AMD Admits Faulty Vapor Chamber Causes RX 7900 XTX Throttling

Radeon RX 7900 XTX
Radeon RX 7900 XTX (Image credit: AMD/YouTube)

The Radeon RX 7900 XTX continues to be one of the best graphics cards on the market. Unfortunately, some reference models have been suffering from overheating. Independent analysis points to a faulty vapor chamber, and today, AMD has publicly acknowledged that the thermal solution is probably the issue behind the Radeon RX 7900 XTX's thermal throttling problem.

"We are working to determine the root cause of the unexpected throttling experienced by some while using the AMD Radeon RX 7900 XTX graphics cards made by AMD. Based on our observations to-date, we believe the issue relates to the thermal solution used in the AMD reference design and appears to be present in a limited number of the cards sold. We are committed to solving this issue for impacted cards. Customers experiencing this unexpected throttling should contact AMD Support." an AMD representative told Tom's Hardware. 

For obvious reasons, AMD didn't divulge how many MBA (Made by AMD) units are affected. "A limited number" is as vague as it can get and could mean anything from a bad batch to a couple of batches. The number will probably tally up to the thousands. On the bright side, the flawed thermal solution doesn't affect every reference Radeon RX 7900 XTX graphics card, only those unlucky ones from specific batches. Since custom designs don't use AMD's reference cooling solution, they're free of the problem. In addition, AMD has confirmed that the flaw only affects the Radeon RX 7900 XTX, so the Radeon RX 7900 XT is unaffected.

In AMD's defense, the chipmaker doesn't manufacture the Radeon RX 7900 XTX's thermal solution. The slip-up is probably due to AMD's supplier not performing proper quality control on the final product. However, it seems strange that an entire batch or batches came out defective. At the end of the day, it's AMD's product, so the company is responsible for its functionality, whether it was a mishap on the manufacturer's part or not. 

What's important right now is that AMD works with its supplier to narrow down the affected batches and issue a recall. Luckily for AMD, only a "limited number" of Radeon RX 7900 XTX graphics cards present faulty cooling, so while the chipmaker will lose some money over the fiasco, it's not like it's the end of the world.

Zhiye Liu
RAM Reviewer and News Editor

Zhiye Liu is a Freelance News Writer at Tom’s Hardware US. Although he loves everything that’s hardware, he has a soft spot for CPUs, GPUs, and RAM.

  • PlaneInTheSky
    AMD: "we are working to determine the root cause"
    Cut the crap AMD. You knew damn well you sold these faulty products. der8auer figured it out after 5 minutes but you were unaware what you were shipping or something? AMD was just hoping no one would notice the defect. Shameful.
    Reply
  • hotaru251
    PlaneInTheSky said:
    Cut the crap AMD. You knew damn well you sold these faulty products
    why?
    it makes no sense for them to do so.

    they gain nothing by selling some faulty coolers as they get to deal with RMA & service fees w/o getting any more $ out of it.

    and if not all batch thing they could of (unluckily) only tested ones w/o issue and deemed em all fine on their end.

    its not liek Nvidia's choice of adapter & melting risk (still bad but gpu still works w/o risk to anything)
    Reply
  • redgarl
    PlaneInTheSky said:
    AMD: "we are working to determine the root cause"
    Cut the crap AMD. You knew damn well you sold these faulty products. der8auer figured it out after 5 minutes but you were unaware what you were shipping or something? AMD was just hoping no one would notice the defect. Shameful.

    AMD took 5 days to issue a statement, Nvidia took a month and didnt issue a statement before GamerNexus did all the work themselves.

    And still, Nvidia didnt issue a replacement cable following standard and the PCI-SIG committee is blaming Nvidia directly.

    So go take a walk NVDelusional...
    Reply
  • OneMoreUser
    PlaneInTheSky said:
    AMD: "we are working to determine the root cause"
    Cut the crap AMD. You knew damn well you sold these faulty products. der8auer figured it out after 5 minutes but you were unaware what you were shipping or something? AMD was just hoping no one would notice the defect. Shameful.

