System RAM (random-access memory) is your computer’s temporary working memory. It holds data that the operating system and active applications need to access quickly. Unlike files saved on an SSD or hard drive, its contents are lost when the computer shuts down.
For local AI, system RAM supports model loading, data processing, and the applications running alongside your AI tools. Models that run on the CPU use system RAM to hold their working data. Your total memory needs also depend on the model, its settings, and the rest of your workflow.
System RAM and GPU memory (VRAM) serve different roles. A dedicated graphics card has its own VRAM, while integrated graphics commonly share system memory. Some AI tools can split workloads between the CPU and GPU, but adding system RAM does not provide the same performance as adding VRAM.
When planning your system, leave room for your operating system, AI tools, and other applications to run together. If physical RAM becomes insufficient, the system may move some data to slower storage, reducing responsiveness. Check the requirements of your intended models and tools before choosing a memory capacity.
