Hardware buying guide by memory tier
How to choose hardware for local LLMs: memory capacity first, bandwidth second, compute third. Tier table, GPU vs Mac vs CPU trade-offs, multi-GPU realities, used cards, power and cooling, and a worked purchase decision.
Do not shop by brand or by gaming benchmarks. For local LLMs the order of importance is memory capacity first, memory bandwidth second, compute third. Capacity decides which models run at all. Bandwidth decides how fast they decode. Compute decides prompt processing speed and how well batching scales. Prices change monthly, so this guide is organised by tier and by reasoning, not by SKU.
Pro content
Keep reading with Inference Academy Pro
Unlock the rest of this lesson, plus every advanced lesson, hands-on lab, the larger playground models and completion certificates.