Monday, September 21st 2026

NVIDIA RTX 5080 Reportedly Dies After DLSS 5 Test in Modded GTA 5

An MSI GeForce RTX 5080 16G Ventus 3X reportedly stopped working after its owner injected an unofficial version of DLSS 5 into a heavily modded copy of Grand Theft Auto 5. Reddit user "wax469" said the card had been running fine for about a month before the crash. According to the post, the screen began flashing uncontrollably about five minutes into a session, and the system refused to boot afterward. Swapping to his Ryzen 9 9950X3D2's integrated graphics let the PC start normally, and reinstalling the RTX 5080 brought the failure back. A replacement card of the same model has apparently worked without issues since.

It's worth being cautious here. NVIDIA has not released DLSS 5 for GTA 5, so the injection relied on unofficial tools, and the user says the card was running at stock settings with no overclock. There's no photo evidence of physical damage, and it's entirely possible a manufacturing defect was already present and just got exposed under load. Whether DLSS 5 had anything to do with it remains unconfirmed. The user was running a heavily modified version of GTA 5 Legacy, and adding unofficial DLSS 5 seems to be what eventually led to the failure.
Still, this isn't the first time DLSS 5 has been tied to a hardware scare. Tom's Hardware has reported that DLSS 5 is linked to a significant increase in power draw. We also reported recently on an RTX 5090 whose 16-pin connector melted during official DLSS 5 testing in NBA 2K27, with power draw spiking past 600 W. That case, too, had more questions than answers. For now, the advice is to stay cautious with unofficial DLSS 5 ports and mods, especially when graphics card prices are at an all-time high right now.
Sources: Reddit, via Videocardz
Add your own comment

53 Comments on NVIDIA RTX 5080 Reportedly Dies After DLSS 5 Test in Modded GTA 5

#26
kazuviking
Vayra86Should... but then there was this event not too long ago
Yeah because it was shitwga. Every single 20 and 30 series gpus they made had mossfet failures, not to mention the ticking time bomb their rgb circuit was. Thank god they dissappeared from the market before butchering even more gpus. And this is ignoring the corrosive thermal pads.
Posted on Reply
#27
No9999on3DMarkSteelNomad
Sounds like bad luck. I have a Gigabyte 5080 that I got because it was returned to the store for whatever reason. Unlocked with mvolt extreme to 600 watts, the 12V is barely plugged in (there’s a noticeable gap but I don’t like to force these plastic connectors), and I’ve been running DLSS 5 with over volt and OC. Runs fine even if it’s a 50% increase in power for certainly not that much performance.
Posted on Reply
#28
JustBenching
Visible NoiseMy car was fine until I went out to eat at P.F. Chang’s. Then I got a flat tire.

Proof that chicken teriyaki causes flat tires. Prove me wrong.
The moment I woke up it started raining. Therefore it's Nvidias fault.

I don't know, are these people for real or are they just throwing stuff and hope it sticks?
AusWolfThat's strange. I wonder if it's a design flaw, or if it's by design. AMD cards work at slightly lower clocks when the RT cores are active, as the power limit is more evenly distributed across the die. Maybe Nvidia doesn't want the same to happen, and lets the cards eat more than they should when the Tensor cores are hard at work with DLSS 5 just to maintain clocks. Just a theory.
It does not. It completely respects the power limit. I've run dlss 5 at a 220w power limit
Posted on Reply
#29
Shihab
TechConsoomerCorrelation != Causation. Clickbait/lazy reporting by TPU. If my GPU died while watching Youtube, will the media outlets send headlines like: "RTX 5090 reportedly dies after playing Youtube video"???
You are implying that there is a correlation here. This anecdote doesn't even reach that bar...
Posted on Reply
#30
Blazeoptimus
I think the main thrust of the article is to 'just be careful', which is always prudent. DLSS5 stresses cards in ways that haven't been done before, so its possible that in some edge cases the card can damage itself if not properly controled.
Posted on Reply
#31
Shihab
BlazeoptimusI think the main thrust of the article is to 'just be careful', which is always prudent. DLSS5 stresses cards in ways that haven't been done before, so its possible that in some edge cases the card can damage itself if not properly controled.
There has been plenty of games that drove cards to their peak power consumption settings. Even before the RT era, it was pretty much the norm for AAA games to do so. And then you have applications a la furmark, bitcoin mining, archviz RT/PT and other gpgpu applications.
I personally despise DLSS5, but it'a hardly setting any new records here.

"Be careful" may as well apply to shadow quality settings, RT/PT toggles, or the good ol' resolution drop down menu.
Posted on Reply
#32
Chomiq
"Leak" DLSS5, burn 50-series GPUs from excessive GPU power consumption, get people to buy replacement 50-series GPUs at inflated prices, profit.
Posted on Reply
#33
Krit
Chomiq"Leak" DLSS5, burn 50-series GPUs from excessive GPU power consumption, get people to buy replacement 50-series GPUs at inflated prices, profit.
Smart move from nvidia.
Posted on Reply
#34
MikOnTopGG
So my GPU dies from an overriding software? I don't think that's how things work

Anyways, what's the credibility of Reddit?

We don't really care about their posts.

Most are fake anyways.
Posted on Reply
#35
john_
We know that DLSS 5 forces the card to consistently consume a high amount of power. As mentioned, FurMark can serve as an example from the past, and let's not forget that, specifically for FurMark, companies had at one point introduced safety features (throttling) in their drivers that activated whenever someone ran this software.

