GitLab Added Microsoft Foundry Support for AI Models

Organizations can now host AI models within their own Azure environments via GitLab Duo Self-Hosted.

Updated on Sept. 22, 2026 in Artificial Intelligence

Isometric editorial illustration of a modular server blade chassis and cabling, representing private enterprise AI infrastructure.
GitLab's updated Duo Self-Hosted platform now supports model deployments via Microsoft Foundry, allowing organizations to manage AI infrastructure within their own Azure environments. AI Illustration. Upload story photo >

Live Poll

Should organizations prioritize controlling their own AI infrastructure over using pre-managed services?

GitLab updated its GitLab Duo Self-Hosted platform to support model deployments through Microsoft Foundry. This integration allows companies to manage AI infrastructure directly within their own Azure environments.

Why it matters

The move provides organizations with increased control over data residency, sovereignty, and network isolation. By routing requests through private infrastructure, teams can better address regulatory compliance requirements.

The architecture comprises three components: a self-managed GitLab instance, a self-hosted AI Gateway, and Microsoft Foundry endpoints. It supports several model families, including OpenAI GPT, Anthropic Claude, Meta Llama, and Mistral.

The players

GitLab

GitLab is a software company that provides a comprehensive DevSecOps platform for project planning, source code management, and CI/CD pipelines.

Microsoft

Microsoft is a multinational technology corporation that provides the Azure cloud computing platform and the Foundry model deployment services.

The details

The AI Gateway acts as an intermediary, enabling administrators to route requests through their private infrastructure rather than relying on external GitLab-managed services. Teams are responsible for managing the specific model deployments, capacity, and networking configurations.

Timeline

  1. GitLab announced the Microsoft Foundry integration on September 22, 2026.

The Tech Race

This move accelerates the industry transition toward private, sovereignty-focused AI hosting, moving away from centralized public model APIs. It positions GitLab as a critical bridge between enterprise software workflows and secure, localized large language model deployments.

Developers and administrators gain the ability to assign specific AI models to designated GitLab Duo capabilities within their secure environments. However, teams must now take on the added responsibility of managing model capacity and infrastructure networking.

The takeaway

Organizations prioritizing data privacy can now leverage familiar model families while maintaining strict network isolation. This shift requires internal teams to become more adept at managing the operational complexities of private AI infrastructure.

Further reading

Learn more about the latest developments in Artificial Intelligence.

Live Poll

Should organizations prioritize controlling their own AI infrastructure over using pre-managed services?