But when the debate in Dallas turned to lithography, much of it was about costs, including mask costs.

Peter Silverman, Intel's director of lithography, started the event off by declaring that extreme ultraviolet scanners must hit cost targets that have become much more stringent due to the changed spending landscape.

The complex engineering challenges facing the EUV development effort are worth it only if they keep the semiconductor makers on track to reduce the cost per function by the historical 25 percent per annum.

The key argument is that EUV is extendable over several process generations, spanning the next decade. You can argue that EUV might not be needed until 2010, but at some point EUV will be inserted. E-beam projection lithography and maskless lithography won't have the throughput; 157-nanometer scanners are limited by the wavelength. And the 157-nm tools won't be cheap, either. The early-use beta 157-nm scanners cost almost $30 million each, though that may decline as volumes ramp.

Silverman said the lens makers might not have access to enough of the high-quality calcium fluoride material needed to make production quantities of the 157-nm scanners.

"The issue for 157-nm lithography is purely a calcium fluoride issue: High-quality calcium fluoride is more expensive than gold," Silverman said. "The yield for 157-grade calcium fluoride is quite, quite low. Several new factories are needed, and while some investments are being made, nothing big enough is being built to supply the industry. So it is an open question whether 157-nm systems will be available in volumes, in the time frame that industry needs them. It also depends on whether EUV will be ready in time."

I equate the EUV effort with something from my teens: the effort to put a man on the moon. The difference is that Uncle Sam paid for the Apollo effort.

For small companies in particular, finding a form of affordable lithography is a life-or-death issue. As consultant Ken Rygler noted, some startups could be driven out of business if they fail to get first-time working mask sets.

Some of those companies will have to add value in ways other than shrinking. Even for Intel, the message from Dallas was that EUV must be affordable.

Route feedback to dlammers@cmp.com.