I will add my observations in here.  I initially thought I had a problem with my card and heat as well.  What I found was that it was entirely the airflow feeding my GPU.  

Besides my FTW3 Ultra I have a 3900x, with PBO on, on a 240mm AIO.  I have 2 Pcie SSD's situated in slots 1 (just above the gpu) and slot 2 (just below the gpu).  I run a positive pressure airflow in my case when at idle, and when both my CPU and GPU are at max, it turns into a slightly negative pressure airflow.  All fans are Noctua's 140mm industrial 3000 rpm pwm server fans (I had about 10 extra sitting around from an old server build).  All of this in a Phanteks p600s case (this becomes relevant to testing).  I live in an area that is warm atm, ambient being around 24C in the room my computer is in.  My cases average ambient is around 32 to 34C. Edit: I had just replaced a 1080 Hybrid card.  
 
Originally I had my AIO as intake.  At stock I noticed my gpu hitting around 83C in most benchmarks.  SO I went digging.  I went in to switch my Radiator to a top exhaust.  During this I found that the bottom 140mm fan on the front had stopped working at some time and I had not noticed.  So I replaced that.  After the rad swap to the top and fixing the fan, my temps went down to 77C.  Not quite where I wanted them just yet.  So I did some more digging. I started adjusting my case fan rpm's as they were still set to run with the AIO at the front and a Hybrid GPU.  Originally I had my front intake at about 800rpm.   To make things easiest on myself I maxed them out for a couple benchmark runs (3000 rpm), I also maxed out the GPU fans at 100% (seems to be 3000rpm).  I found my temperatures stabilizing at 65C.  Thus I assumed it was airflow, so I went to adjusting my case fans to what I consider to be my, personally, noise acceptance level.  For my rear exhaust this was 1800rpm. For my 3 front fans this turned out to be 1000-1200 rpm.  For the GPU I found this to be 80ish %.  At this level I find my GPU under stress was hitting 69-72C.  More than acceptable now.  
 
Now as to the case.  I did some additional testing using my case.  For those who don't know, the p600s has a removable top plate and a removable front plate held by magnets to increase the noise reduction the case is capable of.  Behind the front plate is mesh.  Behind that mesh, just in front of the fans is a dust filter.  If I put the top plate and front plate both on, with my fans at the noise adjusted rpm's stated above, my gpu went back up to around 82C.  Taking just the top panel off (exhaust fans passing through a rad) I saw gpu temps hit 79C.  Taking off the front plate saw temps go back down to around 71C.  I took it further though.  I removed the dust filter that sits between the front panel and the front intake fans.  Doing so saw yet another temperature decrease down to 67C.  Sadly I have pets in the house, so a dust filter is a must. 
 
All that to say, that for me, it was definitely an airflow problem.  
 
Also as per Masterkaj above. I've just now run Superposition at stock at 8k optimized with 50% GPU fans.  Running at my optimized case setup as I would for gaming or rendering tasks, I saw a max temp of 75C (current ambient is 25.5c).  Using my baseline Overclock of +125 core and +250 mem (being what I got doing a quick OC in an hour the other night) and power and CPU temp slider being maxed out, my temp maxed out at 77C. But Superposition doesn't stress my cpu at all, so my AIO exhaust doesn't really ramp up. Probably resulting in a positive pressure environment in my case during these tests.  

73 highest temp.
 

80C for me running the superposition benchmark

We go through this same thread EVERY time there is a GPU release. My FTW3 Ultra on air never goes above 75*C at full load, which for a card consuming this much power is actually kind of amazing. I have a 1000D with 15 fans in it though so airflow is not a problem for me. 
 
These types of threads are almost impossible to get any real data from since everyones case / airflow / fan / ambient temp situation is different. 

When posting temps it is helpful to also say what power limit you are using and if the card is operating on the power limit.  Some lighter loads max out the card's clock and voltage but don't come near the power limit.  
 
Power draw is the biggest driver of temps. 

I'm hitting ~60c under load typically.
 
I'm using a Phanteks P500A case with plenty of airflow and running the OC bios.

masterkaj
So I think if people want to have a more apt comparison for temperatures, we need to run similar benchmarks or stress tests.  A good free one would be Unigine Superposition 8K Optimized and control the settings like Target power 100%, Target Temp 83*C, no overclock, set fan manually to 50%.  This controls for testing variation and limits it to case airflow differences.

