EVGA Is Immediately Replacing All RTX 3090s That Died From New World

EVGA GeForce RTX 3090 FTW3
(Image credit: EVGA)

According to an EVGA discussion with JayzTwoCents, EVGA's customer support is in full swing, replacing all of its customer's RTX 3090s that have died from Amazon's New World game. Customers that RMAed their RTX 3090 yesterday reportedly already have new 3090's being shipped to their doorstep.

This is some excellent customer support coming out from EVGA. Even before the company has a chance to test dead cards, it's shipping out replacements to sad and frustrated consumers. EVGA is even cross-shopping cards when necessary to accelerate delivery times. (Hint: It's always necessary!)

If you have an EVGA RTX 3090 that has died, whether from New World or something else, we would highly recommend getting in an RMA as soon as possible to take advantage of EVGA's warranty and customer support. 

RTX 3080 Ti's and RX 6000 Series GPUs Are Also Dying From New World

Unfortunately, it appears GPU deaths are going far beyond that of RTX 3090 owners. JayzTwoCents reports a bunch of other models, including RTX 3080 Tis and most of AMD's RX 6000 series lineup, dying from playing New World.

In fact, JayzTwoCents delved deeper into the issue by testing two of his own RTX 3090's — one EVGA and one MSI RTX 3090 — and found the game was pushing the EVGA RTX 3090's power consumption over 20% higher than what its power limit should allow. This translates to over 400W of power being used by the RTX 3090 when the stock power limit for the FTW3 is 360-375W at best.

Surprisingly, the MSI RTX 3090 did not have the same issue, with power numbers staying right in line with the power limit's maximum allowed wattage. So it appears some cards are exhibiting abnormal power consumption figures when playing the game, for unknown reasons at this time. That explains why only select cards are dying.

It's an unusual situation to say the least. Back in the day, before AMD and Nvidia started implementing better power limiters in their hardware, it was entirely possible to fry some GPUs by running the FurMark stress test. Protection against 'power virus' workloads like FurMark have been common for several generations of GPU hardware now, however, so it's odd that New World can end up in a similar power overload situation. It's also unclear whether the game causes a VRM, capacitor, VRAM, or some other component to fail. Our advice is to be wary of playing the beta for the time being, at least until others have verified the patch fixes the root issue.

Aaron Klotz
Contributing Writer

Aaron Klotz is a contributing writer for Tom’s Hardware, covering news related to computer hardware such as CPUs, and graphics cards.

  • derekullo
    If they had been mining Ethereum their card might still work.

    Never thought I'd say that lol.
    Reply
  • setx
    It's obvious why it's happening: cards are clearly have inadequate design and cheaping out on power delivery/cooling/protections. And the main blame takes Nvidia as they are limiting stability tests like FurMark in drivers due to their "unrealistic" load. And when some game or program actually does similar load without being known by driver? It fries the card.
    Reply
  • Alvar "Miles" Udell
    Know what this reminds me of?

    https://www.eteknix.com/evga-gtx-1080-caught-on-video-catching-fire/
    Reply
  • setx said:
    It's obvious why it's happening: cards are clearly have inadequate design and cheaping out on power delivery/cooling/protections. And the main blame takes Nvidia as they are limiting stability tests like FurMark in drivers due to their "unrealistic" load. And when some game or program actually does similar load without being known by driver? It fries the card.
    My thought exactly. EVGA replacing the cards is a PR move, so people won't question the quality. Now I know it's better to avoid thier cards.
    Reply
  • King_V
    Absolutely kudos to EVGA for their consumer-friendly action on this.

    At first, I was wondering if it was some kind of error in the power regulation design by Nvidia, or if it was specific to a custom design on the part of EVGA for the 3090. But now that it's exhibiting these symptoms on other models of Nvidia cards, AND on AMD . . I'm wondering exactly WHAT is going on with this game?

    What did they do that's somehow tricking video cards to ignore their power limits?
    Reply
  • hotaru251
    setx said:
    It's obvious why it's happening: cards are clearly have inadequate design and cheaping out on power delivery/cooling/protections. And the main blame takes Nvidia as they are limiting stability tests like FurMark in drivers due to their "unrealistic" load. And when some game or program actually does similar load without being known by driver? It fries the card.



    Have you ever watched XOC?

