• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

RTX 5090 FE Molten 12VHPWR

TaintedSquirrel said:
The guy who hates Nvidia managed to toast both ends of his cable on basically the best GPU and PSU on the market... Amazing, truly.
Click to expand...
You people that keep defending Nvidia and the cable are straight up retarded. The fact It keeps happening and will happen more often as the cards and cable age should be very concerning for all of us.
I never had a card that yearly maintenance was I guess I better change the cable and hope the ends on my GPU don't melt. Or having to buy a new PSU or device I put on my card to tell me if it's going to melt or not.

Something is very wrong with everything I wrote in that paragraph as none of it should happen. The plug should never be a weak point because it's the easiest thing mathematically to calculate the safety factor on and I think we went form like a 1.8 on pcie to I think 12vhp is only 1.2. It's suboptimal power delivery for 600w and really should have been power caped at 500w but Nvidia wanted a 5090 card but at 500w it wasn't fast enough so they upped it to 600w and said o well. All 5090s should have nee two plug cards.
 
tunatime said:
You people that keep defending Nvidia and the cable are straight up retarded. The fact It keeps happening and will happen more often as the cards and cable age should be very concerning for all of us.
I never had a card that yearly maintenance was I guess I better change the cable and hope the ends on my GPU don't melt. Or having to buy a new PSU or device I put on my card to tell me if it's going to melt or not.

Something is very wrong with everything I wrote in that paragraph as none of it should happen. The plug should never be a weak point because it's the easiest thing mathematically to calculate the safety factor on and I think we went form like a 1.8 on pcie to I think 12vhp is only 1.2. It's suboptimal power delivery for 600w and really should have been power caped at 500w but Nvidia wanted a 5090 card but at 500w it wasn't fast enough so they upped it to 600w and said o well. All 5090s should have nee two plug cards.
Click to expand...

They changed the socket and connectors how many times now and they still burn? I mean 8-pins burned once in a while but it was generally the faulty power delivery on the card that caused this rather than the connector itself. I always felt like if they just made the same12hpwr but 30% bigger it would turn this into a non-issue. Pins and wires too small, too close together, no load balancing on the cards that use them like we used to get on older cards. Just a whole bad plan, although I do appreciate the clean look and ease of use.
 
LigTasm said:
They changed the socket and connectors how many times now and they still burn? I mean 8-pins burned once in a while but it was generally the faulty power delivery on the card that caused this rather than the connector itself. I always felt like if they just made the same12hpwr but 30% bigger it would turn this into a non-issue. Pins and wires too small, too close together, no load balancing on the cards that use them like we used to get on older cards. Just a whole bad plan, although I do appreciate the clean look and ease of use.
Click to expand...
They really should only be on cards that are under 500w or like you said a 30% higher safety factor would make this a no issue
 
tunatime said:
The fact It keeps happening and will happen more often as the cards and cable age should be very concerning for all of us.
Click to expand...
There have been more reports of burnt and slightly blown up 9800X3Ds than there has been of RTX 5090's. And yet, no one here sees endless threads raging against the big bad green company.

That being said, no one is defending Nividia or AMD for each of their minuscule issues. And yes, a self immolated CPU or GPU out of tens of thousands is minuscule. It's even more of a nothing-buger when you factor in the fact that some of the CPU and GPU reports are clearly user error.

Oh wait, it twitched! Get'em boys!

1788405123579.png
 
AVATARAT said:
Click bite video or he really do stupid things, I am not sure.
Click to expand...
In his case, I'm inclined to say neither. I'm inclined to say there's a defect in the power cord, or the PSU connector. He said he's used the 5090 before for testing so it is presumably ok. Only 2 wires sending current indicates something pretty serious is wrong. I'd be inclined to blame the cable, since not only is that the part that gets moved around the most but it is also probably the cheapest/least tested.

The only thing he may have done to contribute is if that same cable was getting used over and over for tests that may have worn it out. In general, he seems to be a dude who knows what he's doing and doesn't get involved in a lot of the drama that some of the tech tubers do.


