I saw Jayz2Cents' video before commenting here.  It seems to have some interesting insights:
 
Insight 1: possible reference design for the power capacitors.  Keyword "possible". According to Jay's video the 3080's reference power capacitor setup is 1 MLCC cap and and 5 POS caps.  nVIDIA goes a little above that with 2 MLCC/4 POS and ASUS goes well above it with all 6 MLCC.  The 3090 is 2 MLCC and 4 POS caps.  Interesting if true.  
 
Insight 2: Of the cards reported to have issues officially so far, they all have 6 POS caps and 0 MLCC caps.  Keyword being "reported" as we cannot tell 1 way or the other yet without more investigation that this is the key cause.
 
Insight 3: If the AIB's are going to fix this for the 3080 according to Jay it would be done with a firmware update that lowers likely the GPU Boost after the Boost clock of the cards to keep the card stable.
 
One last thing to mention that everyone reporting is not answering which is the answer to this very important question: Are the cards that are CTDing (Crashing To Desktop) while gaming OCed beyond the factory overclocked or stock speed specs by the user's own discretion or are they running without any additional overclock aside from what the factory might provide?  Either way, that's kind of scary.  Out!   

Having to update the bios to fix this would truly suck and I would advise anyone in future who has to resort to this option to fix their card return it for a full refund straight away.
 
If a company has to intentionally hamstring a card lowering its performance to get it to work properly then that is unacceptable. These cards are not cheap after all.

Avenger411
RTX20 memory fiasco, RTX30 fiasco.
Rip

970 3.5GB VRAM fiasco, 1080 going BOOM (personally had this one), 20 series RAM issues, 30 series.. well that list is growing still.*
 
FTFY

i hope i can get one soon so i can fell the stuggel

Mr.Gisoid
Actually Hardcore Overclocking
https://www.youtube.com/watch?v=GPFKS8jNNh0

Oh buildzoid's thoughts this should be illuminating. Thanks for posting.

people get what they deserve

It seems Production versions of both FTW SKUs have ONLY 4 POSCAPS. Reviewers received pre-production models with 6. Launch was delayed a bit because of this; EVGA knew about it. Pics have been updated on EVGA website. Their may be be other stuff going on too...

All the 3080 FTW3 Ultras on reddit have 2x MLCC's and 4x SP-CAP. I'm pretty confident that anyone currently waiting to get their hands on a 3080 FTW3 Ultra will receive the 2-4 setup.

People keep posting promotional photos which is a moot point as it's subject to revision and editing . And who knows GN may of had a pre-production card sourced by EVGA I don't know. TBH what there saying makes sense and I too believe it's a capacitor issue, but it's still not confirmed.

For now I'll stay on the list for step-up, if my card needs to be RMA'd so be it. EVGA will stand by its customers, that's a major reason a lot of us purchased EVGA in the first place.

Johnny_Utah
 

Mr.Gisoid
https://forums.evga.com/Message-about-EVGA-GeForce-RTX-3080-POSCAPs-m3095238.aspx

Incredibly informative, thank you.
 

It really was! (if you like that extreme level of detail).    one thing I took away is that basically on the SPCAPS, the higher the number, the more the impedance and essentially the better quality and higher cost.  the next concern that will start to raise is that the 220 models are "cheaped out"  and there are 330 470, 560's that are all substantially better quality.
 
the 3090 FTW3 Ultra has 220's - According to the video these are the cheap models with lowest performance for SPCAPS and they will offset any of the higher performance by using the MLCC's if I understood correctly what he was explaining.  Other's are using the 470's. the FE used 2X MLCC, 2X 470, 2X 220.
 
so this may still not be over yet.  I'll say that I personally do get crashing often when the card hits over 2000mhz on GPU even with a 56c core temp. but that could be all sorts of reasons why. 

I do really hope zotac comes out with a statement like jacob did, zotac is pumping those cards out to people

jmehalik
I do really hope zotac comes out with a statement like jacob did, zotac is pumping those cards out to people

In the buildzoid video you can see that zotac also doesn't have the pads for the 10xmlcc under the SP-caps, and they are using smaller ones than some of the other vendors, so even fixing this problem would be impossible for them with out a new pcb.  But its important to note that Zotac gives no guarantee that their cards will be stable at 2000+ MHz.  Its likely to not even be that much different overall in performance.  If you buy a Zotac you should have no expectation that your card is going to be one of the best OC cards out there.  From that perspective you could guess which brands might be better at it, but if you want the highest performing cards you have to wait for reviews.

SP-Caps are also a Panasonic name for their Aluminum based caps, with the POSCAP being Tantalum based
 
That said, SP and POS are definitely different, as tantalum caps usually have a better stability/reliability rate, and are almost as good as MLCC with some of the newer variations

Interesting to read through all of this.  Glad EVGA found the issue.

Lokator
I am mad I just sold my 2080 Ti on ebay.... with this knowledge a week ago I would have sit that out and enjoy my 2080 ti... if I get the 3080 FTW3 via amazon soon and it has low amount or no mlcc, I might consider sending it back after 30 days

NO way I'm selling my 2080Ti now!
 

Does the FTW3 have this issue?