You are talking nonsense. GF 7nm does not exist and would not have existed if AMD didn’t choose TSMC. TSMC saved AMDs bacon big time.
It did exist, was ready (or less than three months away) for tape-out phase and main reason it was cancelled was TSMC.
Just like Intel's 10nm production did exist as early as 2018. There you go
https://ark.intel.com/content/www/u...38121u-processor-4m-cache-up-to-3-20-ghz.html
It took around 3.5 years after that to launch first 10nm desktop CPU's but hey, it did exist back then...
Cool story, bro... Source? AFAIK even GloFo admitted back in 2018 that demands will get higher as they put it in their press release.
“Demand for semiconductors has never been higher, and clients are asking us to play an ever-increasing role in enabling tomorrow’s technology innovations,” Caulfield said.
To what I gather, GloFo stop developing their own 7nm due to lack of financial incentives to do so. high profit yet high entry, so they decided to better support their 12/14nm offering instead.
Lack of demand likely resulted from TSMC's plans to make much bigger 7nm output which they did. At that point Intel's 10nm could produce ~3 GHz dual core chip, so basically only GF's real competition was TSMC. This GF agreed as 7LP dimensions were deliberately made very close TSMC's 7nm process dimensions.
There we basically have it. Because TSMC could produce so many wafers GF feared there are not enough customers for their 7nm and future 3nm node and/or GF's 7nm capacity was too small to be profitable. IF TSMC would have less 7nm production, demand for GF 7nm would have been higher and there would have been less fear about overcapacity.
Also, accordinf to Ian Cutress of Anandtech, AMD has always been designing Zen 2 using TSMC 7nm CLNFF from the start, so the argument of using GloFo are totally 100% what ifs scenario. as AMD never bet on using GloFo 7nm even when they're in development phase.
www.anandtech.com
Not so much:
Keep in mind that AMD’s first 7 nm product was designed for TSMC’s CLN7FF from the beginning, so the company did not bet on GF’s 7LP in late 2018 anyway, and no rush with the manufacturing technology was needed for GF’s key customer.
First, that does not say anything from Zen2. Secondly, AMD's CEO admitted they are planning to use both GF and TSMC for 7nm:
https://www.anandtech.com/show/1231...n-exclusive-interview-with-dr-lisa-su-amd-ceo
LS: So in 7nm, we will use both TSMC and GlobalFoundries. We are working closely with both foundry partners, and will have different product lines for each. I am very confident that the process technology will be stable and capable for what we’re trying to do.
Third, WSA did not allow AMD to use other foundries than GlobalFoundries for 7nm, but GF did gave AMD permission to use TSMC because they did not have capacity for everything AMD needed at that time. Also first GF 7nm tape out was to be AMD processor. Source for both above and also making 7nm process dimensions close to TSMC process there:
https://www.eetimes.com/euv-in-final-push-into-fabs/?page_number=2
From those it's very easy to say AMD was going to make high clock, low core CPU's (=Ryzen) at GF and low clock, high core CPU's (=Epyc chiplets) at TSMC. As GF dropped 7nm, AMD just used backup plan: x570 chipset as IO die and TSMC chiplets as compute die.
