When I did some quick math regarding buying hardware outright vs spinning up cloud versions, cloud seemed to win pretty handily. I've used runpod a bit, they h

When I did some quick math regarding buying hardware outright vs spinning up cloud versions, cloud seemed to win pretty handily. I've used runpod a bit, they have serverless which is good for sporadic inference. For training LoRAs I used their regular cloud offering (which can be switched off), they have volumes so you shouldn't need to re-download models (although if they're stale for a long period of time I think they reclaim them). It's certainly not perfect, their product is also not mature. But I got a lot done, and for not a lot of money.