Let's demystifies the legal and licensing challenges of deploying Large Language Models on‑premise. We will compare major model licenses, clarifies true open‑source options, explains “fair use” and “non‑commercial” pitfalls, and show why Apache 2.0 and MIT licenses underpin reliable, compliant AI stacks for enterprise and edge environments.

As AI adoption accelerates, many teams discover that deploying Large Language Models on‑premise isn’t just a technical challenge, it’s a legal one. Between “open” models that aren’t truly open, enterprise‑friendly licenses, and increasingly restrictive “fair use” or “non‑commercial” clauses, choosing the wrong component can quietly lock you out of production.

In this talk, we’ll decode the licensing landscape behind today’s most popular LLMs, inference engines, and AI tooling. We’ll compare model licenses side‑by‑side, highlight which ecosystems are genuinely open source (and which only look like it), and explain why Apache 2.0 and MIT have become the bedrock of trustworthy, production‑ready AI stacks. You’ll also learn how to navigate “fair use” and “non‑commercial” licenses—what they really allow, where the traps are, and how they affect on‑premise and edge deployments.

If you want to build AI solutions without giving your legal team a headache, this talk is for you.

As legal compliance is important but not the favorite topic of developers, let's do 20min presentation + question, followed by a "Summer community meeting".

So, prepare your webcam and microphone to be part of the discussion.

Session:
Wednesday 22 July 2026
10 AM EDT | 3 PM BST | 4 PM CEST

Speaker:
Antoine Thomas, Community Manager Open Source Solution, Hyland