https://seekingalpha.com/article/40 49095-truth-behind-intel

File: 1500607252344.jpg (146KB, 1170x707px) Image search: [Google]

146KB, 1170x707px

This too

File: nm process details.png (31KB, 707x797px) Image search: [Google]

31KB, 707x797px

>>61596302
They relied so hard on their fabs for everything, then it all went horribly wrong.

>>61596794

That's funny. in the middle of 2018 Samsung MAY release a 7nm process using EUV that matches the 10nm Intel released in Q1-17

>>61596876
Nice alternate reality, bro.

>>61596794
Damn, looks like GF and TSMC 7nm are almost interchangeable.

>>61596914
Until you find out TSMC's process is a low power one while Glofo's one is a high frequency(5ghz) one

>>61596926
Right, but in terms of physical characteristics there's very little difference, even by the standards of modern semiconductor standards.

>>61597029
True, but all of them are rather close in BEOL metrics.

The outlier seems to be Intel's mask count, which seems absurd.

>>61597055
I don't know how did they managed to have MORE masks with lower density node.

>>61597055
That's the reason yields are so shit and 10nm has been delayed for over a year.

I think Intel wants to put this node behind it and get working on 7nm, nothing good comes out of pouring money into a disaster.

>>61596302
The worst thing is, Intel's 10nm still hasn't ramped up to mass production levels, and probably won't until mid 2018. That means 10nm would have been delayed by a whopping 3 years off schedule.

>>61597089
Their 7nm node is III-V FinFET-based?

>>61597106
Starting off with peanut sized cannonlake dual cores for mobile should tell you everything, those things won't be over 90mm^2 and the yields are still fucking awful

>>61597140
I don't think Intel has detailed anything about their next gen node.
That's in like 2020 or something baring no delays? Who actually knows.

>>61597148
Mask count says everything about their 10nm node.
It's a disaster.

>>61597062
Different fin profile, different materials used, the FEOL is exceedingly complex even compared to the sophisticated wizardry that is high frequency signaling in the back end.

Intel's 10nm process is a direct evolution of their 14mm, and subsequently their 22nm Trigate process. They've made strides in the characteristics of their fin profile, they have very high consistency hence why they're trigate, but on the material front I don't think there has been much progress. Use of tungsten? If no other foundry is doing the same then it would be a pertinent question as to why intel is.

>>61597163
They haven't spoken much about their future plans because their 7nm facility in Arizona is still under construction, and its a long ways off from being a completed project. Takes a long time to purge the air of all contaminants, and then clean the air system itself before they can even think about handling wafers there.

Don't worry, Intel can bribe their way back to the top.

>>61598419
Not this time, swetee.
Qualcomm have already shown them their place near the loo.
AMD will eventually do the same.