Sources believe that Japan's Nikon was supposed to ship an alpha or R&D EUV tool to Intel by the end of last year. Then, Nikon reportedly delayed the shipment, sources said. Instead, the company supposedly decided to keep the R&D tool at its headquarters in Tokyo, reportedly because it would take too long to ship the unit, sources said. That machine would still be used for R&D, they added.

However, Nikon denied that rumor and is apparently still looking to the ship the EUV tool to Intel in the future. EUV is a next-generation lithography machine that is aimed for the post-optical era. But EUV, in general, has been delayed by an assortment of problems, namely the mask, power source, resists, among others.

Analysts are under the impression that Nikon will still ship the alpha machine to Intel. "Regarding Intel, Nikon plans to ship a unit next fiscal year, but the shipping process has not started yet," said David Motozo Rubenstein, an analyst with Jefferies Japan Ltd. in Tokyo.

On the bright side, Nikon will shortly ship its first EUV alpha tool, reportedly to Selete, a Japanese chip R&D consortium. The second EUV alpha tool was (and still is) targeted for Intel, the biggest proponent and backer for the technology.

"They are shipping one unit this year to a Japanese customer," Rubenstein said. Sources said the first customer is Selete.

Nikon "cannot book revenues until the whole unit is installed, and it is indeed massive," he said. "It takes almost one year to get it up and running. They expect it to be ready by March, 2008. However, this is, of course, only for R&D use. They don't expect EUV to be used for production until after 2010 because of low throughput, costs, etc."

Officials from Nikon, however, insist the shipment is not late to Intel, said Scott Foster, an analyst with HSBC. In an e-mail, Foster received this response from Nikon regarding the reported tardiness of the alpha EUV tool: "Not true. We will ship the EUV tool as planned. Regarding the size of the tool, it is actually not that big," said Andrew Hazelton of the marketing department at Nikon.

Intel declined to comment. "You know we can't comment on rumors," according to a spokeswoman for Intel.

Troubles for EUV

In any event, EUV is in trouble. The technology was originally slated for chip production at the 45-nm node, but now, EUV is seen as a possible candidate for the latter stages of the 22-nm node in the 2011 timeframe. Some do not believe the technology will ever work in production fabs. It is also too expensive, as a production tool is expected to run $60 million each or more. Some say a tool will run $100 million.

Intel has been the biggest proponent in EUV, ostensibly as a means to extend Moore's Law. The chip giant helped get EUV off the ground, and, in 2004, it was said to have installed an EUV micro-exposure tool from U.K.-based Exitech Ltd. for R&D. The chip giant was also supposed to obtain an R&D tool from Nikon by last year.

Lately, the NAND flash crowd--such Samsung, Toshiba and others--are interested in EUV technology for good reason: leading-edge 193-nm immersion tools are expected to hit the wall at the 32-nm node or so. NAND vendors are pushing the limits of lithography, but the tool suppliers are not staying up with Moore's Law.

Beyond 32-nm, there are a number of lithographic options, including 193-nm immersion with double patterning, EUV, multi-beam e-beam and nano-imprint. Multi-beam e-beam and nano-imprint are not ready for prime time, while EUV appears to have a long ways to go.