Still though, while this is not super common I feel like this could really be made almost zero problem with just 4 more wires: Add two more connection pairs, keep max power draw the same. It doesn't make the connector much bigger, but now you have a huge safety margin. Hell they could even make it backward compatible: You can still run the 6-pin connector in the 8-pin socket just with less safety margin.
 
ZeroBarrier said:
There have been more reports of burnt and slightly blown up 9800X3Ds than there has been of RTX 5090's. And yet, no one here sees endless threads raging against the big bad green company.

That being said, no one is defending Nividia or AMD for each of their minuscule issues. And yes, a self immolated CPU or GPU out of tens of thousands is minuscule. It's even more of a nothing-buger when you factor in the fact that some of the CPU and GPU reports are clearly user error.

Oh wait, it twitched! Get'em boys!

View attachment 823532
Click to expand...
Apples and oranges.

On one side you have a $5000+ card where it's abundantly clear the design is meant to pinch pennies. On the other side you have a $400 CPU where it isn't clear whether the problem is being caused by AMD or motherboard manufacturers. But sure, let's keep making excuses for one of the most profitable companies on the planet, because they clearly need those pennies.
 
gtrguy said:
Could you share those numbers you’ve compiled with us?
Click to expand...
I'm not invested enough in this dead horse to do the research for you or anyone else, if I'm being brutally honest with you. If you are that invested, it won't take you long to find that information; since you seem to not been paying attention the last few years.
 
ZeroBarrier said:
I'm not invested enough in this dead horse to do the research for you or anyone else, if I'm being brutally honest with you. If you are that invested, it won't take you long to find that information; since you seem to not been paying attention the last few years.
Click to expand...
Yes, you have no idea of actual numbers (nor do I) so what you’re saying is simply what you feel the situation is. Got it.
 
ZeroBarrier said:
There have been more reports of burnt and slightly blown up 9800X3Ds than there has been of RTX 5090's. And yet, no one here sees endless threads raging against the big bad green company.
Click to expand...

“Murderboards” literally became popular slang for AsRock AM5 motherboards, but ok.
 
ZeroBarrier said:
I'm not invested enough in this dead horse to do the research for you or anyone else, if I'm being brutally honest with you. If you are that invested, it won't take you long to find that information; since you seem to not been paying attention the last few years.
Click to expand...

So you haven’t checked into the numbers on this, you just know the numbers on this?

I mean you seem to be reasonably invested in this because you enter every thread about it trying to tell everyone this isn’t a big deal.
 
Slapped it into chatgpt and here's the answer. (question was "have there been more reports of burnt 9800X3Ds than there has been of 12vhpwr cables?")

And this is only the officially recognised cases.

No. Based on the publicly documented cases, there have been substantially more reports of melted/burned 12VHPWR connectors than physically burned Ryzen 7 9800X3Ds.

The distinction that matters is that there have been 100+ reports of 9800X3Ds “dying”, but most of those CPUs were not burned. In Gamers Nexus’ early sample of 42 9800X3D failures, only 4 cases showed burns or physical markings. By the end of March 2025, community tracking had grown beyond 100 alleged failures, but reports included BIOS/POST problems and CPUs dying without visible damage, so they can't all be classified as burnouts.

For comparison, the 12VHPWR situation was already much larger:

IssuePublicly reported scale
9800X3D total failure/death reports100+ by March 2025
9800X3D reports with actual burning/physical burn marksAt least several documented; only 4 of 42 in GN's early dataset
RTX 4090 12VHPWR melting — Nvidia acknowledged~50 cases by Nov. 2022
RTX 4090 12VHPWR melting — NorthridgeFix~100 repairs/month in late 2023
RTX 4090 12VHPWR melting — NorthridgeFix later~200 repairs/month by April 2024
Nvidia itself said it was aware of about 50 melted-connector cases globally shortly after the RTX 4090 launch. More importantly, NorthridgeFix later reported receiving 20–25 melted RTX 4090s every week, roughly 100 per month, and by April 2024 said it was dealing with roughly 200 affected cards per month. That's one repair business alone, rather than a worldwide count.

