This year, the IC equipment business is expected to hit $46.3 billion in terms of sales, down 5 percent over 2010, according to the firm. In 2010, the business hit $48.8 billion, up a whopping 99.7 percent over 2009, according to the firm.
On the bright side, capital spending is on the rise. Just in the last week, GlobalFoundries doubled its capital spending in 2011. Intel recently said that its capital spending budget would hit $9 billion, plus or minus $300 million, in 2011, implying 73 percent year-over-year growth.
''With Intel now on the offensive, and including recent capex updates from Samsung and GlobalFoundries, our 2011 capex model moves to plus 12 percent year-over-year verses 5 percent’’ in the original forecast, said C.J. Muse, an analyst with Barclays Capital, in a report.
Still, despite an uptick in spending, the tool supply chain is under duress. Not long ago, chip makers had a choice between five or more vendors in each fab tool segment, said GlobalFoundries’ Armour. Today, ''there are not enough choices,’’ he said. ''Usually, it’s a two-horse race. Maybe three.’’
If that wasn’t enough, Armour is also concerned about soaring tool costs and an ''under-investment in R&D.’’
During a panel discussion at ISS, Intel’s Bruck appeared to be less concerned about the shrinking supply base. ''You can get to the duo-poly’’ in terms of the IC fab tool vendor base, he said.
In a report, VLSI's Hutcheson elaborated on Bruck’s so-called C3 Rule. ''VLSI’s data shows that revenue, R&D, and SG&A per platform have risen three orders of magnitude over the last three decades. What cost tens of thousands to do in 1981 now costs tens of millions. Lithography has risen four orders of magnitude over the last three decades,’’ Hutcheson wrote.
''Consolidation in the semiconductor industry would mean less duplicative SG&A for the equipment industry as well. Hence, tools would be less expensive and semiconductor capital efficiency would be greater,’’ he wrote.
''If the semiconductor equipment industry had not consolidated, R&D would between 50 percent and 100 percent of industry revenues. In order to sustain such an R&D load would mean the equipment industry would have to be sized similar to today’s total IC sales. But that’s not the half of it. Without consolidation, SG&A would approach 10,000 percent, taking the sustainable revenue level to more than $30T, which is more than an order of magnitude larger than the entire electronics industry! In other words, Moore’s Law would have been broken long ago,’’ he said.
Still, there are some alarming trends, especially in lithography, which is turning into a one-horse race on the optical side. In 2009, according to Gartner Inc., ASML held 51 percent of the lithography market, followed by Nikon (39 percent) and Canon (9 percent). Last year, Barclays Capital estimated ASML's current share of the leading-edge, 193-nm immersion market at a whopping 80 percent, leaving Nikon and Canon in the dust.
Canon has already exited the leading-edge lithography business. Nikon, which once filled 100 percent of Intel’s leading-edge lithography requirements, took a body blow from ASML last year when the Netherlands-based vendor won half of Intel's leading-edge lithography business at the 22-nm node. Some believe that ASML could grab Intel’s entire leading-edge lithography business-at the expense of Nikon.
Nikon is falling further and further behind, as it struggles to ship its 193-nm immersion tools. As a result, ASML is commanding a premium for its scanners, partly due to the lack of competition from Nikon.
For the last year or so, ASML has also experienced trouble in terms of keeping up with customer demand, causing some chip vendors-namely second-tier memory makers-from getting their tools on a timely basis.
Clearly, this is not a healthy scenario for chip makers. Simply put, Nikon needs to play catch-up or must join forces with another vendor. One possible idea is a joint lithography venture between Nikon and Applied Materials Inc. Another possibility is that Nikon will need a bail-out from the Japanese government.
Nikon will clearly need help if or when chip makers move towards EUV. Right now, ASML is grabbing most-if not-all of the EUV scanner business. If Nikon can’t deliver EUV, ASML will become the sole source in the arena-a nightmarish thought for chip makers. A production-worthy EUV tool from ASML could soon cost $125 million per unit. And without competition, tool costs could go even higher.
One of the major reasons that nanoimprint and maskless lithography are behind EUV is that early in the development of next generation lithography tools Intel and a few other large companies made sure that most of the resources were funneled to EUV. Nanoimprint and maskless technologies were ignored and starved. Over the last several years there has been so much money and effort thrown at EUV tool and process development that it has become "too big to fail".
Earlier this month DNP announced the purchase of a semiconductor 6025 mask replication tool from Molecular Imprints that uses nanoimprint lithography, making good on the article you wrote in July 2009 (http://www.eetimes.com/electronics-news/4083666/DNP-MII-devise-nano-imprint-mask-technology). They say NIL is progressing to pilot production for semi. Why not follow up and get the latest news from these folks?
Intel is responsible for fabrication equipment industry consolidation to maintain their own process, fabrication, microprocessor and intra platform computing monopolies. In an environment where Intel has destroyed competitors and concentrated their own dealing cartel by racing process destructively. That is at a pace in excess of product organic market efficiencies for nearly two decades. Honest, Intel has never supported the expansion of subordinate economic potentials other than their own. When Moore’s law is an axiom misrepresented to conceal Rock’s enterprise monopoly objective. Where Intel plans in advance the concentration of compliment’s into their Dark hole. In an environment where Intel leads too productize subsequent process regimes, only to move so rapidly to the next, that the prior is prevented organic commercialization. Promoting the very inefficiencies that limit economic profit due fabrication equipment and material design manufacturer’s for reinvestment into a sustainable development practice. Really, today, why doesn’t an Intel microscope kit for young adults come equipped with a barrel etcher and 20 2 inch wafers to fabricate a radio, media player and memory stick? That answer is fabrication process regime never freed from Intel monopoly restraints. And now, at 450 millimeter, under Intel control invites the catalyst for a destructive accelerant. By an executive team that cannot demonstrate management antitrust compliance to free industry from the many form of Intel industrial slave society. Where everyone knows Intel’s objective is to bar others from crossing a very narrow bridge into the new world of molecular electronics. Mike Bruzzone, Camp Marketing Consultancy.
Nano-imprint and maskless are behind EUV. We have covered these topics extensively. Nano-imprint is not ready for prime time in semis. Not sure it will ever work for semis. Maskless or ML2 is still science fiction.
Given the challenges, uncertainty and expense still to come for EUV, I do not understand why other next-gen technologies are not mentioned in such a discussion. What is the status of nanoimprint lithography and maskless writing?
Simple economics says $125M/tool is not a plausible solution.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.