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!
Gelsinger should have been talking about building an ecosystem. TSM has an ecosystem built on iron-clad trust and tight cooperation from design to equipment to actual production. Morris Chang put this in place to where a large group of companies acts as a team to build a world-beating ecosystem that benefits everyone from suppliers to customers. TSM is a finely tuned machine based on trust and respect to achieve ever-higher goals. Gelsinger either has to copy this model or the one Intel was originally built on by Andy Grove. Both are similar but have their differences. The one thing they definitely have in common is success. I hope Gelsinger can pull this off, but I still don't see a plan for an ecosystem and a substantial cultural change that is needed to become a winner once again. This is a process that takes years, not months to establish and build, but taking the first strong steps is key. Ecosystems and culture should have been a central part of his presentation over projects and goals. This will be a battle of ecosystems and Gelsinger must focus on that goal. Any thoughts or additions are welcome. I have placed my substantial bets many years ago and they were on TSM and will continue to be.
I agree as well, which is one reason why they should consider buying GF. Another course would be doing something completely different rather than trying to copy TSMC. The with prices on the rise the foundry business is do for a disruption, absolutely.
Intel needs to have and to own their ip. The tsmc model for Intel will never work because tsmc does not compete with their customers. A samsung-lite model is what I think Intel should try.
I agree as well, which is one reason why they should consider buying GF. Another course would be doing something completely different rather than trying to copy TSMC. The with prices on the rise the foundry business is do for a disruption, absolutely.
What you have to realize is that Intel success as a foundry provider was not good enough in the last 10 years. The foundry was used to tell design teams internally that “this is it” manage your design to fit the process. The word “service” meant different things than in TSMC case. If they want to succeed in this business the entire Foundry has to be an “independent” unit with service to customers as a priority not as a “favour”!
They have to work on advancing down the gate size as their processors and other infrastructure chips depend on their competitiveness. US consider having foundries on their land, time to help them do that. But the ball is in their hands to change attitude. They also have to provide “free” IP so design houses will come to their foundry by having all interfaces covered! Toronto team had a lot of IP built but this is not enough to draw the customers.