VRAM (Video Random Access Memory)
VRAM is the high-speed, dedicated memory built directly onto a graphics card (GPU). While regular system RAM holds your operating system, browser tabs, and everyday apps, VRAM acts as a private, ultra-fast workspace exclusively for graphics and AI calculations.
When an AI model generates text, it has to read billions of internal numbers from memory for every single word it produces. Because VRAM sits right next to the GPU chip and moves data 10x to 20x faster than standard motherboard RAM (often 500 to 1,800 GB/s compared to 50 to 90 GB/s), fitting the entire model and your chat history inside VRAM is what makes replies stream onto your screen smoothly instead of crawling at a few words per second.
For local AI buyers, VRAM capacity is the most important hardware spec to watch because it sets a hard ceiling on which models you can run at full speed. An 8GB graphics card easily handles compact 8B models, while stepping up to a 24GB card unlocks much smarter 32B models. Unlike regular desktop RAM sticks that you can swap out later, VRAM is permanently soldered onto the graphics card, meaning the only way to add more is to buy a larger GPU or link a second card.
Looking for verified hardware rigs?
Check our real-time £/GB VRAM index to find systems optimized for local model inference.