# LoRA Hot-Swap Lab: Single Base + N Adapters — 50 Customers Served on a Single 4090

> Source: https://sukruyusufkaya.com/en/learn/fine-tuning-cookbook/ftc-lora-hot-swap-multiplexing-vllm
> Updated: 2026-05-14T14:43:00.953Z
> Category: Fine-Tuning Cookbook (Model-by-Model)
> Module: Part XV — Serving Engineering
**TLDR:** vLLM 0.3+'s killer feature: single base + N LoRA adapters, runtime hot-swap. Separate LoRA per customer, all on same 24GB. Llama 3.1 8B base (~5 GB AWQ) + 30+ adapters (~40 MB each) → 50 customers on single 4090. QPS-vs-latency curve.