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!
Pat Gelsing will have a very limited amount of time to show his ability to turn Intel around. The street now gives Intel zero credibility, and this could be dangerous for a company that requires large amounts of capital and talent to compete. The talent aspect will become key if Intel starts to become conceived as a sinking ship. At this point, Gelsinger has to hit a home run or even two if he wants to avoid Intel becoming a has been. I hope he pulls a rabbit out of the hat, for the more successful companies there are in this area, the more motivated all semi companies will be. Culture is everything and it looks like Gelsing will have to start from scratch. I feel he needs a partner relationship, like Apple and TSM, which is one of the greatest business/tech relationships in history. Any thoughts or comments appreciated.
Pat Gelsing will have a very limited amount of time to show his ability to turn Intel around. The street now gives Intel zero credibility, and this could be dangerous for a company that requires large amounts of capital and talent to compete. The talent aspect will become key if Intel starts to become conceived as a sinking ship. At this point, Gelsinger has to hit a home run or even two if he wants to avoid Intel becoming a has been. I hope he pulls a rabbit out of the hat, for the more successful companies there are in this area, the more motivated all semi companies will be. Culture is everything and it looks like Gelsing will have to start from scratch. I feel he needs a partner relationship, like Apple and TSM, which is one of the greatest business/tech relationships in history. Any thoughts or comments appreciated.
Does intel design not do everything for IFS that Apple does for TSMC and more? Based on what Dan has said about Qualcomm helping develop the PDKs for 18A and intel mentioning that they are using direct print/flexible design rules for 18A per IFS customer request, it would seem they also have an external partner that is in their inner circle.
It is not just the commercial relationship, it is the learning and balance in evaluating decisions and making future plans which comes from having multiple knowledgeable external partners. TSMC benefited greatly from needing to listen to diverse customers, rather than the echo chamber of a tied in-house customer. That probably was key to UMC survival and prosperity too, and has been healthy (though late coming) for GF.
There is " 10000 hour rule" to become master. In semiconductor manufacturing, there is also a rule of ~100k(?) wafer output learning curve to get stable HVM. The output wafers could be company dependent. IMO, to speed up, intel needs high intensity in wafer manufacturing. Will the market, investment, talent and WFE support it? Let's see.
It is not just the commercial relationship, it is the learning and balance in evaluating decisions and making future plans which comes from having multiple knowledgeable external partners. TSMC benefited greatly from needing to listen to diverse customers, rather than the echo chamber of a tied in-house customer. That probably was key to UMC survival and prosperity too, and has been healthy (though late coming) for GF.
Without a doubt having a more diverse product portfolio will help intel fabs be better and more efficient. Intel design increasing external foundry usage can also allow design and fabs to better understand the areas in which they are deficient.
Pat's a very good guy but the reality is he has <10% chance of pulling Intel thru this and back to to its previous lofty position. He'd need another 10 years to change the culture & weed out the resistance but doesn't have that.
They'll absolutely still be around in 10 years but not as we know them. It's sad but in fairness, its been coming.
Personally I would like to see Intel design AND manufacture chips for others. That is how Apple started with Samsung. That is how many fabless companies started (using the ASIC business model). The number of systems companies that are looking at chip design is staggering. Intel needs to look for the next Apple.
Personally I would like to see Intel design AND manufacture chips for others. That is how Apple started with Samsung. That is how many fabless companies started (using the ASIC business model). The number of systems companies that are looking at chip design is staggering. Intel needs to look for the next Apple.
Not mentioned is how Samsung burned Apple so bad Steve Jobs commented on his death bed to spend every cent suing Samsung all over the world. His death occurred on the day when hundreds of employees of TSM where coordinating with Apple employees to speed their collaboration. He insisted on the collaboration going forward as speed despite his condition.
Not mentioned is how Samsung burned Apple so bad Steve Jobs commented on his death bed to spend every cent suing Samsung all over the world. His death occurred on the day when hundreds of employees of TSM where coordinating with Apple employees to speed their collaboration. He insisted on the collaboration going forward as speed despite his condition.
Not to be that guy, but they went back 2 years latter to split production with Samsung due to TSMC having a slow ramp on 16FF. Obviously emotions play a part in Apple exclusively using TSMC toady, however at the end of the day these are businesses. TSMC simply is the logical choice for Apple given the consistent cadence/roadmap, preferred treatment, best in class foundry services, and at least for the past few years unquestioned process leadership (or at the very least tying Samsung but coming out 1+ years sooner). If the conditions change and TSMC no loner is the "best choice" the TSMC-Apple duo will end just like the Samsung 28LPP and 14LPP adventures ended. Of course I don't see any events shy of a travesty of epic proportions causing TSMC to stop becoming the best choice anytime soon. But I did want to bring it up that all rock solid relationships can eventually be ended if there is no longer a profit motive behind them; after all nobody expected the truly excellent Apple-Samsung duo to ever end. Without the cheap components/collaboration with Samsung many of Apple's most innovative products might have never left the lab (check out Dan's A Detailed History of Samsung Semiconductor, it's a great read). Alas they (and the rest of the CPA) blew 28nm and Samsung destroyed their reputation ending that duo.
