Nvidia is finally stepping on the gas! RTX 5070 Ti Super will offer the desired graphics memory increase

  • The new RTX 5070 Ti Super will offer 24 GB of GDDR7 memory instead of the original 16 GB
  • The number of compute cores remains at 8,960, but higher clock speeds and a power consumption of 350 W are expected
  • This is a reaction to criticism that the current RTX 50 generation offers the same amount of memory as the previous generation

Sdílejte:
Jakub Kárník
Jakub Kárník
5. 7. 2025 12:30
nvidia rtx 5070 ti

Known leaker @Kopite7kimi has once again brought interesting information from behind the scenes at Nvidia. This time it’s about the upcoming graphics card RTX 5070 Ti Super (designation PG147-SKU55), which represents an improved version of the RTX 5070 Ti model launched earlier this year.

Same core, but much more memory

This card is expected to use the same GB203-350-A1 Blackwell architecture core as the standard RTX 5070 Ti, including the identical number of 8,960 CUDA cores. These are essentially small processing units that perform parallel computations when rendering graphics or, for example, performing artificial intelligence calculations. The more there are, the faster the graphics card can process data.

The biggest upgrade concerns memory. While the standard RTX 5070 Ti has 16 GB of GDDR7 memory, the Super version will offer a full 24 GB. This represents a significant 50% increase, which will be especially appreciated by players of demanding games at 4K resolution. At high resolutions, much more textures and other data need to be loaded, which must be stored in the graphics card’s memory. More memory means smoother gameplay and less stuttering when loading new areas.

Why is so much memory actually needed?

You might be wondering why someone would need a graphics card with 24 GB of memory. It’s simple – modern games are increasingly demanding. At 4K resolution, very detailed textures are loaded, which take up a lot of space. When you add various ray-tracing effects (realistic shadows and reflections), the need for memory increases even further.

Think of it as a large library of pictures. The more detailed and larger pictures you want to view at once, the larger library you need. If the graphics card’s memory is too small, it constantly has to “discard” some pictures and load others, which causes tearing and a reduction in game fluidity.

Besides gaming, content creators will also appreciate the large memory. When rendering videos, 3D models, or working with large photographs, 24 GB of memory is a huge advantage that can significantly shorten project processing time.

Higher power consumption, but same memory bandwidth

Interestingly, despite the larger memory, its speed and bus width will remain the same as the standard version. The RTX 5070 Ti Super will retain memory with a speed of 28 Gbps and a 256-bit bus, giving a total bandwidth of 896 GB/s. For comparison, the upcoming RTX 5080 Super is expected to offer faster 32Gbps modules, providing 1 TB/s bandwidth.

The power consumption of the new card should increase from 300 W to 350 W, which represents a 16% increase. This shift is likely due to a combination of more memory modules and potentially higher core clock speeds, although specific values are not yet known.

Reaction to criticism and competition

The massive increase in memory on the Super models can be seen as a direct reaction by Nvidia to the criticism the company faced after the launch of the RTX 50 series. Many users were disappointed that the new generation (with the exception of the top-tier RTX 5090 model) did not offer more memory than its predecessors, even though games and applications are increasingly demanding.

Furthermore, as AMD pushes hard and offers generous amounts of memory in its cards, the green team had to react. Right now, it looks like Nvidia has decided to rectify this “mistake” through the Super models.

Summary of changes compared to the standard model

To summarize, the RTX 5070 Ti Super compared to the standard RTX 5070 Ti will offer:

  • The same number of cores (8,960 CUDA)
  • O 50 % více videopaměti (24 GB místo 16 GB)
  • Stejnou propustnost paměti a šířku sběrnice (256 bitů)
  • Stejnou rychlost paměti (28 Gbps)
  • O 16 % vyšší spotřebu (350 W místo 300 W)

Očekává se, že cena zůstane podobná jako u běžné verze, tedy kolem 749 dolarů (v Česku pravděpodobně něco málo přes 20 tisíc korun). Vzhledem k výraznému navýšení paměti by to mohla být zajímavá nabídka pro hráče, kteří hledají kartu pro hraní ve 4K bez omezení způsobených nedostatkem videopaměti.

Are you eyeing new graphics cards from Nvidia?

Source: wccftech, Kopite7kimi

About the author

Jakub Kárník

Jakub is known for his endless curiosity and passion for the latest technologies. His love for mobile phones started with an iPhone 3G, but nowadays… More about the author

Jakub Kárník
Sdílejte: