Gargini gave a keynote speech here at Semicon West, saying "we think 2012 would be the year for the 450-mm wafers, so we need to start discussing it pretty soon, in the next six months. We need to start planning."

An ITRS roadmap meeting, being held here starting Wednesday (July 14), will consider updates to the 2003 edition of the ITRS. That version of the roadmap included the 450-mm wafer size for 2012, but little public attention has been directed at the issue until now. At Semicon West, several participants said while the larger wafers may be needed by Intel and other companies with big die sizes and large volumes, the larger part of the chip industry will have little enthusiasm for the 450-mm wafers.

According to the roadmap, the 300-mm wafer generation kicked off in 2001. About 14 percent of total wafer capacity this year will be served by 300-mm wafers, though most of the new equipment being developed now is targeted at the larger platters, according to Applied Materials Inc.

The last wafer transition was rocky, with a serious downturn affecting the need for the 300-mm wafers. For the 450-mm generation, "we need to work even more closely together to have a clear understanding" of the requirements of the various parts of the industry, he said.

It takes as many as five years to prepare the standards and prototypes, before final equipment design can begin, he added.

Not everyone is convinced that the 450-mm generation is needed. Already, a single 300-mm wafer can hold as many as 10,000 reasonably sized die, and larger wafers would make it even more challenging for foundries that need to supply small numbers of chips to customers at the most advanced design rules.

Also, a single 300-mm wafer fab can cost more than many semiconductor companies can afford. By one estimate, a company needs to have about $5 billion in annual revenues in order to justify the expense of building and operating a 300-mm wafer fab.