this videos breaks down the cost of owning your own GPUs to run local AI models, which ones to buy, and which models you can run on them. with 3x RTX 3090s of ~$800 each, you can run some of the top models in the 30B-70B range such as Qwen or Gemma. https://youtu.be/z83-C4iwwEk?is=OLo2fuhR…