With the release of the RTX 50 series cards on the horizon, the very first image of the RTX 5090 graphics card has surfaced online.

The graphics card used to light up the monitor in the picture is the RTX 5090. Hong Kong-based graphics card manufacturer, Pannon Group, is setting up a factory in Indonesia, where the image captures the production line being commissioned.

Though the image quality is somewhat blurry, the substantial size of the RTX 5090 is apparent—and it appears to follow a triple-slot design. It seems that AIB partners are currently in the trial testing phase, post which full-scale production will likely ensue.

According to previous insights shared by related bloggers, the GeForce RTX 5090 GPU's cost won't significantly exceed that of the RTX 4090, which was initially priced at around $20,000.

Previously revealed specifications indicate that the RTX 5090 will boast a 14-layer PCB design, while its flagship chip, tagged GB202-300-A1, features up to 21,760 CUDA cores, marking a 33% increase compared to the RTX 4090.

The memory has seen an upgrade to 512-bit 32GB GDDR7, translating to a 50% enhancement in both bit width and capacity. However, the exact core and memory frequencies are yet to be unveiled.

The TGP power consumption is expected to hit 600W—a 33% increment—yet it will remain air-cooled.

At least for the meantime, the power supply connector continues to be a single 12V-2x6 16-pin, though rumors suggest it might eventually shift to using two connectors, potentially allowing a theoretical output power of up to 1200W.