So in raw numbers I'd characterize it approximately as:

12VHPWR/RTX 4090 melting: hundreds to potentially thousands of real-world incidents over its lifetime.

9800X3D physical burning: probably dozens rather than hundreds based on publicly identifiable cases, although there have been considerably more non-burning CPU failures.

There's another important difference: the infamous first severely burned 9800X3D that Gamers Nexus physically investigated appears to have been caused by the CPU being misaligned in the AM5 socket, allowing VDDCR pins to contact ground, rather than the CPU spontaneously overheating. Other later 9800X3D failures—particularly on ASRock boards—appear to be a separate phenomenon and remain harder to categorize.

So if someone is claiming “more 9800X3Ds have burned than 12VHPWR connectors,” I don't see evidence supporting that claim. The available evidence points quite strongly in the opposite direction.

If you want, I can also try to produce a best-estimate total number of burned 9800X3Ds vs melted 12VHPWR/12V-2×6 connectors through September 2026, going through the reported cases and eliminating duplicates.
Click to expand...


The extra thing it asked if I wanted to do in the bottom line.

Yes. After digging through the available case compilations, repair-shop data, manufacturer statements and Reddit reports, the gap looks much larger than the raw headlines suggest.

My best estimate as of 4 September 2026 is:

Failure typeBest estimate of physically burned/melted incidentsConfidence
Ryzen 7 9800X3D physically burned/scorched~20–50 publicly documented casesMedium-low
12VHPWR / 12V-2×6 melted connector incidentsat least ~450–500 documented casesHigh for the minimum
Likely real-world 12VHPWR / 12V-2×6 total~1,000–2,000+ incidentsMedium-low
So I would estimate there have been roughly 10–50 times as many melted 16-pin connector incidents as physically burned 9800X3Ds.

9800X3D: probably ~20–50 actual burn cases​

The biggest source of confusion is that people routinely describe a dead 9800X3D as “fried”, even when there is no physical burning.

Gamers Nexus' original carefully reviewed dataset contained 42 reported 9800X3D failures. Only four of those 42 had burns or physical markings — about 9.5%. GN specifically separated the scorch-mark cases from BIOS/no-POST failures.

By March 31, 2025, the community database had grown to 108 9800X3D failure reports, but Tom's Hardware explicitly corrected its coverage to point out that those weren't necessarily dead CPUs; some were boot problems.

The problem continued growing. By October 2025, reporting around the ASRock situation was discussing roughly 300 affected systems, although again this was the broader CPU-failure problem rather than 300 burned CPUs.

There's also a particularly useful later dataset. In June 2026, one Reddit researcher searched a full year of Reddit posts and then manually removed false positives and duplicates. They found 207 AM5 X3D CPU-death reports, of which 139 involved the 9800X3D. The search criteria included dead, fried, RMA, no-POST, black-screen and burn-mark reports, so the 139 number absolutely cannot be treated as 139 burns.

And physical-burn reports definitely continued into 2026. For example, a January 2026 9800X3D showed a visible frying mark after roughly a year in an ASRock X870 Nova. Another August 2026 case showed visible burned patterns on a 9800X3D used with an ASUS X670E board.

Using the approximately 5–10% physical-damage proportion seen in the better-documented early cases against the several hundred overall failures gives a central estimate around 25–40 visibly damaged CPUs.

I'd therefore use:

9800X3D physically burned/scorched: ~20–50 cases

rather than claiming that the hundreds of "dead 9800X3D" reports were all burnouts.

12VHPWR: we can establish nearly 500 without much estimating​

This side is much easier because there are some unusually good hard numbers.

First, Nvidia itself acknowledged approximately 50 melted RTX 4090 connector cases globally in November 2022. Gamers Nexus calculated that this represented roughly a 0.04–0.05% failure rate among approximately 125,000 early RTX 4090s.

