It's funny, I did not see anyone mention that Intel is basically forced to build advanced FABs in the U.S. because they are restricted due to US export control regulations. Just a few years ago (less than 3yrs ago), no technology smaller than 65nm could be sent outside the US by a US company. Companies based in other countries of course could develop the technology on their own (i.e. Samsung, TSMC, etc.).
Thanks for your article. I think Intel management has proving why they are leaders in their market. The selection of Hillsboro, Ore. is a good location for a R&D center. Why would Intel want to build in California when the cost of living is higher and that state has one of the highest home foreclosure rates in the Country? Also, with advance technology, you donít have to be physically in California to be a key player in that State. Intel is during the right thing. When you are a leader, you set standards, you donít follow the crowd. What a wise decision by Intel.
True, that's a good point. I was just thinking, even one 450 mm line might cost a lot, just simple-mindedly going by proportion to footprint.
Which node was Intel planning 450 mm for? I don't think 15 nm or 11 nm. I don't think anything conventional beyond 11 nm as we would expect from Si wafer processing.
The upgrade of the production fabs would include new tools for many lines in the fabs where the R&D fab would only be one set of equipment for a pilot line to start building the new process. I can see how the upgrades will be more money than a single 450mm line.
The Oregonian newspaper reports that 800+ researchers and technicians will be hired between Oregon and Arizona with the bulk of them being in Oregon. I don't know about Arizona but the technicians hired in Oregon are predominately non benefit temps that are only allowed to work for a year before being subject to layoff for at least 6 months. Not very steady employment.
I will hope to see more "manufacutring" jobs created in US no matter in CA, OR or AZ. $8B is a big investment, there are quite a few of places in the world will pay $8B to have Intel build a fab in their countries and tax-free in Paul's lifetime. I will hope US politians can start thinking how to keep manuafacturing jobs in US, whatever it takes to keep jobs in USA, not to push "profitable" companies going abroad by taxing them. Give Intel's Tax back so they can create more jobs in US. Give companies incentives to stay in US, not penalty for going abroad. Intel happened to dominate 90% CPU market w/ technology lead so they can afford such uncompetitive US manufacturing environment. How about the smart phone-iphone we use today? I don't recall much of the content made in USA. One of these days smart phones start replacing PC, what kind of jobs will be left in USA? My best wishes for Intel. Also wish Apple/Google can create some real manufacturing jobs in USA. They do earn majority money from US consumers and their profit is really good, much better than Intel.
If D1X is to start up in 2013, that is too late for 15 nm (already in production). Even 11 nm (2015) would be cutting it close. It would have to be beyond 11 nm, which is most likely new transistor architectures.
Blog Doing Math in FPGAs Tom Burke 15 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...