I have been saying this for a long time.Now that Rocket Lake has been released, looking back at our pages and pages of discussion, leaks, rumors, and now verified tests Intel's decisions are looking just a little clearer to me.
I think it really comes down to production. Intel has billions and billions of dollars of fabs and they can't let them sit idle. It's easy to say, "Why didn't they just release Tiger Lake 8 core 10SF instead of Rocket Lake?" They simply didn't have the production capability for 10SF to make this happen now. They have orders to fill and they only way they were going to be able to fill them is by utilizing their 14nm fabrication process. I assume once they produce enough Rocket Lake parts to fulfill order to Dell, HP, Acer, etc.. and enough stock for retailers they will begin to transition their fabs to 10SF.
Thing is by the time they are in full swing with 10SF/ESF, AMD will be full swing with 5nm.
I'm wondering how Intel 10ESF will compare to TMSC 5nm in terms of transistor density and power? Will the gap be larger than 14nm vs. TMSC 7nm which exists currently?
Intel’s bigger issue is going to be AMD’s 5nm refresh. Intel should really push hard to deliver 7nm earlier than planned. I doubt they will though. Their only real saving grace is that AMD is not expected to deliver client parts on 6nm or 5nm before late Q1/early Q2 of next year.