Then came CableMod.

The U.S. Consumer Product Safety Commission recorded 272 reports of CableMod 12VHPWR angled adapters becoming loose, overheating and melting into GPUs. This isn't a Reddit estimate — it's the incident count used in the official February 8, 2024 recall.

That's already:

50 + 272 = 322 incidents

before counting the huge number of ordinary PSU cables and Nvidia adapters.

There's now also an ongoing community database maintained by MODDIY. Importantly, it keeps the recalled CableMod adapter incidents in a separate category, reducing the duplication problem. It currently lists 232 recalled CableMod-adapter melting cases separately.

For ordinary connector/cable incidents, its database recorded 64 12VHPWR melting cases during 2025 alone.

I counted its 2026 entries through September 3, 2026 and get approximately 81 additional 12VHPWR/12V-2×6 incidents. The database includes everything from RTX 4090s to RTX 5090s and even some Radeon cards using the connector.

It also contains another roughly two dozen qualifying cases from 2022–2024 outside its special CableMod-recall section.

After deliberately allowing for possible duplication between the Nvidia, CableMod, repair-shop and community figures, I think a defensible statement is:

At least ~450–500 distinct 12VHPWR/12V-2×6 melting incidents are publicly documented.
Click to expand...
And that's a floor, not an estimate of the true worldwide number.

The repair-shop numbers suggest the actual connector total is much higher​

This is where the difference becomes striking.

NorthridgeFix was receiving approximately 20–25 RTX 4090s with melted connectors per week by November 2023 — roughly 100 per month.

By March/April 2024, that one repair business reported replacing approximately 200 melted RTX 4090 connectors in a single month and showed a box containing the removed connectors.

These weren't 200 internet posts. They were actual GPUs that had arrived at one repair facility.

And the problem didn't simply vanish. NorthridgeFix was still publishing repairs of melted RTX 4090 connectors in January 2026.

So even though I wouldn't multiply "200/month" by several years — that would make an unjustifiably large assumption — the fact that one shop alone physically handled hundreds of affected GPUs makes a worldwide total exceeding 1,000 melted connectors very plausible.

My working estimate would therefore be roughly:

12VHPWR/12V-2×6 actual melting incidents worldwide: ~1,000–2,000+, possibly substantially higher.

Bottom line​

I'd revise my previous comparison to approximately:

9800X3D

  • Several hundred reports of CPUs dying/no longer POSTing.
  • Probably ~20–50 reports involving genuine visible burning/scorching/socket damage.
  • Some of those physical failures have plausible causes unrelated to the wider ASRock issue; GN's famous catastrophically burned MSI example was most likely caused by the CPU being improperly seated.
12VHPWR / 12V-2×6

  • ~450–500+ incidents can be substantiated relatively conservatively from documented datasets.
  • Hundreds of physical repairs were handled by a single repair business.
  • A reasonable worldwide estimate is 1,000–2,000+ actual melting incidents.
So no — the 9800X3D has not come close to overtaking 12VHPWR in physical burn incidents. Even if I took the high end of my 9800X3D estimate (50) and the absolute conservative floor for 12VHPWR (~450), the connector has about 9× as many documented physical incidents.

If we compare my central estimates — roughly 30–40 burned 9800X3Ds versus 1,000+ melted 16-pin connectors — you're looking at something closer to 25–35× as many connector failures.
Click to expand...
 
Last edited:
zehoo said:
Slapped it into chatgpt and here's the answer. (question was "have there been more reports of burnt 9800X3Ds than there has been of 12vhpwr cables?")

And this is only the officially recognised cases.




The extra thing it asked if I wanted to do in the bottom line.
Click to expand...
And Nvidia mods have been actively deleting burned connector posts on the main Nvidia subreddit. It's been rare for me to catch one before those mods delete them, but they still pop up there now and then.
 
