GIGABYTE Gears Up to Launch a Whopping 14 GeForce RTX 40 SUPER Series GPUs

GIGABYTE accidentally leaked on their official Russian website the number of GeForce RTX 40 SUPER graphics cards they will release. Specifically, NVIDIA will launch three models, and among various cooling systems and overclocking, GIGABYTE will release no less than 14 different models.

It should be noted that by the end of January, NVIDIA will launch its GeForce RTX 4080 SUPER, GeForce RTX 4070 Ti SUPER, and GeForce RTX 4070 SUPER. However, it will be much earlier, during the second week of January, when NVIDIA officially announces these graphics cards at CES Las Vegas.

GIGABYTE will launch 14 different GeForce RTX 40 SUPER graphics cards

To be exact, GIGABYTE will release four GeForce RTX 4080 SUPER graphics cards. These are from the AERO, AORUS, GAMING, and WindForce 3 series. The most repeated model is the GeForce RTX 4070 Ti SUPER with five different models. Thus, we have the 4070 Ti SUPER AERO, AORUS, EAGLE, GAMING, and WindForce 3. We end with the RTX 4070 SUPER, which also repeats with five different models, the same as its older sister. Visually, they will even be identical since each model is linked to the cooling system used by each GPU.

Certainly, along with GIGABYTE, other companies will follow this strategy, although most likely with a significantly lower number of variants to avoid confusing consumers. Also, releasing 14 different variants demonstrates that GIGABYTE believes they will be popular; otherwise, they would limit themselves to 2 or 3 variants of each model, as will most other brands.

This is what we expect from the GeForce RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER

The NVIDIA GeForce RTX 4080 SUPER would arrive with 10,240 CUDA Cores. This represents only a 5% increase in cores compared to the 9,728 CUDA Cores of the RTX 4080. On the memory side, it maintains the same 16 GB GDDR6X capacity with the same 256-bit memory interface. However, this memory will reach 23 Gbps (vs 22.4 Gbps) to increase the bandwidth up to 736 GB/s (vs 717 GB/s).

We continue with the NVIDIA GeForce RTX 4070 Ti SUPER, which also disappoints with its 8,448 CUDA Cores. This implies 10% more cores (768 CUDA Cores) than the original GeForce RTX 4070 Ti. At least on the memory side, there are much more noticeable improvements. We go from 12 GB GDDR6X to 16 GB GDDR6X. Although the same speed of 21 Gbps is maintained, we go from a 192-bit memory interface to a 256-bit one. This implies a significant improvement in bandwidth, going from the current 504 GB/s to 672 GB/s.

We finish with the NVIDIA GeForce RTX 4070 SUPER. This GPU improves the most, as its graphics chip will have 7,168 CUDA Cores. This implies a 21% increase in cores (1,280) compared to the original GeForce RTX 4070. On the memory side, it’s disappointing. It maintains the same 12 GB GDDR6X memory at 21 Gbps and keeps the same 192-bit memory interface, making the bandwidth the same: 504 GB/s.

In this way, the RTX 4070 SUPER greatly improves at the silicon level. The RTX 4070 Ti SUPER significantly improves in terms of memory, and the RTX 4080 SUPER is the big disappointment, with a slight improvement in GPU and memory. Obviously, they want the GeForce RTX 4090 to remain the best GPU in the RTX 40 series.