    Those guys push these gpu's WAAAY past what any game could want. yet..they dont die (usually and when do its users fault for failing to keep soemthign in control but thats besides point)


    power delivery is likely fine. (in fact if that was what killed em you'd easily be able to tell as something would be fried usually)
    cooling is likely fine (while the cards do get hot their built in protections throttle if you get too hot)
    protections work, but are not infallible. If you keep relying on it then it will eventually fail and they can not protect vs everything (as they have to react and if the dmg is faster than that they cant do anything)


    issue is likely what happened in past where GPU were allowed to go too far and caused issues. (this happened in past and vnidias fix was to lower clocks in firmware to prevent the issue from happening)

    saying its on nvidia is liek saying its an engine makers fault for their engine dyign when yo never let off gas and burn it out until its damaged and dies.

    No game should have uncapped menu screen (no reason to EVER push more than ur monitors refresh rate when its just pictures and txt usually static) as casues un-needed heat/power usage.

    Mentioning FurMark is pointless too as thats a situation that wont ever be realistic especially since w/o OCing your gpu its never gonna be an issue (and if ur unaware... any OC over nvidias specs is NOT their fault as thats out of spec)

    and if you seriously think amazons game is actually as hard as furmark ur mistaken. The games specs are nowhere near as demanding..heck ashes of singularity is harder on gpu.

    Called Amazon's game has issues which if u watched video on the post is shsown...the game is drawing more PWR, yet not using said power to do anything about clock speeds/performance...its drawing power for literally no reason.

    tommo1982 said:
    My thought exactly. EVGA replacing the cards is a PR move, so people won't question the quality. Now I know it's better to avoid thier cards.
    please read before you post false info.

    if you read (can even go reddit for other cases) page you will see it s not only 3090's and it isnt only EVGA gpu's.
    and effects AMD gpu as well (and if ur unaware...evag does NOT deal with amd gpus so those would be on rest of the brands)

    Alvar Miles Udell said:
    Know what this reminds me of?

    https://www.eteknix.com/evga-gtx-1080-caught-on-video-catching-fire/
    thats kind of different issue.
    that WAS due to cooling issues...and that lead them to make their ICX update (to either 2.0 or 3.0) and monitor more than just the core of gpu. As back then they didnt bother to give the vrm any worry (as those were usually not an issue in past)



    TO those who saying its a PR move on EVGA you have likely never owned anything of theirs. Their Customer Support is likely best in the GPU space. Timely and helpful.


    Again issue is likely due to amazon's poor choice of limits (as w/o drivers from nvidia its ALL up to them to control the hardware) and that GPU being OC'd and just running wild w/o drivers to tell em "oh sh*t" and back off.
    Reply
  • daworstplaya
    Again Evga shows they have excellent customer service. I'm willing to bet a new vBIOS comes to fix said issue. The only thing I can think of, is the game has found a bug in the vBIOS of the cards that allows it to go beyond the allowed power specs.
    Reply
  • Metteec
    Oh goody, another VBIOS update. The last one I installed in my EVGA 3090 XC3 Ultra bricked my card. It cost me over $100 to ship the card back to EVGA because of the insurance.
    Reply
  • warezme
    EVGA's replacement of cards is not a PR move. That's what they do. They have one of the best RMA and customer support in the biz. I have bought tons of EVGA stuff and they have never given me crap over anything from motherboards to video cards even cross shipping as in these cases and at least with their older products not sure since I haven't purchased EVGA in a while they have a life time warranty as long as you register online.
    Reply
  • setx
    hotaru251 said:
    saying its on nvidia is liek saying its an engine makers fault for their engine dyign when yo never let off gas and burn it out until its damaged and dies.
    You have no idea what you are talking about. No user-mode software shall ever have the possibility to damage hardware. Period. No matter by accident or being malicious. That's how properly designed system shall behave.

    hotaru251 said:
    and if you seriously think amazons game is actually as hard as furmark ur mistaken. The games specs are nowhere near as demanding..heck ashes of singularity is harder on gpu.
    I seriously think that concept of blacklisting specific programs in drivers instead of implementing proper hardware limits (that would react faster than software and work even if software part is somehow screwed) is fundamentally flawed. Yes, it's cheaper, bit it's the wrong way.
    Reply