Nvidia reportedly discontinues RTX 5090. The RTX 5080 with 24 GB is set to become the 2027 flagship

Why is Nvidia taking the RTX 5090 away from gamers?

The whole thing started with a post that MEGAsizeGPU published on X on October 10. According to him, all future shipments of GB202 chips will be reserved for professional RTX Pro cards. Benchlife elaborated on the information: card manufacturers are reportedly no longer receiving chips for the RTX 5090, nor for the Chinese variants RTX 5090 D and RTX 5090 D V2. According to the server, Nvidia does not consider this the end of the product’s lifespan, but merely shifting production capacity to where it can earn more.

The GB202 chip doesn’t just power the gaming RTX 5090 with 32 GB of memory. The same base is used for professional RTX Pro 6000 with 96 GB, RTX Pro 5500 with 84 GB, and RTX Pro 5000 with 48 or 72 GB of memory, which sell for many times higher prices. Furthermore, companies developing artificial intelligence have been buying up RTX 5090s in recent months as a cheaper alternative to professional cards. At a time when memory is scarce and demand for AI chips continues to grow, the choice for Nvidia is relatively simple.

Once manufacturers sell out their stock, new RTX 5090 units will reportedly no longer be produced, so prices are likely to rise even further. The card already costs a multiple of its original price. Upon its release in January 2025, Nvidia recommended a price of approximately 60,000 Czech crowns, but today, only two desktop models are available on Heureka, with the cheaper one costing around 170,000 Czech crowns.

Nvidia has not yet officially confirmed the end of production. Wccftech admits that this might only be a temporary redirection of chips, and the company will return to gamers once the memory market stabilizes.

What will the RTX 5080 with 24 GB offer?

According to Benchlife, Nvidia plans to introduce the RTX 5080 with 24 GB of GDDR7 memory in the first quarter of 2027, replacing the current 16 GB version. However, card manufacturers only know about it from verbal communication so far. They reportedly haven’t received the documentation for developing their own models yet, only a promise that it will arrive soon. Thus, the timeline is far from certain.

Technically, a simple solution is offered. Instead of eight 2GB memory chips, the card will receive the same number of 3GB modules, which will increase the memory by half and leave the 256-bit bus unchanged. This is exactly what leaks surrounding the RTX 5080 SUPER, which we wrote about in June, promised. However, the SUPER refresh reportedly stalled in the summer due to expensive memory. Whether the new RTX 5080 is just a rebranded SUPER or a different product is currently unknown. Leaks do not mention frequencies or power consumption.

Will it catch up to the RTX 5090 at least a little?

The memory alone will not make the RTX 5080 a successor to the RTX 5090, and the outgoing flagship will remain significantly ahead in performance. However, the higher capacity will help in 4K resolution and in games with demanding ray tracing or path tracing, which can easily fill up graphics memory.

The new product won’t be cheap either. The current RTX 5080 sells in the Czech Republic for approximately 37,000 Czech crowns, and 3GB memory modules are more expensive and harder to find. Wccftech therefore expects that the price of the new version will be around 2,000 dollars, which translates to over 40,000 Czech crowns before tax.

The question remains how long the new product will stay at the top. According to earlier leaks, Nvidia is targeting the first half of 2027 for the new RTX 60 generation, so the RTX 5080 might only hold the position of the most powerful GeForce for a few months.

Would you buy the RTX 5080 with 24 GB, or would you rather wait for the RTX 60?

Sources: MEGAsizeGPU/X, Benchlife, Wccftech, Notebookcheck