Reading between the lines on this one it looks lik there is going to a hiatus in Moore's Law while Intel, ASML and others reapply themselves to this.
It also looks like the transition to 450-mm may be easier to implement than the one to EUV, which has got Intel thinking.
Moore's law is about doubling the number of transistors AND the new chip still costing the same (I.e. the cost per transistor goes down). What the industry is finding is Moore's Law is breaking down due to lithography cost (tools used to print the chips way too expensive). Next generation EUV steppers cost ~$150M and can only process 10 wafers per hour versus expected 160.
Intel needs to move its business model off Moore's law. Intel is spending more and more on a new silicon technology that is delivering less and less.
Only ASML has the edge technologies in lithography. ASML is still investing a lot in human and finacial resources in EUV. If Intel, Samsung and TSMC want 450mm fab, they have to involve .... push ASML to do it.
450mm was already considered close to hell freezing over, and it's already ahead of EUV. This might be a sign of realization by the industry. And the earliest production date has slipped to 2014, another sign of realization.
Uncertainty cannot help lower cost. Delay is the greatest cost of all. It may be time to really look at other options. So what happened to these other options like DSA, imprint, maskless, etc.?
Intel must double-down on 450mm because growing millimeters of silicon diameter is cheaper and far more predictable than reducing electromagnetic wavelengths to X-ray range. Presumably, ASML can build a 450mm stage (other, much smaller companies are) with greater success than the SXPL machine they are trying to build. Meanwhile, they sell multiple profitable immersion tools for optical multi patterning.
I can smell the desperation of intel, while competitors is chasing closer from everywhere, there is only one direction, up. and there is only sky there and it don't have a wing...
they better invest in CERN as well, they could offer quark level tech for them.
I don`t see samsung joining, simply bcoz IBM, GLOFO and the rest will not be happy and its not certain ASML will produce the results, if ASML fail, thIeR shares will tank and the money invested will be lost..
I think TSMC will join (best case scenario)
Moore's law is not so much about getting more transistors on a wafer by making them smaller, it's basically about reducing the cost per transistor.
Making transistors smaller using EUV is at the moment too expensive so reducing the cost per transistor by moving to 450nm is the logic way to go and meet Moore's law.
The main issue is that at the 300-mm transition the chipmakers pressured the equipment makers to take on all the risk saying that the first companies to offer 300-mm versions in various equipment categories would "clean up" and get lets of orders.
The the 300-mm transition stalled (I forget why, but i think there was some economic crash or others) and was long and drawn out. The second movers were able to catch up and a lot of equipment makers got burned.
This time around we are down to close to monopoly suppliers...ASML, Applied Materials ...and they are starting to tell customers.....you want this work doing...you got to foot the bill...up front....and sign some purchase orders.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...