Thursday, February 19th 2026

NVIDIA GeForce RTX 5090 Melts 12V-2×6 Connector Despite -100 W Power Limit

The melting and burning 12V-2×6 connectors on high-end NVIDIA—and some AMD—GPUs is well documented at this point, and although there is mounting evidence to support the fact that the new connector standard itself is partially to blame, there is still the idea that high power draw is to blame for these meltdowns. This is not necessarily the case, according to one user of the Mobile01 forum. According to the forum post about the incident, the top row of pins in a Gigabyte Aorus Master Ice RTX 5090's GPU power connector has melted down, despite the user having implemented a steep undervolt and power limit.

When installed, the user had limited GPU power to 500 W and dropped the voltage limit down to 0.9 v—that's a reduction of 100 W and 0.205 v compared to stock settings. It's worth noting that the GPU in question was connected to the PSU using a 16-pin adaptor, and the connector melted at the GPU side, not the PSU side, and the first indication of the meltdown was frequent crashing during gaming sessions. Both connector ends at the GPU side seem to be affected, but it's entirely possible that the GPU would have suffered even more damage had the power limit been higher. Currently, independent research seems to indicate that the 12V-2×6 cable itself seems to suffer from an unstable or inadequate connection, which increases the current and temperature across individual pins. Unofficial adaptors also seem to exaggerate these contact issues.
Source: Mobile01 forum
Add your own comment

133 Comments on NVIDIA GeForce RTX 5090 Melts 12V-2×6 Connector Despite -100 W Power Limit

#76
Dr. Dro
lexluthermiesterNot nearly as many. It was a rare occurrence and usually associated with poor quality PSUs.
Ugh. Something just dawned on me, because while you're not wrong, the same actually applies here, with less margin for error (which is what causes the cards to burn anyways) - and that people may have started to skimp on the quality of other components, such as the power supply, to accommodate rising RAM prices.
Posted on Reply
#77
dom1tian
Dr. DroUgh. Something just dawned on me, because while you're not wrong, the same actually applies here, with less margin for error (which is what causes the cards to burn anyways) - and that people may have started to skimp on the quality of other components, such as the power supply, to accommodate rising RAM prices.
I learned that lesson very, very early on before widespread testing of modern powersupplies, back in the day. Dont mess around and get a cheap powersupply. Look at the data. Just my 2 pennies
Posted on Reply
#78
Dr. Dro
dom1tianI learned that lesson very, very early on before widespread testing of modern powersupplies, back in the day. Dont mess around and get a cheap powersupply. Look at the data. Just my 2 pennies
Yup. I've insisted on purchasing a god tier power supply years ago, and I still use it today. Next full rebuild I'll have to retire it, sadly. It'll be coming up on 10 years old soon.
Posted on Reply
#79
steve1974
if any consumer product, these products would have been banned from sale until there was a fix what makes a gpu exempt from this ?
Posted on Reply
#80
AVATARAT
The main problem with this design is that too much heat is generated in the contact area, which is always a problem with this type of product. This is the reason for their failure, as copper degrades more quickly when exposed to heat.
That is why this standard is bad.

