Skip to main content
TechnologyJun 5, 2026· 3 min read

NVIDIA GeForce RTX 5000 SUPER: Is the refresh with more GDDR7 memory back on track?

Rumors about the potential GeForce RTX 5000 SUPER have resurfaced after several months of apparent stagnation, with various reports several times suggesting a possible cancellation due to the "memory crisis." According to leaker MEGAsizeGPU, NVIDIA has reportedly resumed plans for a refresh of the Blackwell family aimed at the consumer market, with a debut still expected during 2026.

The company has never officially announced the arrival of SUPER models based on the Blackwell architecture, and currently, there are no concrete indications of a launch window. However, the leaker claims that the project is active again and that the range might also include a GeForce RTX 5060 with 12 GB, which could potentially be named GeForce RTX 5060 SUPER.

RTX 50 Super is back on track. This time includes 5060 12G (or maybe it will have a new name as 5060 super) — MEGAsizeGPU (@Zed__Wang) June 5, 2026

Previous rumors had already outlined a strategy focused primarily on increasing video memory capacity. In particular, there was talk of a GeForce RTX 5080 SUPER equipped with 24 GB of GDDR7, an RTX 5070 Ti SUPER also with 24 GB, and an RTX 5070 SUPER with 18 GB. Compared to the models currently on the market, these increases would be possible through the adoption of new 3 GB GDDR7 chips instead of the 2 GB solutions used today.

The possible RTX 5060 with 12 GB represents a novelty compared to previous rumors but appears plausible from a technical standpoint. A card with a 128-bit memory bus could indeed reach this capacity using four 3 GB GDDR7 modules, without needing to modify the width of the bus itself. This is a different approach compared to past rumors that had hypothesized configurations with 9 GB for some variants of RTX 5050 and RTX 5060 Ti, which were achieved with a 96-bit memory configuration.

According to information released in previous months, the RTX 5000 SUPER project was temporarily slowed down due to the high demand for GPUs intended for artificial intelligence and the difficulties related to the availability of memory. MEGAsizeGPU also claims that the situation regarding VRAM supply has not yet been fully resolved, but believes NVIDIA can manage the situation.

The technical specifications attributed to the RTX 5070 SUPER and RTX 5080 SUPER versions remain unchanged from previous leaks. The former is credited with 6,400 CUDA Cores, 18 GB of GDDR7 memory on a 192-bit bus, and a TGP of around 275 watts. The latter is expected to offer 10,752 CUDA Cores and 24 GB of GDDR7 on a 256-bit bus, with a declared consumption of about 415 watts. For the RTX 5070 Ti SUPER, there are talks of 8,960 CUDA Cores and 24 GB of memory.

It remains prudent to keep caution. NVIDIA has never confirmed the existence of the RTX 5000 SUPER range, and during recent public events such as GTC and Computex, the company's focus has been almost exclusively on artificial intelligence and new initiatives related to the NVIDIA RTX Spark platform, with no reference to future updates in the GeForce family.