If this is not an isolated incident, we can expect Nvidia to appropriately modify the firmware of its cards at some point to prevent such occurrences. Perhaps more immediately and quietly, they might introduce a check in their driver to detect if someone is running DLSS 5, or more generally, if the GPU is operating at a level that demands more power than it is allowed to draw. Therefore, I imagine power consumption limits in the drivers will be reinforced, while there might also be further provisions for when a card consumes its maximum power capacity for a sustained period of time.
Posted on Reply
#36
windwhirl
MikOnTopGGAnyways, what's the credibility of Reddit?
It's "Ok". Not blind trust, not disregard.
Posted on Reply
#37
Shihab
Chomiq"Leak" DLSS5, burn 50-series GPUs from excessive GPU power consumption, get people to buy replacement 50-series GPUs at inflated prices, profit.
You're [probably] being sarcastic, but one must point out that 50 series is 1 year old. Nearly all of them would still be under warranty. This scheme would cost Nvidia and its AIBs, not benefit them.
Plus, their gaming-related sales are a tiny fraction of their total revenue. Even if this scheme ended with a net profit, it wouldn't make sense to risk damage to your brand just to add 0.0000001% to your overall net...
Posted on Reply
#38
Chomiq
ShihabYou're [probably] being sarcastic, but one must point out that 50 series is 1 year old. Nearly all of them would still be under warranty. This scheme would cost Nvidia and its AIBs, not benefit them.
Plus, their gaming-related sales are a tiny fraction of their total revenue. Even if this scheme ended with a net profit, it wouldn't make sense to risk damage to your brand just to add 0.0000001% to your overall net...
Hate to break it to you but when we see RAM and NVMe RMAs being denied due to increased prices in comparison to original purchase prices these GPUs are next in line.
Posted on Reply
#39
MikOnTopGG
BlazeoptimusI think the main thrust of the article is to 'just be careful', which is always prudent. DLSS5 stresses cards in ways that haven't been done before, so its possible that in some edge cases the card can damage itself if not properly controled.
Well, it can only damage itself from overheating or from some internal malfunction. A software can't break something physical, like a GPU Core.
Posted on Reply
#40
outlw
No9999on3DMarkSteelNomadthe 12V is barely plugged in (there’s a noticeable gap but I don’t like to force these plastic connectors)
Yeah, this is a time you should make an exception to your rule.
Press that connector in until it is fully seated!
MikOnTopGGWell, it can only damage itself from overheating or from some internal malfunction. A software can't break something physical, like a GPU Core.
It can, though, cause the card to have such a high sustained power draw as to cook the power delivery components on the PCB and/or the VRAM chips...
Posted on Reply
#41
Shihab
ChomiqHate to break it to you but when we see RAM and NVMe RMAs being denied due to increased prices in comparison to original purchase prices these GPUs are next in line.
The cases I am aware of and in which replacement was denied, still involved refunds. Even at the original selling price, that's still a loss to the seller.
Posted on Reply
#42
AusWolf
outlwIt can, though, cause the card to have such a high sustained power draw as to cook the power delivery components on the PCB and/or the VRAM chips...
Or the way some cards cooked themselves with Furmark back in the days - individual components on the GPU die getting the whole power budget of the entire card, creating hotspots that degrade the chip.
Posted on Reply
#43
outlw
AusWolfOr the way some cards cooked themselves with Furmark back in the days - individual components on the GPU die getting the whole power budget of the entire card, creating hotspots that degrade the chip.
Too bad there is not some kind of sensor reading that shows the hottest spot on the GPU...
:slap:
Posted on Reply
#44
MikOnTopGG
outlwYeah, this is a time you should make an exception to your rule.
Press that connector in until it is fully seated!


It can, though, cause the card to have such a high sustained power draw as to cook the power delivery components on the PCB and/or the VRAM chips...
Yeah, exactly. So, my point is, some lines of code can't damage the Core itself. Only the enormous sustained wattage required for the stability of the calculations.
Posted on Reply
#45
outlw
But...
The enormous sustained wattage is caused by some lines of code.

Either directly from the program, or from the firmware/drivers permitting that high of a sustained load.
Either way, it is being caused by software, and could be fixed by changing said lines of code 🤷🏻‍♂️
Posted on Reply
#46
Hecate91
No9999on3DMarkSteelNomad12V is barely plugged in (there’s a noticeable gap but I don’t like to force these plastic connectors)
The connector needs to be pushed in all the way, 12vhpwr is a tight fit but leaving it loose is a higher chance for it to melt.
Shihabthat's still a loss to the seller.
Oh no the poor seller. /s
Posted on Reply
#47
AusWolf
outlwToo bad there is not some kind of sensor reading that shows the hottest spot on the GPU...
:slap:
1. There wasn't always.
2. There's only so many sensors on one die.
Posted on Reply
#48
outlw
No, there was not always a Hot Spot Sensor, but the GPU still has it, and it would very likely be extremely useful in cases like this.
It really would be great if nVidia Allowed Access to it again.
Posted on Reply
#49
MikOnTopGG
outlwNo, there was not always a Hot Spot Sensor, but the GPU still has it, and it would very likely be extremely useful in cases like this.
It really would be great if nVidia Allowed Access to it again.
They also added the unlocked Voltage and Wattage control for a day or two, and then took it down as well, don't know if you can remember. Im gonna try to find that post. Was some time ago (1-2 months.)
Posted on Reply
#50
Nhonho
It's so simple to solve this problem. It would suffice, for example, to add four more pins and mandate connectors with metal locking clips on both the graphics card and the cable, along with a data connection between the power supply and the graphics card (and management software to prevent burnouts and short circuits).

Furthermore, switching the voltage to 24 V would also greatly help prevent these burnouts.
Posted on Reply
Add your own comment
Nov 22nd, 2026 01:16 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts