According to the latest leaks, NVIDIA plans to limit the launch of its new GeForce RTX 50 series to its flagship model, the RTX 5090. The company would therefore postpone the release of the other models to 2025.
NVIDIA is preparing to once again revolutionize the graphics card market with the imminent launch of RTX 5000. After introducing the RTX 4000 series, all eyes are on the next generation which is codenamed “Blackwell“.
The flagship model, the RTX 5090 will surely be equipped with the GB202 GPU, an advanced graphics processor that promises marked improvement performance for games and professional applications. Although the expected gains are not as spectacular as in previous transitions, they still mark an important development in terms of speed and of graphics capacity.
Only the RTX 5090 will be released at the end of 2024
According to information from the Moore’s Law Is Dead YouTube channel, NVIDIA plans to launch the new cards RTX 5090 Blackwell through the end of 2024, strategically positioned to impact sales during the holiday season. This timing is also chosen to face competition from the latest RDNA 4 technology, developed by AMD. Details on this series will be further highlighted at CES 2025, where the company will focus its presentation on energy efficiency and the performance of these new models.
Even if NVIDIA limits itself to the launch of the RTX 5090, we can expect a good overview of the new architecture “Blackwell.” This presentation should cover the new features and give an idea of the performance of the other models of the series when they are launched in 2025.
NVIDIA’s decision to proceed with a more conservative launch for the GeForce “Blackwell” series appears to be dictated by the need to manage inventory of the RTX 40 series and prioritize GPUs dedicated to AI, which are more profitable for the company. These factors influence the go-to-market strategy, balancing production and demand. This targeted launch of the RTX 5090 should allow the company to test the reception of its new technology before deploying the entire series.