Just when we thought we’d seen the worst of high-powered GPU issues, the NVIDIA GeForce RTX 5090 has reportedly struck again. A recent post on the NVIDIA subreddit by a user named Roachard shows photos of a melted power connector and a completely fried power supply. This isn’t just a case of a damaged cable, this time, the entire PSU went down with it.
The build in question was packing some serious hardware: an AMD Ryzen 9 9800X3D processor, MSI’s beefy RTX 5090 Gaming Trio OC, and a Corsair SF1000L SFX power supply. Despite being fully aware of previous incidents where 5090s melted third-party cables, Roachard played it safe and used the stock NVIDIA-supplied power connector. Unfortunately, even that wasn’t enough to avoid disaster.
This isn’t the first time we’ve seen reports of cable and connector issues with NVIDIA’s latest-gen GPUs. The RTX 4090 had similar drama shortly after launch, with the new 12VHPWR connector (also known as the 16-pin PCIe 5.0 power connector) being blamed for melted plugs and even minor fires in some extreme cases. NVIDIA later released updated guidelines, and many assumed the problems were ironed out with the 5090 series. But clearly, the issue hasn’t gone away entirely.
Related articles you may find interesting
What makes this case more alarming is that it’s not just a cable this time, the whole PSU was apparently killed. That’s not just frustrating, it’s a costly and potentially dangerous failure. Given the 5090’s massive 600W power draw, thermal management and connector quality are more critical than ever. But even high-end setups with proper airflow and official cables aren’t immune, it seems.
So what’s going on? There’s still no official word from NVIDIA or board partners about widespread issues with the 5090 connectors, but incidents like these raise concerns. Some suspect that the 12VHPWR connector design might still be prone to poor seating, especially in compact cases where cables are bent tightly. Others wonder whether some of the high-wattage spikes during heavy workloads are simply too much for some PSUs to handle, especially in SFX form factors.
Until we get clarity, the best advice is to double-check your power connections, avoid excessive bending near the connector, and consider using angled adapters or cable extensions to relieve stress on the plug. And most importantly, keep an eye on your system temps and power behavior, especially if you’re running a 5090 or any GPU with similar draw.
Let’s hope this is just an isolated case, but if more users start experiencing similar failures, NVIDIA might be in for another round of PR damage control.