ZeroBarrier said:
There have been more reports of burnt and slightly blown up 9800X3Ds than there has been of RTX 5090's. And yet, no one here sees endless threads raging against the big bad green company.

That being said, no one is defending Nividia or AMD for each of their minuscule issues. And yes, a self immolated CPU or GPU out of tens of thousands is minuscule. It's even more of a nothing-buger when you factor in the fact that some of the CPU and GPU reports are clearly user error.

Oh wait, it twitched! Get'em boys!

View attachment 823532
Click to expand...
Nope not at all the same. Nice try though. A CPU or mb having a problem inside the silicone or part of the power delivery system is something I can go well shit happens as it is a lot harder to deal with manufacturing defects that only effects some units (which btw is the only reason I didn't go with asrock and went with msi) then it is to ues math and know electrical engineering to make a power delivery system have a higher safety factor then 1.2. that's only 20% overhead when it works perfectly. So let's say each connection point is slightly worn out and at 95% of spec. Well now that 1.2 sf down to right around 1......now the plug starts heating up slightly and it gets worse till it melts.
 
zehoo said:
Slapped it into chatgpt and here's the answer. (question was "have there been more reports of burnt 9800X3Ds than there has been of 12vhpwr cables?")

And this is only the officially recognised cases.




The extra thing it asked if I wanted to do in the bottom line.
Click to expand...
Either he will question this research as AI Slop or he will go back into hiding until he sees this


Use AI Image Sep 4, 2026, 09_57_46.jpeg
 
heck i'm still wondering why they made the new connectors pegs half the size as the 'ol standard size and at the same time, spec'd it to handle double the power?

i guess the more they burn, the more they sell? :greedy:

mpg850gs.jpg
 
zehoo said:
Slapped it into chatgpt and here's the answer. (question was "have there been more reports of burnt 9800X3Ds than there has been of 12vhpwr cables?")

...
Click to expand...
Not disagreeing or agreeing with the conclusion, but it would be nice if the response had direct link references to cross check all of the data.
 
serpretetsky said:
Not disagreeing or agreeing with the conclusion, but it would be nice if the response had direct link references to cross check all of the data.
Click to expand...
If you cut/paste the question I asked into chatpgt it'll probably come up with the same/similar response with references. 4.6 set to high.
 
d3athf1sh said:
heck i'm still wondering why they made the new connectors pegs half the size as the 'ol standard size and at the same time, spec'd it to handle double the power?

i guess the more they burn, the more they sell? :greedy:

View attachment 823787
Click to expand...

My guess is the initial idea was to make it more compact, but it doesn't do any good if the gage is insufficient for the load, and that's particularly true when your load balancing sucks to the point where the PSU/card decides to send all of the current down two wires.
 
The other thing that makes this even dumber, is Nvidia prevents OEMs/partners from offering a choice of connectors. If Asus or whoever offered a model with multiple 8 pins, or an actually better connector, I would have paid more given the chance.
At this rate, it looks like Nvidia is just going to keep doubling down on this failure.
 
Aside from the crazy price, the power draw and 12vhpwr connector issue prevents me from ever getting an RTX 4090 or RTX 5090. I don't feel too worried with a RTX 5080 or RTX 5070Ti using the 12vhpwr since they're just drawing less power.
 
zerogg said:
The other thing that makes this even dumber, is Nvidia prevents OEMs/partners from offering a choice of connectors. If Asus or whoever offered a model with multiple 8 pins, or an actually better connector, I would have paid more given the chance.
At this rate, it looks like Nvidia is just going to keep doubling down on this failure.
Click to expand...
2 12v2x6 connectors for any card that can use more than 450 watts, and 3 connectors for more than 900 watts. That's what they should do. The other thing is for the custom cards, don't tie everything into a common bus. It can even be advertised as a feature- far more robust power delivery than reference cards.
 
