Rumors about the NVIDIA GeForce RTX 50 Super series are heating up again. Multiple reports indicate that the new cards are expected to hit the market between the end of Q1 and the beginning of Q2 2026, somewhere between March and May. However, they won't be released at CES 2026. While NVIDIA has yet to provide an official notification to partners, some indications suggest that the Super Series does exist. For example, entries for the RTX 5070 Super and RTX 5070 Ti Super have already appeared in the Seasonic Power Calculator, and noted tipster @Kopite7Kimi revealed similar news a few months ago.

The RTX 50 Super series will continue its strategy of improving upon existing models by expanding memory and boosting performance. The focus remains on the 70 and 80 series products, with the inclusion of the 60 series yet to be confirmed. Considering that the 5090 already features 32GB of GDDR7 memory, the Super version is more likely to target the mid-range and high-end segments to meet the memory needs of high-resolution, demanding games.

Leaked information suggests that the RTX 5070 Super might see an increase in memory from 12GB to 18GB, while the 5070 Ti Super could see a rise from 16GB to 24GB, with both models experiencing an increase in TDP, hinting at greater power consumption and enhanced performance. Another major point of interest is the RTX 5080 Super, expected to come with 24GB of GDDR7 memory, a 50 percent increase over the base 5080's 16GB. This aligns with the upgrade pattern of previous Super series, emphasizing memory expansion, bandwidth enhancement, and slight frequency increases, while maintaining the same architecture.

According to the product list at this stage, the RTX 5080 Super will be based on the GB203-450 core, featuring 84 SM units for a total of 10,752 CUDA cores, a 256-bit bus, a 32 Gbps memory rate, and bandwidth of up to 1024GB/s. Comparatively, the RTX 5080 will have the same core but with 16GB of memory and 960GB/s of bandwidth. The RTX 5070 Super and 5070 Ti Super are based on the GB205 and GB203 cores, featuring adjusted memory and power consumption. These Super models aim to address the shortcomings of their predecessors in high-resolution ray tracing scenarios by adding more memory and enhancing bandwidth.

As for pricing, NVIDIA has not yet disclosed the price points for the Super Series. Historically, new cards tend to maintain a similar price point to the prototype model, or even coexist with the original for some time after launch, before gradually replacing older models. With current suggested retail prices of $999 for the RTX 5080, $749 for the RTX 5070 Ti, and $549 for the RTX 5070, it's plausible that the Super Series will be priced in a similar range, but with higher memory as a key selling proposition.

The market context highlights the strategic importance of the Super Series, as NVIDIA continues to dominate high-end performance with the 5090 and 5080, while the mid-range and high-end markets remain crucial sales drivers. Given that AMD's RDNA 5 series and Intel's Battlemage GPUs are closing the gap, NVIDIA needs to ensure its competitiveness in the mainstream price bracket. Memory expansion remains one of the fastest and most effective methods to achieve this, particularly with the growing demand for AI local processing and high-resolution gaming, making the allure of additional memory increasingly significant to users.

To summarize, while the RTX 50 Super Series wasn't unveiled at CES, several indications point toward its development, with a potential release window set for Spring 2026. Should they be launched, the RTX 5070 Super, 5070 Ti Super, and 5080 Super will be crucial choices for mid-range and high-end gamers, reinforcing NVIDIA's market dominance through diversified product offerings.