That is why some adapters from the secondary market have cooling, because there is an understanding of the problem.
That is why, when it has been turned on 30 times, it is recommended to replace it with a new one, because the surface gets scratched and starts to generate more heat.
Posted on Reply
#81
MondoMike
Darmok N JaladBut it looks like the splitter cable provided by Gigabyte.
Provided by the GPU manufacturer typically doesn't mean "more better" the issue is its a splitter with an unknown PSU in China, no doubt. Without the direct 12v-2x6 it is likely an outdated PSU non-atx 3.0 and probably lacks good fail safes and transient loads. Imagine spending $3-4k on a GPU but not $100-$200 for a newer and better PSU.
Posted on Reply
#82
Aleksandar_038
MondoMikeProvided by the GPU manufacturer typically doesn't mean "more better" the issue is its a splitter with an unknown PSU in China, no doubt. Without the direct 12v-2x6 it is likely an outdated PSU non-atx 3.0 and probably lacks good fail safes and transient loads. Imagine spending $3-4k on a GPU but not $100-$200 for a newer and better PSU.
The issue is with bad electrical design of 12 HPWR and that is it. When you design something, you define baseline and try to predict all usage conditions. On account of HPWR connector, both failed very miserably.
Posted on Reply
#83
MondoMike
Aleksandar_038The issue is with bad electrical design of 12 HPWR and that is it. When you design something, you define baseline and try to predict all usage conditions. On account of HPWR connector, both failed very miserably.
How many native ones have burned up? I feel like I've only ever seen adapters burned up. Didn't we do away with most adapters anyway? such as 2x SATA to 6/8pin or 2x 6 pin to 8 pin? I feel like adapter hate is more inline with the reality of it but maybe that's just me.
Posted on Reply
#84
Aleksandar_038
MondoMikeHow many native ones have burned up? I feel like I've only ever seen adapters burned up. Didn't we do away with most adapters anyway? such as 2x SATA to 6/8pin or 2x 6 pin to 8 pin? I feel like adapter hate is more inline with the reality of it but maybe that's just me.
Does not matter. Unless explicitly forbidden to be used with adapter, and warranty voided if used with adapter - it will be used with adapter. And when designing such connection, you must take into account that somebody will use adapter.

Basics of engineering.
Posted on Reply
#86
Darmok N Jalad
MondoMikeProvided by the GPU manufacturer typically doesn't mean "more better" the issue is its a splitter with an unknown PSU in China, no doubt. Without the direct 12v-2x6 it is likely an outdated PSU non-atx 3.0 and probably lacks good fail safes and transient loads. Imagine spending $3-4k on a GPU but not $100-$200 for a newer and better PSU.
Well, we don’t know that to be true. What if OP had a high-quality PSU from before this new standard? What if his PSU had the new cable but he elected to use the splitter in the box instead? What if he used a crappy 1000W PSU that came with the new cable? Those are all conjecture. If “don’t use old PSUs that lack this new cable” is the rule, then the manufacturer shouldn’t put a splitter in the box that allows someone to bypass said rule. We’re all fairly accustomed to looking at the PSU recommendation, and Gigabyte just plainly states that a 1000W PSU is recommended.
Posted on Reply
#88
OkieDan
CrackongThe connector I recommended is the XT90, 30mm long male plug, typical paired with 2x10AWG wires with a recommended bending distance of 10mm away from the connector and bending radius of 10mm. So it is 50mm in total.

12vhpwr had a 14 mm long male plug, paired with 12x16AWG wires and with a huge recommended bending distance of 35mm away from the connector and bending radius of 10mm, So it is 59mm in total.

If strict safety regulations are followed,
The much safer XT90 solution actually takes LESS vertical space than the 12vhpwr solution.

And the XT90 solution only needs 2 wires, that makes it infinitly more clean, tidy, and easy to manage.
10 Awg is not rated to handle enough current for 12v 600W.
Posted on Reply
#89
AVATARAT
Hecate91The included adapter should be sufficient for the GPU, but I find it interesting the goal posts have moved from only use the included adapter, to only use an ATX 3.0 PSU with 12v2x6.
And the "just plug it in all the way" excuses don't apply anymore either, since Dell has added their own retention mechanism to the connector.
videocardz.com/newz/dell-pc-teardown-shows-mechanically-locked-12v-2x6-gpu-power-connector
Yes, this is a smart choice on Dell's part, which only shows that there is a problem, and part of the problem is the "best contact surface," which will not move even a millimeter.
The only things missing from this locked connector are cooling and a temperature sensor, but in the end, they just want to eliminate "some user errors" rather than extend the life of the card too much.
Posted on Reply
#90
Sora
Aleksandar_038Basics of engineering.
of which you have no background in.
Posted on Reply
#91
Crackong
OkieDan10 Awg is not rated to handle enough current for 12v 600W.
Fair enough.
Then use 8 AWG
Posted on Reply
#92
Aleksandar_038
Soraof which you have no background in.
Butthurt too much?

