Nvidia RTX 50 Super GPUs delayed by prohibitive GDDR7 memory pricing
Nvidia has reportedly delayed its RTX 50 Super GPUs because 3GB GDDR7 memory modules cost up to three times more than 2GB variants, disrupting board partner supply chains.
Nvidia's upcoming Super refresh of the RTX 50-series graphics cards is reportedly on hold due to soaring costs for 3GB GDDR7 memory chips, according to Tom's Hardware citing a VideoCardz source. One board partner already has the hardware on hand, but the company has allegedly been told that the launch is delayed because of prohibitive memory pricing. This situation threatens to push retail prices far above Nvidia's targeted MSRP, or force board manufacturers to sell units at a loss if they stick to set price points.
The rumored RTX 50 Super lineup relies on 3GB GDDR7 chips to provide 50 percent more capacity than the 2GB modules used in current-generation cards, avoiding changes to memory bus configurations. Models expected in the pipeline include the RTX 5080 Super and RTX 5070 Ti Super with 24GB of VRAM across a 256-bit bus, alongside the RTX 5070 Super featuring 18GB of VRAM on a 192-bit bus. However, these 3GB components cost twice or three times as much as 2GB variants, creating severe economic hurdles for the deployment of these configurations.
This component pricing crisis coincides with a structural shift in Nvidia's supply chain operations. According to Tom's Hardware, Nvidia altered its policy in late 2025 as a memory chip crisis unfolded, ending its practice of supplying VRAM chips alongside GPU dies to board partners. Consequently, the companies completing final graphics card assembly must independently source their own memory chips within a fiercely competitive market.
The broader memory market faces severe constraints that complicate hardware deployment strategies for infrastructure and consumer hardware alike. SK hynix, one of the major memory manufacturers, has indicated that the memory shortage will persist until 2030, projecting that 2027 will represent the worst year of the crunch. These supply chain dynamics have already impacted wider product roadmaps, such as Nvidia skipping new GPU announcements at CES 2026 for the first time in five years.
With memory costs surging significantly and accounting for a large fraction of the bill of materials in modern systems, the economics of silicon deployment are under intense pressure. While board partners already possess GPU dies, the timeline for resolving the GDDR7 pricing standoff remains uncertain. Industry watchers will need to monitor how Nvidia and its partners navigate these component constraints to maintain cost-effectiveness without abandoning their planned product cadences.