Nvidia’s upcoming GeForce RTX 50 SUPER refresh has reportedly reached final production readiness, with at least one board partner already holding retail units. The launch, however, is being held back by component costs rather than engineering delays. Reports indicate that Nvidia has asked its partners to pause the rollout, citing the persistently high price of high-density 3 GB GDDR7 memory modules as the primary obstacle.
Pricing Pressure Delays Finished Hardware
The core issue revolves around the 3 GB GDDR7 packages needed to deliver the memory configurations expected across the SUPER lineup. While the cards are described as ready to ship, add-in board manufacturers are effectively unable to proceed until the cost of these memory modules falls to a more viable level. Nvidia has not yet publicly confirmed the RTX 50 SUPER series.
A Closer Look at the SUPER Specifications
Months of leaks have nonetheless formed a consistent outline of the planned models. The GeForce RTX 5070 SUPER is anticipated to combine a GB205 GPU with 18 GB of GDDR7 memory on a 192-bit interface. Moving up the stack, the RTX 5070 Ti SUPER is said to retain the GB203-350 chip with 8,960 CUDA cores while upgrading video memory from 16 GB to 24 GB. At the high end, the RTX 5080 SUPER is expected to utilise a full GB203-450 GPU featuring 10,752 CUDA cores, paired with 24 GB of GDDR7 across a 256-bit bus. These expanded capacities would address one of the most frequent criticisms of the standard RTX 50 series.
AMD’s Alternative Path Could Shift Timelines
The delay arrives at a strategically sensitive moment, as AMD develops its next-generation RDNA 5 architecture. Rumours circulating in the industry suggest AMD may explore a hybrid memory system combining standard GDDR7 with onboard LPDDR5X. Such an approach could offer a path to higher effective memory capacity and bandwidth without depending solely on high-cost GDDR7 packages. If AMD’s strategy proves practical, it may limit Nvidia’s room to postpone launches while waiting for memory prices to settle, potentially prompting a broader reassessment of memory configurations for future GeForce designs.
Source: videocardz.com