 
My result, ambient 22 °C - 86 °C at the end and also still climbing. My boost was also never higher then 1950 Mhz.
 

Mkilbride2599
Wow these temps are worrisome. I ran Superposition 8K on a loop for like 20 minutes. I maxed out at 61c. 69% fan speed. Couldn't hear it at all.  Meanwhile people here at posting temps 80c and calling it "fine". While also being 100$ more expensive...
 
Gah, if only I had a FTW3 myself to test. My Tuf results are the same as other Tuf reviews, of which there are a lot. But I see such deltas here.  I love EVGA; but there's no denying based on posts here it runs significantly cooler, which is better for performance and lifespan. Because 70c+, GPU boost starts downclocking your card.

 
We were testing at 50% fan speed... Not 69%...
 
 

charmin35

fofal

Look at his temps. That's strix bios? Power limit is higher then 107.

Oof that video is painfully laggy, it's like the 3080 is sucking the performance from his phone

Funny you mention that.  JayzTwoCents mentioned in one of his recent videos that they were seeing Phone to GPU interference while benchmarking.

kevinc313
 
 
Funny you mention that.  JayzTwoCents mentioned in one of his recent videos that they were seeing Phone to GPU interference while benchmarking.

Oh yeah that is crazy stuff, I think it was his wireless microphone? I wonder if that is still happening after the Nvidia driver update. That was an interested drama that heated up about the power filtering caps or whatever they were. 

SprayingMango
We go through this same thread EVERY time there is a GPU release. My FTW3 Ultra on air never goes above 75*C at full load, which for a card consuming this much power is actually kind of amazing. I have a 1000D with 15 fans in it though so airflow is not a problem for me. 
 
These types of threads are almost impossible to get any real data from since everyones case / airflow / fan / ambient temp situation is different. 

I agree to an extent, I think the biggest difference is the cooler on the XC is very minimal compared to the FTW3, I think the FTW3 is fine and the XC cant cool at stock settings. 

Mkilbride2599
 
You see no issue with it being 75c, while a card 110$ cheaper never goes above 60?

Seems like an FTW3 design flaw.  No one should be seeing temps this high when other AIBs are 10-15 degrees cooler.

I think the Asus Tuf specifically cools better because it has a separate heatsink for the memory, the FTW3 uses one copper plate for GPU and memory together.
 
https://www.techpowerup.com/review/asus-geforce-rtx-3080-tuf-gaming-oc/2.html
 
the 2080ti and 1080ti FTW3 also used design more similar to this

mjm5625

 It sounds like a jet engine or something

lol after months of battling my Super black's temps I think that line there should be the catchphrase of any evga fan.    Hybrid or Copper only for me from this day forward.

SprayingMango
We go through this same thread EVERY time there is a GPU release. My FTW3 Ultra on air never goes above 75*C at full load, which for a card consuming this much power is actually kind of amazing. I have a 1000D with 15 fans in it though so airflow is not a problem for me. 
 
These types of threads are almost impossible to get any real data from since everyone's case / airflow / fan / ambient temp situation is different. 

The voice of reason....110% agree.  Case design, fans, airflow, ambient temps, etc, etc......If you want a fair comparison, you need to limit the variations to a minimum.  
 
From what I saw in that one video of that was posted a few times, that youtuber was running ASUS Strix bios on his EVGA card, if I read this all right?  I'd like Jacob to chime in here, but, if that is correct, what would EVGA's warranty stance be if someone wanted to RMA their card for issues and they loaded other AIB's BIOS on the card (assuming you knew about it! ).  Also, in that same video he seems to be running at 71% power target.  I'm going to take an educated guess, THAT is the reason he has lower temps during the test?  It does not make sense to me that ASUS's bios would LOWER the temps that much, unless they are cutting back clocks, power limits or voltage, or any combination of those WITHOUT sacrificing much performance.  If that is true, then EVGA should be able to optimize their BIOS to get better as well.  Just a thought, maybe I'm crazy...it is 4AM here and I need to go to bed!