rzb, This would be a very bad idea for TSMC. Funny thing is you have no idea why. Is it because of the redundancy inherent in FPGA designs and therefore the ability to "fuse out" or otherwise mitigate defects?
No ... That would be an advantage for Xilinx to build in redundancy to their FPGA designs. I was talking about advantages for TSMC to have an earlier adopter being an FPGA.design.
While I generally would agree with your comment, one might note that Intel and AMD/GF have generally had CPUs as the lead product on each new node for quite some time. Way back, Intel might have done SRAMs (not for test, actual product), but that has been WAY back. So if Intel and AMD have done it that way, why can't a foundry? Its really a matter of what customers are willing to pay in the early ramp. If SOC vendors are more eager than FPGA vendors, well, why would TSMC care?
Agreed. If there were only do overs. The only hard lesson I've learned if I can pass it on is to never buy a stock until you tell yourself a price when to sell if it drops. I broke my own rule recently and it cost.
I started watching the tape lately, which way the markets themselves are moving.
I wish I could find it again, I agree with an article I read recently. Basically a lot of money is sitting on the sidelines after the recent 2008/2009 market dump. Bernanke has said interest rates will remain at zero for years to come, the result being people's greed will kick in and the money they see sitting in the bank doing nothing will start coming back into the markets. I think that's what you were saying with the markets racing to new highs, I think they'll keep going...
Then again, I can write about why you shouldn't be in the markets here at all... :)