Tsumi said:
The other thing is for the custom cards, don't tie everything into a common bus. It can even be advertised as a feature- far more robust power delivery than reference cards.
Click to expand...
Has anyone actually run the numbers on this idea. Or run a spice simulation?
Every time i see this idea I have a hard time seeing how this would help. The theory is that each VRM gets its own pin from the connector, and the VRMs would be setup to load balance, meaning each pin would be forced to use the same amount of current even if the pin has a lot of contact resistance? Did I get that right?

edit: also realized I dont know how multiphase VRMs load balance. Is the input current the same? Probably not right? probably the output current is roughly the same on each VRM. So if there is a voltage drop on the input the input would actually need to pull even more current to make up for the power loss from the voltage drop.
 
Last edited:
serpretetsky said:
Has anyone actually run the numbers on this idea. Or run a spice simulation?
Every time i see this idea I have a hard time seeing how this would help. The theory is that each VRM gets its own pin from the connector, and the VRMs would be setup to load balance, meaning each pin would be forced to use the same amount of current even if the pin has a lot of contact resistance? Did I get that right?
Click to expand...
That does seem to be an issue for the "just do separate pin loads!" crowd. The way a modern card works, I can't think how that would be done without more expensive hardware. There's no way to evenly divide the component load across pins, particularly since there are different VRMs for different things (core, memory, etc) that aren't all loaded evenly. So you'd realistically have to have the card made where the pins come in, go through a current balancing setup, then get hooked to a combined bus that goes out to all the components. It wouldn't be as cheap or easy as just knocking a resistor on a pin or separating VRMs per pin as people like to say. It would be a whole thing to balance everything.

Probably cheaper to just make a connector with a higher safety factor via more pins than to do that.
 
zandor said:
Ouch. Hopefully it's repairable. I'll be sure to have the Asus monitoring utility going whenever I get around to trying out DLSS5 & set it to shut down my rig if current goes too high on a pin.
Click to expand...
something just kept smelling funny, i thought it was me

but decided to check that 12VHPWR

it was extremely hard to remove and i could tell it melted and began plastic welding itself together
 
zandor said:
Ouch. Hopefully it's repairable. I'll be sure to have the Asus monitoring utility going whenever I get around to trying out DLSS5 & set it to shut down my rig if current goes too high on a pin.
Click to expand...
it was certainly solid streaming consistently pulling above 600 watts

image0.jpeg
 
Blurry photos, but RIP.

I sleep a bit better at night after having added a TG Wireview Pro but eh, I'm still not confident I won't have issues as I intend to keep the card for a long time in this economy.
 
El_Capitan said:
I don't know why it's been so slow to adopt 24V's.
Click to expand...
You'd need new, incompatible, PSUs. With any new 12v standard, adapters are a thing. You can make one cable format work with another. Sure eventually you end up not being able to do that, you aren't getting old school 4-pin Molex to 12vhpwr. However it is a slow transition, so people don't need a new PSU.

You make a 24v GPU? You HAVE to get a new 24v PSU and people won't like that. What's more, it'll be an extra expensive PSU because it probably will want to provide lots of 12v power too, which means a big set of DC-DC converters. Makes it something no company wants to jump on.

That said, I think it needs doing. I don't know what specific voltage, I'd leave that up to engineers and we don't need to have it be a specific multiple with how converters work now, it could be 30v or 42v or whatever ends up being a good choice for the balance of reasons. But I think it needs doing, and while the transition will be painful, it'll mean cleaner wiring all over once it is done in addition to fixing the GPU issue. IF boards move to a higher power, it either means they can have less power connectors or that they can provide more power through the board. The second is what I'd like to see, where the board has big power traces and we start powering more things off the board itself, rather than wires running all over the computer.

But the issue of it being a "new PSU or bust" standard are what holds it back. Cars are the same way. Newer cars need a LOT more electrical than old ones and it would make sense to move to a higher voltage, which has been proposed. However nobody has wanted to do it because it means that all existing components would not work or would need a converter. So we have cars with 300a alternators and massive fucking cables still on 12v.
 
You must log in or register to reply here.
Back
Top