Nvidia’s Power Connector Melting Issue Persists Into RTX 50-Series Generation

Remember when the most expensive consumer graphics processing units in the world actually melted? Or, more accurately, when their high-voltage power connectors melted into blackened pieces of plastic and copper? For many PC hardware enthusiasts and builders, the memory remains remarkably fresh. Unfortunately, despite high hopes and hardware redesigns, this hazardous hardware phenomenon is far from a thing of the past.

Even though Nvidia’s latest RTX 50-series was widely expected to sidestep the notorious melting saga that plagued its predecessors, reality has proven far more stubborn. Within weeks of the hardware hitting store shelves and review benches, striking photographs of scorched RTX 5090 plugs began circulating across online tech communities, enthusiast forums, and social media platforms. Sadly, these unsettling hardware failures continue to surface, raising serious questions about the ongoing reliability of high-wattage computing power delivery.

Nvidia’s Fix Solved a Problem, But Not the Only Problem

Thinking back over the hardware landscape, it is hard to believe that roughly four years have passed since the modern "melting connector saga" first shocked the computing world. The initial high-profile victims were Nvidia’s flagship RTX 4090 cards, which, at the time of their release, stood comfortably at the absolute pinnacle of consumer graphics hardware.

When consumers invest significant sums—such as the lofty $1,600 baseline price tag of the original RTX 4090—they naturally expect enterprise-grade reliability and industrial safety. Buyers place immense trust in flagship hardware to perform flawlessly under any circumstances, assuming that high-end engineering can comfortably handle rigorous workloads. Yet, within a remarkably short window following launch, reports of catastrophic failures began cropping up rapidly across the globe. Disturbingly, these graphics cards seemed to fail even when they were not subjected to particularly heavy or sustained computing workloads. The thermal damage typically manifested both on the graphics card side of the power receptacle and on the modular power supply unit side, instantly triggering lengthy Return Merchandise Authorization processes and forcing hardware manufacturers to issue replacements.

Nvidia launched a thorough internal investigation into the widespread hardware failures. Rather than pointing the finger directly at the underlying hardware design, the corporation focused its preliminary findings on user error—specifically, the physical method by which consumers plugged these massive graphics cards into their power supplies. The logic seemed sound on paper: extensive testing revealed that when the proprietary high-voltage connector was not pushed all the way into the receptacle, high electrical resistance created localized heat and a legitimate fire hazard. Bending the heavy power cables too sharply near the connection point was similarly identified as a major design taboo. At the time, Nvidia indicated it was aware of only about fifty confirmed cases worldwide, representing a microscopic margin of overall sales. Independent hardware analysis groups like Gamers Nexus backed that general assessment, estimating the overall hardware failure rate at a remarkably low 0.04 percent. Meanwhile, the manufacturer worked to streamline and fast-track the RMA process for affected consumers.

The formal follow-up to the saga eventually arrived with the commercial introduction of the updated 12V-2×6 connector standard. This revised power delivery design featured physical modifications, including sense pins designed to be 1.5 millimeters shorter and power terminals lengthened by 0.25 millimeters compared to the original, highly problematic 12VHPWR specification. The engineering update was intended to comprehensively solve the systemic reliability problem. The logic dictated that if the physical plug was not completely inserted into the socket, the shortened sense pins would fail to make contact at all, preventing the graphics card from drawing any power through a poorly seated connection.

Nvidia expressed immense confidence that the 12V-2×6 standard would permanently answer the industry’s prayers regarding high-wattage power delivery. Corporate representatives publicly stated that they did not expect the powerhouse RTX 5090 to suffer from the same cable melting issues that blighted the previous generation, despite the new flagship GPU drawing a massive 575 watts of electrical power under load.

The underlying trouble, however, lies in a fundamental limitation of the physical design. Those shorter sense pins can successfully confirm whether a plug is fully inserted into the socket, but they are entirely incapable of monitoring whether electrical current is being distributed evenly across the internal array of twelve individual pins. Consequently, a graphics card equipped with a properly seated plug will continue drawing immense amounts of raw electrical power even if a single internal wire ends up shouldering the vast majority of the load. That exact technical oversight brings the industry directly to the ongoing issues seen today.

The Real Culprit is Where the Current Goes After You Plug It In

To truly understand why the thermal melting episodes continue to plague modern systems even after the introduction of a redesigned connector, it helps to examine the underlying electrical engineering. The issue comes down to what happens to the electrical current once it successfully reaches the internal circuitry of the graphics card, a design evolution that can be traced back through successive generations of hardware.

