Speculation surrounding the RTX 5090 has ramped up recently, and we’re starting to get more definitive insights into Nvidia’s highly anticipated RTX 50-series GPU. Renowned hardware insider Kopite7kimi reports that the RTX 5090 is expected to demand an impressive 600 watts of power.
The leaker disclosed specifications for both the RTX 5090 and RTX 5080 on X (formerly Twitter). While previous hints about RTX 50-series graphics cards were mostly vague, this marks the first time that complete details have been revealed. According to the information shared, the RTX 5090 will indeed require 600 watts and feature an astounding 32GB of GDDR7 memory, paired with a 512-bit bus.
GeForce RTX 5090
PG144/145-SKU30
GB202-300-A1
21760FP32
512-bit GDDR7 32G
600W— kopite7kimi (@kopite7kimi) September 26, 2024
Though whispers had circulated regarding the RTX 5090’s high power requirements, this is the first occasion that specific metrics have surfaced alongside other specifications. As a point of comparison, the RTX 4090 requires 450 watts. However, its 16-pin power connector can deliver up to 600 watts, suggesting that Nvidia may fully utilize this capacity with the next-generation RTX 5090. Interestingly, despite its demanding power needs, Kopite7kimi indicates that the RTX 5090 will utilize a dual-slot cooler design.
Details about the RTX 5080 were also shared by the leaker, stating that it would need 400 watts of power and feature 16GB of GDDR7 memory operating on a 256-bit bus. This represents only a slight upgrade from the RTX 4080 Super, hinting that Nvidia may choose to maintain a notable performance disparity between the RTX 5090 and RTX 5080, similar to the previous RTX 40-series lineup.
GeForce RTX 5080
PG144/147-SKU45
GB203-400-A1
10752FP32
256-bit GDDR7 16G
400W— kopite7kimi (@kopite7kimi) September 26, 2024
So far, Nvidia has not officially acknowledged the existence of the RTX 5090 or RTX 5080, although he have acquired substantial knowledge regarding the forthcoming Blackwell architecture, which the GPUs will employ. Nvidia has already unveiled this new architecture for data centers, although its enterprise version was postponed due to design issues.
As for when Nvidia will roll out its next-generation graphics cards to the gaming market, we typically expect announcements in October or November. However, delays in the enterprise sector could result in the launch being pushed back to 2025.