I am looking into renting Hetzner GEX44 dedicated server to run a couple models on with Ollama. I haven't done the arithmetics yet but I wouldn't be surprised t

I am looking into renting Hetzner GEX44 dedicated server to run a couple models on with Ollama. I haven't done the arithmetics yet but I wouldn't be surprised to see a 100x cost-decrease compared to OpenAI APIs (granted the models I'll run on the GEX44 machines will be less powerful)