During the era of the RTX 30-series, Nvidia treated the multi-pin power connector as three distinct and separate 12V electrical sources. Each individual source was equipped with its own dedicated shunt resistor. Those specialized shunt resistors allowed the intelligent graphics card to actively monitor and balance its electrical load across each pair of wires, ensuring that power distribution remained stable and safe during operation.

That crucial design methodology changed dramatically with the launch of the RTX 40-series and carried forward into the current RTX 50-series hardware generation. In modern high-end cards, all six of the individual 12V power lines are lumped together into a single unified electrical input. As a result, the graphics card is only made aware of the total overall electrical draw, remaining completely blind to imbalances occurring across individual wires. This lack of granular monitoring creates significant vulnerabilities because tiny, unavoidable variations in contact resistance between individual pins are entirely normal in mass-manufactured electrical connectors. Because electricity naturally follows the path of least resistance, an imbalance can easily cause a disproportionate amount of current to surge through a single wire.

This exact discrepancy has been repeatedly highlighted in real-world testing by hardware analysts. Well-known tech creator der8auer tested a reference design RTX 5090 and discovered through thermal imaging and current clamps that two of the six individual wires were carrying over 22 amperes of electrical current, while the remaining wires sat at a much cooler 2 amperes. While the majority of the wires remained comfortable, the physical connector itself did not: on the power supply side of the connection, local temperatures skyrocketed to an alarming 150 degrees Celsius.

Context remains vital when evaluating these findings, as each individual pin is officially specified to handle approximately 9.5 amperes safely. The root cause of the hardware failure becomes glaringly obvious in light of such extreme imbalances. While it is worth noting that identical extreme imbalances have not been successfully replicated by every single hardware reviewer, the mere fact that such severe discrepancies appeared at all during standard testing remains deeply alarming to industry observers.

2026 Has Made the Problem Impossible to Wave Away

At a time when high-end graphics hardware pricing has reached unprecedented heights, consumers naturally expect their expensive components to offer absolute stability and safety. Yet the persistent connector melting problem has stubbornly refused to fade away, even as broader market conditions have grown increasingly difficult over recent years.

Professional hardware reviewers continue to report troubling instances of thermal degradation with the RTX 5090, a flagship component commanding astronomical market prices ranging from $5,000 to $7,500. Prominent hardware outlet Club386 suffered a catastrophic failure when an RTX 5090 Founders Edition and a high-end power supply were permanently damaged by a melted power cable. Nvidia subsequently replaced the ruined graphics card under warranty.

Meanwhile, independent testing channels like Hardware Unboxed caught a power cable on a high-end ROG Astral RTX 5090 test bench reaching approximately 175 degrees Celsius. Curiously, the underlying mechanics of the failure mirrored the earlier findings published by der8auer, with a single internal wire pulling roughly 22 amperes while the companion wires remained largely inactive.

Adding fuel to the fire, the recent arrival of DLSS 5 has brought new variables into the power equation. While public discussion often centers on debates surrounding artificial intelligence features and rendering technologies, many observers overlook the immediate physical consequence: a dramatic increase in overall power draw. Industry outlet TweakWatch documented a significant power spike, noting that system power consumption jumped from an already substantial 465 watts up to a maximum peak of 575 watts when running DLSS 5 workloads in modern titles like NBA 2K27. Shortly after those performance tests were published, an online hardware reader reported that their RTX 5090 power connector had thermally melted while evaluating the new software features.

As much as the broader computing community wishes that melting power connectors were firmly a thing of the past, the reality on the ground proves otherwise. While such failures remain statistically rare overall, they are far from impossible, and given the staggering cost of hardware like the RTX 5090, no user wants to risk destroying their investment.

Current best practices for mitigating risk remain consistent with advice from previous hardware generations. Users must ensure that power plugs are firmly and completely inserted into their sockets, leaving adequate straight cable length before any tight bends occur near the rear panel. Checking the physical connection periodically after initial gaming sessions can help ensure stability. Many builders also choose to avoid third-party extension cables or complex angled adapters that can introduce additional points of electrical resistance. Finally, undervolting the graphics card remains a popular strategy among enthusiasts seeking to optimize power delivery, reduce thermal output, and potentially enhance long-term hardware longevity.