    Are you really that oblivious to how things work in the world. Shameful to be posting such BS.

    No mass consumer products are tested to a degree where something like the issue at hand is going to be picked up with certainty, it is why there is warranties and the like.

    As describe in the article the issue is with the vapor cooling part, clearly some parts from a supplier has turned out to be faulty as in it not being a design error but a production error at the supplier. It might be some machines wasn't setup right, a part failed or it could be a operator error.
    Essentially something relative simple, something guys like der8auer could have guessed - in fact given the limited number of faulty reports and the way they came about he should have done so. Only it was of course better for him to do a massive stink and calling for AMD to issue a recall, seems like he found his target audience with you.
    Reply
  • spongiemaster
    redgarl said:
    AMD took 5 days to issue a statement, Nvidia took a month and didnt issue a statement before GamerNexus did all the work themselves.

    And still, Nvidia didnt issue a replacement cable following standard and the PCI-SIG committee is blaming Nvidia directly.

    So go take a walk NVDelusional...
    AMD has known about this issue for more than 5 days.


    How much have you heard about the 16 pin connector since it was determined to be user error? Once people were made aware of the issue and you only need to make sure the cable is properly seated, it magically stopped happening.
    Reply
  • ohio_buckeye
    In case anyone feels like playing lottery....

    SAPPHIRE Radeon RX 7900 XTX Video Card 21322-01-20G - Newegg.com

    Actually in stock.
    Reply
  • mr.stealth
    spongiemaster said:
    AMD has known about this issue for more than 5 days.


    How much have you heard about the 16 pin connector since it was determined to be user error? Once people were made aware of the issue and you only need to make sure the cable is properly seated, it magically stopped happening.
    Can't really blame it all on user error when the design of the connector makes it both inordinately difficult to fully insert, and to determine if it's fully inserted. And things tend to not be in the news as much once they are figured out, and new developments stop coming in.

    It will continue being a problem until the connector design is improved. Unless you have some way to guarantee that every future 40 series owner knows to use a degree of force that the average person would reasonably assume to be damaging to their new $1500 GPU. A note in the install directions to make sure it's fully inserted doesn't really cover that.
    Reply
  • TheOtherOne
    PlaneInTheSky said:
    AMD: "we are working to determine the root cause"
    Cut the crap AMD. You knew damn well you sold these faulty products. der8auer figured it out after 5 minutes but you were unaware what you were shipping or something? AMD was just hoping no one would notice the defect. Shameful.
    I remember not in very distant past, a certain company did exactly that with their highly anticipated and very overpriced product and interestingly enough that company has yet to publicly admit mistakes were made. 🤨

    I don't think AMD is that company tho! 🤔
    Reply
  • mathew7
    To all that compare this problem with the nvidia power pins:
    what do you prefer, a car that suddenly accelerates or a car that goes into "limp home" after 1 hour? Which do you think is really dangerous on the road?
    My point is that the nvidia power pins poses a potential fire hazard. This one just reduces it's performance (plenty of warning in case it DOES kill it).
    Reply
  • KyaraM
    redgarl said:
    AMD took 5 days to issue a statement, Nvidia took a month and didnt issue a statement before GamerNexus did all the work themselves.

    And still, Nvidia didnt issue a replacement cable following standard and the PCI-SIG committee is blaming Nvidia directly.

    So go take a walk NVDelusional...
    GN also concludes that it needs a very stupid or incompetent user to actually make them melt. You need to pull out the cable several millimeter and additionally tilt them for anything to happen. It's also such an obscure and specific user error that he himself needed a while to figure out what actually happened. Furthermore, he concluded that the occurrence of the burning issue is less than 1%. Compared to that, this vapor chamber is the bigger design flaw since it hits more people. And that has nothing to do with being delusional, that is just how it is. But I mean, if you have no argument, do an ad hominem against your opponent and call them delusional I guess?

    To refresh your memory, here is the testing with conclusion.
    ig2px7ofKhQView: https://www.youtube.com/watch?v=ig2px7ofKhQ
    Reply