NVIDIA’s latest GPUs are melting their power connectors, and I’m frankly astounded that we’re still dealing with this issue. After watching Linus and his team perform surgery on an RTX 5090 to replace the problematic connector with something more robust, I’m convinced that NVIDIA’s approach to power delivery is fundamentally flawed.
The story begins with NVIDIA adopting not one, but three different GPU power connectors over their last three generations. The RTX 4090 used the 12VHPWR connector, which had well-documented melting issues. Then came the RTX 5090 with the supposedly improved 12V-2×6 connector from the ATX 3.1 spec—and yet, the melting problems persist.
This is unacceptable for products that cost thousands of dollars.
The Root of the Problem
CT scans of the RTX 5090 reveal the fundamental design flaw: despite having 12 pins (six for 12V and six for ground), they all connect to a single metal strip that attaches to the board at just one point. This means the GPU has no way of knowing if one or more pins has a poor connection, forcing the remaining pins to handle more current than they should.
When a connector is improperly seated or when using a poorly made adapter in a cramped case, some pins may lose contact. The GPU continues drawing the same power, but now through fewer pins—a recipe for overheating.
What’s particularly frustrating is that NVIDIA didn’t even design these problematic connectors. The 12VHPWR was part of the ATX 3.0 spec, and the 12V-2×6 is part of ATX 3.1. But NVIDIA is taking the heat because they’re the only GPU manufacturer forcing adoption of these standards.
A Better Solution Exists
The makeshift solution demonstrated—replacing the stock connector with a robust RC battery connector rated for 60 amps continuous—proves that better options exist. This connector provides:
- Higher continuous power rating (720 watts)
- More secure connection that won’t get bumped loose
- Better physical feedback when properly connected
The modified card worked perfectly in testing, maintaining normal temperatures and performance while drawing 575 watts during benchmarks.
Why Doesn’t NVIDIA Just Use Better Connectors?
The answer, as usual, comes down to cost. The difference between the connectors themselves might be pennies, but the real expense is in the wiring. Using twelve 16-gauge wires (the current approach) is cheaper than two thick 8-gauge conductors when multiplied across millions of systems.
Thinner wires are also more flexible for cable management and easier to work with during manufacturing. But these cost savings come at the expense of reliability and safety.
Not All Cards Are Created Equal
Interestingly, not all RTX 5090s share this vulnerability. ASUS has implemented per-pin monitoring in their cards and even created an in-game overlay to show that everything is working properly. This proves that the problem can be addressed with better engineering.
NVIDIA needs to take responsibility for this issue. Whether through a recall, a redesign, or some other solution, they can’t continue selling products with a known safety risk.
For now, if you own a Founders Edition card, be extremely careful with your power connector. Ensure it’s properly seated, avoid tight bends near the connector, and consider using a case with ample clearance around the GPU.
The gaming community deserves better than melting connectors on premium hardware. It’s time for NVIDIA to prioritize safety over saving a few dollars on wire gauge.
Frequently Asked Questions
Q: Are all RTX 5090 cards affected by the melting connector issue?
No, not all cards have the same vulnerability. Some manufacturers like ASUS have implemented additional safety features such as per-pin monitoring that can detect connection problems before they cause overheating.
Q: What causes the connectors to melt in the first place?
The melting occurs when some pins lose contact due to improper seating or physical stress on the connector. Since all pins connect to a single point on the board, the remaining pins must carry more current than designed, causing them to overheat.
Q: Can I prevent connector melting on my RTX 5090?
You can reduce the risk by ensuring the connector is fully seated, avoiding tight bends in the power cable near the connector, using only high-quality cables/adapters, and making sure your case provides adequate clearance around the GPU power connection.
Q: Why did NVIDIA choose these problematic connectors?
While NVIDIA didn’t design the connectors (they’re part of the ATX 3.0/3.1 specs), they chose to adopt them primarily for cost reasons. Using multiple smaller wires is cheaper than fewer thick wires when manufacturing at scale, and the thinner wires offer better flexibility for cable management.
Q: Is NVIDIA planning to address this issue?
As of now, NVIDIA hasn’t announced an official solution to the melting connector problem on the RTX 5090. Given the potential safety concerns, they will likely need to take some action, whether through a design revision, recall, or other remediation program.
Finn is an expert news reporter at DevX. He writes on what top experts are saying.
