NVIDIA is reportedly planning to release ‘Super’ variants of its RTX 5070 and RTX 5070 Ti GPUs, featuring significant increases in VRAM. These rumored upgrades, including 18GB for the RTX 5070 Super and 24GB for the RTX 5070 Ti Super, aim to enhance performance for demanding gaming and professional workloads.
NVIDIA’s Next-Gen ‘Super’ GPUs: More VRAM on the Horizon
Rumors are circulating that NVIDIA is preparing to launch ‘Super’ refresh models for its GeForce RTX 5070 and RTX 5070 Ti graphics cards. These new iterations are expected to bring substantial upgrades, particularly in terms of Video Random Access Memory (VRAM), addressing a common concern among high-end users and gamers.
Key Takeaways
- Increased VRAM: The RTX 5070 Super is rumored to feature 18GB of GDDR7 VRAM, a significant jump from the non-Super RTX 5070’s 12GB. The RTX 5070 Ti Super is expected to boast an impressive 24GB of GDDR7 VRAM, up from the RTX 5070 Ti’s 16GB.
- Enhanced Performance: While core counts might see minimal changes, the increased VRAM, coupled with potential clock speed boosts and higher Total Graphics Power (TGP), is anticipated to improve performance, especially in 4K gaming and memory-intensive applications.
- Strategic Move: This refresh could be a strategic response to competition, particularly from AMD’s RDNA 4 GPUs, aiming to offer more compelling options in the mid-to-high range GPU market.
Detailed Specifications and Improvements
The rumored specifications highlight a clear focus on VRAM capacity:
| GPU Model | Current VRAM | Super VRAM (Rumored) |
|---|---|---|
| GeForce RTX 5070 | 12GB GDDR7 | 18GB GDDR7 |
| GeForce RTX 5070 Ti | 16GB GDDR7 | 24GB GDDR7 |
Beyond VRAM, other potential changes include:
- RTX 5070 Super: Expected to have a slight increase in CUDA cores (around 6400 from 6144) and a TGP of 275W (up from 250W).
- RTX 5070 Ti Super: While maintaining similar CUDA core counts (8960), its TGP might increase to 350W (from 300W) to support the larger VRAM and potentially higher clock speeds.
These VRAM increases are reportedly achieved by utilizing 3GB memory modules, allowing for greater capacity on the same memory bus interfaces (192-bit for 5070 Super, 256-bit for 5070 Ti Super).
Market Implications and Release Timeline
The introduction of these ‘Super’ variants suggests NVIDIA’s intent to bolster its offerings and maintain competitiveness. The increased VRAM could significantly improve the longevity and performance of these cards in future titles and professional workloads that demand more memory.
While no official announcement has been made, industry speculation points towards a potential debut around CES 2026. The pricing strategy for these new GPUs will be crucial, with some rumors suggesting NVIDIA might aim to keep prices competitive, possibly even at similar MSRPs to their non-Super counterparts, to counter market pressures.
