A recent video released by the notorious rumour channel Moore’s Law is Dead delves into the potential release date of the RTX 5090 and highlights an interesting factor that might influence Nvidia’s decision. According to a source claiming to be from inside Nvidia, the launch of their next-gen cards in 2024 will largely depend on the competitiveness of AMD’s RDNA 4 GPUs. If AMD’s next-gen cards prove to be a formidable contender in the mid-range market, Nvidia might feel compelled to respond with a new flagship GPU to maintain their dominance. However, if AMD’s RDNA 4 GPUs fail to have a significant impact on Nvidia’s sales, it is possible that Nvidia will hold off on releasing the RTX 5090 until 2025.
The upcoming consumer architecture from Nvidia is said to be based on the Blackwell GPU design. While the source did not provide detailed specifications, they mentioned that it is being prepared for a potential launch in Q4 2024. This indicates that Nvidia is actively working on the development of the RTX 5090, but its release is contingent upon various factors.
Two key factors will reportedly influence Nvidia’s decision regarding the release of the RTX 5090. Firstly, they will closely monitor the sales performance of the current RTX 40-series throughout the next year. If the sales remain strong, Nvidia might see no immediate need to introduce a new flagship GPU. Secondly, the state of AMD’s next-gen card launch will play a crucial role. If AMD’s RDNA 4 GPUs achieve a successful launch and start to impact Nvidia’s sales significantly, it might prompt Nvidia to accelerate the release of the RTX 5090 as a response.
While it is unlikely that AMD’s RDNA 4 GPUs will directly compete with Nvidia’s high-end offerings, a successful launch could potentially impact Nvidia’s volume sales. In such a scenario, Nvidia may feel the need to release the RTX 5090 to recapture market share and remind consumers of their upcoming GeForce generation. Therefore, those eagerly anticipating a next-gen Nvidia GPU in 2024 should keep an eye on the success of AMD’s RDNA 4 launch.
Alternatively, if Nvidia proceeds with their proposed launch of the RTX 40-series Super cards in early 2024, it could lead to decreased prices and improved performance. This would solidify the Ada generation as the go-to graphics cards for the following 12 months. In this scenario, it becomes difficult to comprehend why Nvidia would introduce the RTX 5090 at the end of the year, potentially cannibalizing the market for their own products.
Speaking about Nvidia’s future plans, the source suggests that the company intends to highlight the efficiency of the RTX 5000 at CES 2025. This indicates that Nvidia is currently planning to launch their next-gen GPUs by the start of 2025. While this may seem like a significant delay, it is earlier than what was initially speculated. Some feared that Nvidia might delay the release until 2025, leaving users to wait an extended period before experiencing the Blackwell architecture in desktop GPUs.
The key question that arises is how the Blackwell architecture will perform compared to its predecessors. According to the MLID quote, Blackwell as an architecture is not expected to deliver the same level of pure raster uplift observed when transitioning from the RTX 30-series to the RTX 40-series. However, it is important to note that the final specifications are not yet finalized, and there is a possibility that a full GB202 chip could provide a significant performance boost over the current RTX 4090. If this happens, it would create a similar level of anticipation and excitement as the jump from the RTX 3090 to the RTX 4090.
The potential release date of the Nvidia RTX 5090 hinges on various factors, including the sales performance of the RTX 40-series and the success of AMD’s RDNA 4 launch. While it is uncertain when exactly the RTX 5090 will arrive, Nvidia seems to be actively working on the Blackwell architecture and plans to make a significant announcement regarding RTX 5000 efficiency at CES 2025. Regardless of the release date, users can anticipate a leap in performance, although the magnitude will depend on the final specifications of the Blackwell architecture.