Basics of engineering in any field - predict, prevent, overengineer for safety. None of those are fulfilled in this case, therefore, 12HPWR is simply JUNK.
Posted on Reply
#93
lexluthermiester
Aleksandar_038Basics of engineering in any field - predict, prevent, overengineer for safety. None of those are fulfilled in this case, therefore, 12HPWR is simply JUNK.
This! Well said.
Soraof which you have no background in.
I do. He's not wrong.
Posted on Reply
#94
Aleksandar_038
lexluthermiesterI do. He's not wrong.
I am not an electrical engineer... But I am an engineer, just in different field. And we all fight same battles in our fields, it is not all that different. Basics are always basics.
Posted on Reply
#95
Gucky
Crackongup-to-spec XT90 connectors are built for handling instant discharge of batteries. with very high C value.
XT90's limit is 72V 90A, a continuous 12V 50A is a piece of cake for XT90.
There is your thinking problem. INSTANT is not continuous. 12VHPWR do not burn instantly, they take hours to melt. Same can happen to the XT90.
And about the temperature, a GPU is a difference starting point compared to a battery. A GPU already is ~50-60°C (or more) to begin with. That heat spreads through the copper layers to the connector.

But yes using 2 cables instead of 12 WILL fix the power inbalance problem in the cables.
Posted on Reply
#96
Besensac
Rate my 12V cable bend 🤣
Using the seasonic psu cable so I guess it’s fine 🤔
Posted on Reply
#97
DudeBeFishing
GuckyThe specs vary widely... some say 40A continuous, some say 90A. Also 8AWG to 12AWG, which is also a big range. The 12VHPWR needs 50A btw.
And it also gets quite warm, warmer then the 12VHPWR (if it works how it supposed to).
The cables 8 and 10AWG, which is 5-7 times the thickness of the 16AWG the 12VHPWR uses. You can't bend 8-10AWG in tight spaces... There is a difference between several cables bundled togather and a single thick one...
Free hanging are rated as they're named. XT90 is 90A. The PCB mounted have lower current ratings. I think it's due to the single small PCB pins. More PCB pins will fix that, or use two XT60. The 10 awg wires on my RC car and battery backup are more flexible than most 12VHPWR wire bundles.

PCB mount:
XT30 - 15A - 180W @ 12V
XT60 - 30A - 360W @ 12V
XT90 - 40A - 480W @ 12V

Connecting six 12V wires to a single XT60 pin will be challenging cost wise. You'd probably need to solder them together to a single 10 awg wire, then that 10 awg wire to the XT60. You can't expect people to have a PSU with the new connector. That connector didn't exist when they bought their previous GPU.

Another option is using the existing 8-pin ESP connector, the one the CPU uses. I'm not sure what the actual rating for these are, but 300W shouldn't be an issue. There's varients with thicker pins that can handle it. It will be confusing for non tech savvy people when they can't put their 8-pin PCIe connector into the GPU's 8-pin ESP. People will try using PCIe to ESP adapters which may fail. There's the unbalanced load problem without individual wire current balancing.

Maybe Nvidia shouldn't design a +300W GPU with no room for wire connectors.
Posted on Reply
#98
Aleksandar_038
Still
DudeBeFishingMaybe Nvidia shouldn't design a +300W GPU with no room for wire connectors.
Generally way too much of our performance gains in laat decade came from increased watts. Both for CPU and GPU.

Back in the day, overclocking was the thing.

These days, it is proper underclocking so you can protect your expensive hardware and keep it running for longer.

So now I need two 360 AIO to cool high end CPU and GPU.

Connector issue is conseuqence, but it still couls have been prevented if someone was not cheaping out...
Posted on Reply
#99
AVATARAT
Aleksandar_038Back in the day, overclocking was the thing.
Overclocking is still a thing, but nowadays we just have to replace the crappy cable from time to time (sometime and connector :D ) :)
Posted on Reply
#100
Aleksandar_038
AVATARATOverclocking is still a thing, but nowadays we just have to replace the crappy cable from time to time (sometime and connector :D ) :)
:) Well, yes, it is, but now performance gains are minor and overclocking is more of competition and benchmark, and raising bar for yourself, than there is real need / gain in performance. CPUs and GPUs come clocked usually to the limits.

Back in the day, OC Celeron 300 to 450 MHz was huge :) Difference in performance was also day and night... Best of all, you needed just carefully to chose mobo and buy a bit better cooler. These days, it is much more expensive game :)
Posted on Reply
Add your own comment
Nov 15th, 2026 19:46 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts