You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
How much in the US and how much in Europe? Japan is getting fabs fast so Asia is getting even stronger. I don't see half coming from the US and EU. Are we only counting EUV fabs? Or all FinFET fabs?
Like me, he doesnt count 12nm or 14nm as advanced... N, N-1, N-2 is advanced IMO. N-2 might be 7nm today.... it wont be next year. it definitely aint 14nm, 16nm, etc. 128Layer NAND isn't leading edge. 1Y DRAM isnt leading edge, 2019 tech isnt leading edge. just an opinion.
50% is a goal. There are a ton of milestones to hit before we get there. Micron had a great comment on building in the US vs Asia last week.
TSMC leadership is but one Intel 10nm like mistake away from losing it all, 3nm was close. I’d say almost line they are on Intel’s 14nm and than 10nm trajectory
TSMC leadership is but one Intel 10nm like mistake away from losing it all, 3nm was close. I’d say almost line they are on Intel’s 14nm and than 10nm trajectory
I agree that I don’t yet see an intel level face-plant in N2. But his point was that N3 feels reminiscent of i14nm. Also i10nm wasn’t 5 years late. It was 2-3 years late depending on if you want to go off their original promise of 2017, or if you say it should have come out in 2016 (2 years after 14nm).
IF TSMC did have a 10nm level fiasco, I think momentum will make them do just fine. Sure they will lose marketshare due to the lost confidence and inferior PPA, but I don’t know if face planting only one node is enough of a gap for Samsung or intel to orchestrate a role reversal. I think TSMC would need to royally bungle somewhere between 2 out of 3, or 3 in a row for a real coup to occur.
I do think TSMC is worth watching now. A small slip every now and again is understandable and forgivable (especially when you have the market power of TSMC). But TSMC must ensure any and all semblances of arrogance are purged out of their company. There seem to be some signs of an arrogant culture, but hopefully N3 and Arizona have beaten any such notion out of them. Nobody could have foreseen what happened to intel. Yet the slow acting poison of arrogance and the siphoning off of resources can bring even the largest titans to their knees.
Success breeds complacency, and complacency breeds failure. Only the paranoid survive…
I agree that I don’t yet see an intel level face-plant in N2. But his point was that N3 feels reminiscent of i14nm. Also i10nm wasn’t 5 years late. It was 2-3 years late depending on if you want to go off their original promise of 2017, or if you say it should have come out in 2016 (2 years after 14nm).
IF TSMC did have a 10nm level fiasco, I think momentum will make them do just fine. Sure they will lose marketshare due to the lost confidence and inferior PPA, but I don’t know if face planting only one node is enough of a gap for Samsung or intel to orchestrate a role reversal. I think TSMC would need to royally bungle somewhere between 2 out of 3, or 3 in a row for a real coup to occur.
I do think TSMC is worth watching now. A small slip every now and again is understandable and forgivable (especially when you have the market power of TSMC). But TSMC must ensure any and all semblances of arrogance are purged out of their company. There seem to maybe be signs of arrogance but hopefully N3 and Arizona have beaten that out of them. Nobody could have foreseen what happened to intel. Yet the slow acting poison of arrogance and the siphoning off of resources can bring even the largest titans to their knees.
Success breeds complacency, and complacency breeds failure. Only the paranoid survive…
you think TSMC 3nm is like Intel 14nm-10nm? Seriously? People were concerned on 3nm before they ramped it on one of the world highest volume products from the most demanding customer. Intel started 10nm with technology lead. ended it 2 generations behind....
NOW could TSMC drop the ball like Intel? yes.... but the chances of TSMC failing, Intel and Samsung executing and TSMC falling behind are <2%. And reminder, you need to actually make money at technology. If Intel ramps 18A and the margins are negative for Intel manufacturing division, that might be problematic
