What is a private LLM?

A private LLM is an AI language model set up so that a company’s prompts and data stay under its own control. Usually that means a downloadable open-weight model running on the company’s own servers or in its own private cloud account. The term is also used loosely for a company-only AI assistant on an enterprise plan. There, data does go to the provider, but under a contract that rules out training on it.

Why it matters for a business

The 2 meanings answer different questions. An enterprise plan settles training and retention by contract, and the data still sits on someone else’s infrastructure. A self-hosted model keeps the data inside, which some regulated work requires, and hands the company everything else: capacity paid for whether busy or not, upgrades, the security of its own logs, and someone on call. Self-hosting answers whether the data leaves the company. It does not answer who inside the company can read the logs.

Go deeper

Open-source vs. closed AI models: what is actually open

On Atlas: LLM serving and inference, compared

Related terms

• Open-weight model
• Open-source AI
• Inference
• Shadow AI

All terms, A to Z