NVIDIA RTX 5090 Production May Be Winding Down as RTX 5080 24GB Rumors Gain Momentum
NVIDIA’s GeForce RTX 50 series could be heading for a major shake-up, according to new industry rumors. The latest reports suggest that NVIDIA may be preparing to end or sharply reduce production of the GeForce RTX 5090, its current flagship gaming graphics card, while shifting more of its high-end Blackwell GB202 GPU supply toward professional RTX Pro products.
At the same time, a new GeForce RTX 5080 24GB model is reportedly being prepared as a more gamer-focused alternative for enthusiasts who want more video memory but cannot justify the extremely high prices now attached to the RTX 5090.
The GeForce RTX 5090 launched in January 2025 as NVIDIA’s fastest gaming GPU, featuring 32GB of GDDR7 memory and the powerful GB202 Blackwell graphics chip. Since launch, there have been several claims that the card could face production cuts or even be discontinued earlier than expected. Until now, however, those reports have remained unconfirmed.
The newest rumors claim that NVIDIA may soon stop producing the RTX 5090 in meaningful volume. Instead, the GB202 GPU used inside the RTX 5090 would reportedly be redirected almost entirely toward professional workstation and AI-focused graphics cards.
That would be a major move, but it is not difficult to understand why NVIDIA might consider it.
The GB202 chip is the flagship GPU in NVIDIA’s Blackwell family. In the consumer gaming market, it powers the GeForce RTX 5090 32GB, currently one of the most powerful gaming graphics cards available. In the professional market, the same chip is used across multiple RTX Pro models, including high-memory cards such as the RTX Pro 6000, RTX Pro 5500, and RTX Pro 5000.
These professional GPUs are in high demand because of AI workloads, workstation use, content creation, simulation, rendering, and enterprise computing. With AI demand continuing to surge, NVIDIA may be able to generate more revenue by prioritizing GB202 chips for professional cards rather than gaming products.
Another major factor is memory pricing. GDDR7 and high-capacity DRAM remain expensive, and rising memory costs are putting pressure on both graphics card makers and buyers. The RTX 5090 has already reached extreme retail pricing in many markets, with some models reportedly selling for more than $6,000. At that point, the card is no longer priced like a traditional gaming GPU and has moved into professional workstation territory.
This price jump has also created an unusual situation. Some AI users have been buying GeForce RTX 5090 cards instead of professional RTX Pro models because the gaming card can offer strong compute performance and 32GB of VRAM at a lower price than certain workstation alternatives. In some regions, RTX 5090 cards have reportedly been modified with custom cooling solutions and redesigned boards to make them more suitable for AI workstations and servers.
NVIDIA does offer region-specific and export-compliant professional and gaming variants, but those models may include reduced AI performance, locked features, or lower memory configurations. Because of that, full-performance RTX 5090 cards have remained attractive to users outside the gaming market.
For gamers, the potential loss of RTX 5090 supply would be frustrating. The card already suffers from high prices and limited availability in many areas. If production is reduced or redirected, prices could stay inflated or climb even higher.
However, NVIDIA may have another option ready: the GeForce RTX 5080 24GB.
According to current rumors, NVIDIA is preparing a 24GB version of the RTX 5080 aimed at enthusiast gamers. This card could offer a larger VRAM buffer than the existing RTX 5080 16GB while avoiding the extreme cost of the RTX 5090.
The current RTX 5080 uses 16GB of GDDR7 memory on a 256-bit memory bus. The rumored 24GB version is expected to use higher-density 3GB GDDR7 memory modules instead of the 2GB modules found on the standard model. That would allow NVIDIA to increase VRAM capacity to 24GB while keeping the same 256-bit bus.
Possible RTX 5080 24GB specifications include:
GPU: GB203-based graphics processor
Memory: 24GB GDDR7
Memory bus: 256-bit
Memory speed: potentially around 32Gbps
Memory bandwidth: up to roughly 1TB/s
Power target: possibly higher than the standard RTX 5080
Performance: expected to sit well below the RTX 5090 but above the current RTX 5080 in memory-heavy workloads
The rumored card may resemble earlier reports of an RTX 5080 Super-style model. While it would not match the RTX 5090 in raw performance, the extra VRAM could make it more appealing for 4K gaming, high-resolution texture packs, ray tracing, AI-assisted creative tools, game development, 3D rendering, and productivity workloads.
Pricing will be critical. The standard RTX 5080 launched with a $999 suggested price, but real-world retail pricing has often been much higher, commonly landing between $1,499 and $1,999 depending on region, model, and supply. If NVIDIA releases an RTX 5080 24GB, it could realistically arrive in the $1,999 to $2,999 range if market conditions remain unchanged.
That would still be expensive, but it may look reasonable compared with the RTX 5090’s inflated pricing. For many buyers, an RTX 5080 24GB could become the more practical enthusiast-class GPU, especially if they need more than 16GB of VRAM but do not want to pay workstation-level prices.
Still, nothing is official yet. NVIDIA has not confirmed the end of RTX 5090 production, nor has it announced an RTX 5080 24GB model. For now, these details should be treated as industry rumors rather than confirmed product plans.
Even so, the situation highlights the growing tension in the graphics card market. Gaming GPUs are no longer competing only for gamers. The same high-end chips and memory stacks are increasingly valuable for AI, professional visualization, machine learning, rendering, and workstation deployments. That puts consumer graphics cards in direct competition with enterprise demand.
If NVIDIA does shift more GB202 supply toward RTX Pro cards, it could mark a turning point for the GeForce RTX 50 lineup. The RTX 5090 may become even harder to find, while the RTX 5080 24GB could step in as the new high-end option for gamers who want strong performance and more memory without entering full professional GPU pricing.
For now, PC gamers and hardware enthusiasts will be watching closely. The next few months could determine whether the RTX 5090 remains NVIDIA’s dominant gaming flagship or becomes a rare, ultra-premium product overshadowed by professional AI demand.
