Nvidia RTX 5050 to have faster GDDR6 memory than RTX 4060

RTX 5050 matches Radeon RX 9060 XT memory speeds with 20 Gbps GDDR6 memory

Thanks to @Zed__Wang on Twitter/X, we have memory specifications for Nvidia’s upcoming RTX 5050 graphics card. Based on this leak, Nvidia’s GeForce RTX 5050 will feature 8GB of 20 Gbps GDDR6 memory. This is the same memory speed that AMD uses on its RX 9000 series GPUs.

Strangely, this means that AMD’s RX 9060 XT 8GB has the same memory configuration as Nvidia’s RTX 5050. Both feature 8GB of GDDR6 memory over a 128-bit memory bus. This suggests that AMD may now have stronger memory compression and bandwidth efficiency than Nvidia. Traditionally, this has been a strong point for Nvidia.

By the way, Desktop RTX5050 uses 20Gbps GDDR6, the same as the RDNA4 family. https://t.co/Va2Qj7ZRIe

— MEGAsizeGPU (@Zed__Wang) June 13, 2025

Nvidia’s RTX 5050 will likely be Nvidia’s most affordable RTX 50 series GPU model. This new GPU reportedly features 2560 CUDA cores and 8GB of GDDR6 memory. This gives Nvidia’s RTX 5060 50% more CUDA cores and much more memory bandwidth. This new model also has fewer CUDA cores than Nvidia’s last-generation RTX 4060 (which has 3072 CUDA cores).

With more memory bandwidth and fewer CUDA cores than an RTX 4060, it will be interesting to see how Nvidia’s GeForce RTX 5050 performs. Rumour has it that Nvidia plans to launch its RTX 5050 next month. Given its specifications, we expect this GPU to cost around $200. After all, a GPU with two-thirds of a $300 GPU’s CUDA cores should cost $200 (preferably less).

You can join the discussion on Nvidia’s rumoured RTX 5050 graphics card on the OC3D Forums.