The GPU's running LLM's might be having their moment. We'll see how long it lasts as the cost to consume comes down.