Another question regarding Helix - its being built as a platform for private air-gap-ready ai agents that can work against private LLM models.

Are there appliances or easy to deploy hardware that allow one to run these private models on-premise vs cloud