineers are asking entirely different questions. How do I finish this design with half the staff? How do I cut the up-front design cost to zero in case we never sell this generation? How do I afford a tapeout when I only have $50k available to go to early production?

In this cold season, survivors will shift their attention to multichip solutions in older processes, to more extensive use of programmable logic and other off-the-shelf components, to outsourcing and creative financing of design budgets. It's a long time till summer.