Electronics Weekly Electronics Design & Components Tech News
TSMC says it intends to start trial production on 450mm wafers in 2013-14 and to be in volume production in 2015-16.
Intel says it will be getting first generation 450mm tools in 2015 and will start making ICs on it. In 2016/2017 second generation 2 tools will be delivering real productivity and Intel will then start upgrading its existing fabs.”
But, before this week’s announcement of the Intel investment in ASML, ASML’s position on 450mm tool development was that 450mm manufacturing will not be introduced until 2018 “or thereabouts”.
ASML’s 2018 is rather significantly later than TSMC’s 2013-14 and Intel’s 2015-16.
Not for the first time, there is a significant miss-match between the aspirations of the chip-makers and the timetables of the tools-makers.
As Samsung and TSMC ponder on whether to join in the investment in ASML, or throw their weight behind alternative lithographic tool-maker Nikon, the world can see the unfolding of the Intel master-plan for world domination.
Always the most pushy proponent for 450mm, Intel can see itself becoming undisputed leader in process technology – both for wafer size and feature size.
If it can achieve that, then the most powerful electronics manufacturers in the world will have no choice but to go to Intel for foundry, Intel’s microprocessors will be the only option for every leading performance product from smartphones to servers and Intel will transform from being dominant chip-maker to owning the industry.
Is that a good thing?
I don’t think so.
At Semicon West Conference last week its apparent Intel needs 30% cost savings of 450 mm wafers on 35% of its microprocessor production priced below cost.
The question is will Intel continue to produce 30% more product than real time demand? Which since Intel time began is the essential element of strategy used to bank the exponential capital requirement of Rock’s Law to pay for continuing along Moore’s Axiom.
I personally believe Intel will harvest 193i learning at the conference, in a lithography panel including Intel personnel, that directed self assembly is the evolutionary patch Intel can employ to continue 193i toward 14 nm.
Traditional scaling ended long ago. Now operating in the realm of applied science will Intel wait for the next disruptive lithography innovation to sort out before investing in the new commercial industrial art?
Intel needs 450 more than they need EUV. In the Semicon Litho panel EBEAM was positioned as the compliment technology to 193i and DSA.
What Intel may be saying with its ASML investment, to IBM, Global Foundries, Samsung, TSMC, is pony up with us for lithography. And if not 450 mm wafer handling systems first and foremost integrated into our current fabrication line will be used against you.
Mike Bruzzone
Camp Marketing
I’m told Nikon hasn’t got a hope of catching up in 450mm, Martijn, but who knows what a collaboration with Canon and a few billion dollars from the industry and the Japanese government could achieve?
that is assuming 450 brings indeed significant cost saving AND will be inserted in the leading edge node.
I expect the first 450mm tools will not meet the same performance specs as high end 300mm tools at that time.
Could we see a split between N/N-1 nodes in cheap HVM on 450, and leading edge N+1 node on 300 mm until 450 technology catches up?
So far haven’t heard any comment from Nikon on 450mm. Even without the Intel boost, it was questionable if Nikon could keep up with ASML in new research (see EUV), what will be their strategy now? Only make dry 450mm tools?