Intel Foundry’s Post

View organization page for Intel Foundry, graphic

26,034 followers

Thank you, Forbes and TIRIAS Research, for this in-depth look at how the Intel 3 node is a significant step in our five nodes in four years (5N4Y) journey! Intel 3 delivers notable enhancements over the previous Intel 4 node, including an up to 18% improvement in performance. Learn more below. #Intel #IntelFoundry #Foundry

Jeff Morrison

Financial Cultural Operational and Technical Consultant - Alpha Sense Financial Consulting

3w

When do you expect to reach parity in terms of yields. Ireland is reporting Intel 4 and 3 aren't ready for primetime. Intel 4 is at 560 ISO and 3 is at 530. TSMC is at an average weekly ISO of 670. For a 12mm x 12mm die this translates to 120% more good die on 3nm. The process isn't well characterized and the TSMC folks say your at a local maximum They've comprehensively examined the response surface and feel they are near a global maximum and have better control of the E-test parameter and each layers critical dimensions have 6 sigma control. We reckon that you're at 4.2 for the 3nm, and haven't optimized the target CD's especially on the critical gate layer and ultra fine pitch metal layers. Also registration between layers with stacked tolerance is a big limiting factor to better yields What's the recovery plan to get economicallly viable utilization and yields. With equipment utilization at sub 50% and yields less than 45% your true over all efficiency is less than 25%. TSMC has 90% utilization and approximately 85% yields they are realizing roughly 80% productivity in their 3nm factories. Moreover your transistor density on 3nm is 25 to 30% lower and consumes 2.2x the power/transistor.

Like
Reply
Sharisse Wong

Partner Collaboration | Business Growth & Transformation | Accounts & Programs Management | Strategy & Execution | Supply Chain Management

2w

Well said!

Like
Reply
See more comments

To view or add a comment, sign in

Explore topics