{"id":17941,"date":"2024-10-10T11:14:12","date_gmt":"2024-10-10T11:14:12","guid":{"rendered":"https:\/\/igamemag.com\/nvidia-geforce-rtx-5080-vram-beats-the-rtx-4090\/"},"modified":"2026-10-16T16:22:42","modified_gmt":"2026-10-16T16:22:42","slug":"nvidia-geforce-rtx-5080-vram-beats-the-rtx-4090","status":"publish","type":"post","link":"https:\/\/igamemag.com\/en-gb\/nvidia-geforce-rtx-5080-vram-beats-the-rtx-4090\/","title":{"rendered":"RTX 5080 vs RTX 4090 VRAM: 16GB GDDR7 vs 24GB GDDR6X Explained"},"content":{"rendered":"\n

The RTX 5080 VRAM<\/strong> is 16GB of GDDR7<\/strong> on a 256-bit bus with 960 GB\/s<\/strong> of bandwidth, according to NVIDIA. That’s less memory than the RTX 4090<\/strong>, which has 24GB of GDDR6X on a 384-bit bus, but it’s newer, faster memory. So the RTX 5080 doesn’t beat the RTX 4090 on VRAM capacity: it trades size for speed and Blackwell features. Here’s how the two compare on NVIDIA’s official specs.<\/p>\n

RTX 5080 vs RTX 4090 specs<\/h2>\n\n
<\/td>RTX 5080<\/strong><\/td>RTX 4090<\/strong><\/td><\/tr>\n
Architecture<\/strong><\/td>Blackwell<\/td>Ada Lovelace<\/td><\/tr>\n
VRAM<\/strong><\/td>16GB GDDR7<\/td>24GB GDDR6X<\/td><\/tr>\n
Memory bus<\/strong><\/td>256-bit<\/td>384-bit<\/td><\/tr>\n
CUDA cores<\/strong><\/td>10,752<\/td>16,384<\/td><\/tr>\n
Boost clock<\/strong><\/td>2.62GHz<\/td>2.52GHz<\/td><\/tr>\n
AI TOPS<\/strong><\/td>1,801<\/td>1,321<\/td><\/tr>\n
Total graphics power<\/strong><\/td>360W<\/td>450W<\/td><\/tr>\n
Recommended PSU<\/strong><\/td>850W<\/td>850W<\/td><\/tr>\n
DLSS<\/strong><\/td>DLSS 5 incl. Multi Frame Generation<\/td>DLSS Frame Generation (no Multi Frame Generation)<\/td><\/tr>\n
PCIe<\/strong><\/td>Gen 5<\/td>Gen 4<\/td><\/tr>\n<\/tbody><\/table>\n

How much VRAM does the RTX 5080 have?<\/h2>\n

16GB of GDDR7.<\/strong> NVIDIA lists 960 GB\/s of memory bandwidth for the RTX 5080, up from 717 GB\/s on the RTX 4080 it replaces, which also had 16GB (of GDDR6X).<\/p>\n

Does the RTX 5080 have more VRAM than the RTX 4090?<\/h2>\n

No. The RTX 4090 has 24GB<\/strong>, 8GB more than the RTX 5080. For games that need a lot of memory at 4K with high-resolution textures, or for AI and creative work with large models, the extra capacity of the RTX 4090 (or the RTX 5090’s 32GB) can matter. For most games, 16GB of faster GDDR7 is plenty.<\/p>\n

Is the RTX 5080 faster than the RTX 4090?<\/h2>\n

On paper they’re very different. The RTX 4090 has many more CUDA cores (16,384 vs 10,752) and a wider memory bus. The RTX 5080 has a higher boost clock, more AI performance (1,801 vs 1,321 AI TOPS) and supports DLSS Multi Frame Generation<\/strong>, which the RTX 40 Series doesn’t. In games that support DLSS 5 frame generation, the RTX 5080 can produce higher frame rates; in raw rendering without it, the RTX 4090’s bigger chip still has the edge. It also uses less power: 360W vs 450W.<\/p>\n

How much does the RTX 5080 cost?<\/h2>\n

NVIDIA lists the RTX 5080 from $999<\/strong> in the US and an MSRP of \u00c2\u00a3909<\/strong> in the UK. Partner cards vary in price.<\/p>\n

Which should you buy?<\/h2>\n