GitLab launches MSP-delivered managed DevSecOps on Google Cloud

GitLab has launched a managed version of its DevSecOps platform on Google Cloud, delivered by MSPs.

GitLab-certified MSPs, including Beyond and Digital Future, will deliver and run the managed platform for organisations subject to country-specific sovereignty, data-residency and other regulated requirements.

The company said customers keep full control over where their code, pipelines and security data reside while running the complete platform on Google Cloud, with compliance teams retaining audit visibility into every agent action, merge request and security finding through GitLab’s policy controls.

The companies also deepened their AI roadmap collaboration - the latest Gemini models, including Gemini 3.5, are now available in the GitLab Duo Agent Platform, with GitLab being in Google’s Gemini early-access program means new models are added as they ship.

Google’s Gemma models, including Gemma 4, are also now available for GitLab Duo Self-Hosted, giving customers in self-managed and regulated environments model options beyond Gemini.

The move builds on an April 2026 collaboration that let customers call Google models through GitLab Duo and count the usage toward their existing Google Cloud commitments.

“AI agents are reshaping how software gets built, and the platform at the centre of that shift needs to be one that enterprises can trust with their most sensitive workloads,” said Manav Khurana, chief product and marketing officer at GitLab.

“This partnership with Google Cloud gives enterprises access to the right deployment option, models and cost controls that they need to run DevSecOps at scale, on infrastructure they control, with governance that their compliance teams can audit.”

"From streamlining workflows to writing code, Gemini models are delivering real-world value across a wide variety of use cases,” said Brian Goldstein, vice president and general manager of strategic AI and ISV GTM at Google Cloud.

“By deepening our integration with GitLab, we are making it easier than ever for developers to leverage our most capable models on the platform where they already do their daily work."