This can happen to anyone, even with the safest settings
In the quiet routine of a Monday gaming session, a user discovered that even the most deliberate caution — undervolting, liquid cooling, premium components — cannot always shield us from the hidden fragilities within our machines. A melted 16-pin power connector on an RTX 5090 setup, untouched by excess or recklessness, reminds us that the boundary between reliability and catastrophic failure is thinner than the cables we trust to carry our power. MSI's quiet pivot to a revised PSU model with added safeguards speaks to an acknowledgment that the original design carried a vulnerability the market had not yet named.
- A black screen with sound — the ghost of a working system — was the only warning before the user found plastic connectors fused and destroyed by heat.
- The detail that unsettled the enthusiast community most was the undervolting: a conservative, safety-first choice that should have reduced risk, yet the cables burned anyway.
- The GPU socket survived intact, leaving open the possibility that the card itself could be saved, but the power supply and cable were beyond recovery.
- MSI moved quickly to replace the damaged PSU under warranty, offering the newer Ai1600TS model — a unit that ships with a physical GPU Safeguard feature the original lacked.
- The community is now asking whether this is an isolated manufacturing defect or a systemic vulnerability lurking inside the RTX 5090's power delivery ecosystem.
A Reddit user sat down to game on a Monday and found only a black screen — audio intact, display dead. He restarted, played briefly, and the next day the problem returned. When he finally opened the case and looked, he found his 16-pin power cable had not merely scorched — it had melted. The connector on the GPU end was destroyed; the one on the PSU end was worse. He had owned the RTX 5090 Founders Edition for over a year, running it with a custom liquid cooling loop and, crucially, undervolted — a technique that reduces voltage and heat output, the conservative choice of someone trying to protect their hardware.
The detail that caught the community's attention was precisely that undervolting. If anything, it should have made the system safer. Yet the cables had still burned. The GPU socket itself survived intact, which meant the card might be salvageable, but the MSI MEG Ai1600T power supply and cable were finished.
MSI acknowledged the failure and offered a warranty replacement — not with the same unit, but with the newer Ai1600TS, a model that includes a physical protection feature branded as GPU Safeguard. The existence of that feature in the revised model implied the company had already recognized a vulnerability in the original design. Whether the root cause was a manufacturing defect, a compatibility issue, or something else remained unanswered.
The user closed his post with a simple warning to the community: be careful, because this can happen to anyone. For owners of high-end hardware who believe premium components and cautious tuning are enough to guarantee safety, the incident was a sobering reminder that some failures arrive without warning signs, strange smells, or any signal at all.
A Reddit user sat down to play a game on Monday and got nothing but a black screen. The sound came through fine—he could hear the music, the effects, everything—but the display was dead. His first instinct was to blame the RAM, maybe an overclock gone wrong. He restarted the machine. The game booted. He played for a while, then shut it down. The next day, he tried again. Same problem. Black screen. This time, he decided to look at the hardware itself.
What he found stopped him cold. The 16-pin power cable connecting his RTX 5090 Founders Edition to his MSI MEG Ai1600T power supply had melted. Not partially scorched. Melted. The plastic connector on the GPU end was destroyed. The connector on the power supply end was worse. He'd owned this graphics card for over a year. It had a custom liquid cooling loop mounted on it. There had been no warning signs, no strange smells, nothing that suggested the power delivery system was about to fail catastrophically.
The user posted about it on Reddit on May 18th, and the detail that made other enthusiasts take notice was this: he'd been running the card undervolted. Undervolting is a technique where you reduce the electrical voltage feeding a chip, which lowers both power consumption and heat output. It's a conservative approach, the opposite of pushing hardware to its limits. If anything, it should have made the system safer, not more vulnerable. Yet the cables had still burned.
The melted connectors raised an immediate question about whether the problem lay with the graphics card, the power supply, or the cable itself. The user's GPU socket—the actual connector on the card where the cable plugged in—had survived intact. That was fortunate. It meant the card itself might still be salvageable. But the power supply and cable were clearly done.
MSI acknowledged the failure and initiated a warranty replacement for the damaged power supply. More tellingly, the company offered to swap the unit for the newer Ai1600TS model, a variant that includes additional physical safeguards branded as GPU Safeguard. The existence of these new protections suggested that MSI had recognized a vulnerability in the original design and moved to address it. Whether the problem was a manufacturing defect, a compatibility issue between the PSU and this particular GPU, or something else entirely, remained unclear.
The user's final message to the community was a warning: "Be careful and protect your equipment. This can happen to anyone." For owners of high-end graphics cards and premium power supplies, the incident served as a reminder that even expensive, well-regarded hardware can fail in ways that are difficult to predict or prevent. The question now was whether this was an isolated case or a sign of a broader reliability issue in the RTX 5090's power delivery ecosystem.
Citas Notables
I thought maybe it was a processor or RAM overclock issue, but no—it was the cable and power supply burning— Reddit user reporting the incident
Be careful and protect your equipment. This can happen to anyone.— Same user, closing statement
La Conversación del Hearth Otra perspectiva de la historia
Why would undervolting make this worse, not better? Shouldn't lower voltage mean less stress on the cables?
That's the unsettling part. Undervolting should theoretically reduce strain. The fact that it happened anyway suggests the problem wasn't about pushing the card too hard—it was something structural. Either the cable itself was defective, or there was a compatibility issue between the PSU and the card that no amount of conservative settings could prevent.
Did the user have any warning before the cables melted? Any burning smell, any visible damage?
Nothing. He only discovered it when the black screen problem forced him to physically inspect the connectors. A year of use with no red flags, then sudden catastrophic failure. That's what made people nervous—there was no way to see it coming.
The GPU socket survived. Does that mean the card is still usable?
Possibly, yes. The actual connection point on the graphics card wasn't damaged, which is why he had hope. But the power supply and cable were destroyed, and he'd need to replace both before he could even test whether the card still works.
Why did MSI offer the newer Ai1600TS model instead of just replacing the damaged one with an identical unit?
Because the newer model has physical safeguards—GPU Safeguard—that the original didn't. By offering the upgrade, MSI was essentially admitting they'd identified a design weakness and fixed it. That's not something a company usually does unless they've seen the problem before.
What does this mean for other RTX 5090 owners?
It means they should pay attention. This might be rare, or it might be a sign of a larger batch issue. Either way, the user's warning—check your cables, know what to look for—became the most practical advice anyone could give.