GitLab routes agentic DevSecOps to Google Cloud

GitLab said on April 14 that its artificial intelligence coding agents can now run through Google Cloud’s Vertex AI, with the bill counting toward existing Google Cloud commitments. (about.gitlab.com) The change puts Google’s Gemini models and other foundation models behind GitLab Duo Agent Platform, which GitLab released in general availability on January 15, 2026. GitLab said the agents can handle planning, coding, review, security remediation, and other software delivery tasks inside its platform. (about.gitlab.com) GitLab said those agent calls will run through its artificial intelligence gateway on Google Cloud services such as Google Kubernetes Engine and Cloud Run, so customers do not have to stand up separate model infrastructure. The company said the same setup works for GitLab-managed and self-hosted deployments, with a bring-your-own-model option for self-hosted customers. (about.gitlab.com) Vertex AI is Google Cloud’s platform for accessing and serving large models, including Gemini, through one managed control plane. Google says its Model Garden catalog includes Google models and other foundation models, while Vertex AI provides deployment, serving, and governance tools. (cloud.google.com 1) (cloud.google.com 2) Google Cloud commitments are long-term spending or resource contracts, usually for one or three years, that companies use to lower cloud costs. Google says those commitments can be spend-based or resource-based, so letting GitLab usage count against them moves artificial intelligence agent spending into an existing procurement bucket. (docs.cloud.google.com) GitLab framed the deal around governance as much as model access. In its announcement, the company said agent actions in Duo Agent Platform keep GitLab’s built-in compliance and audit controls, a selling point for large companies that want coding automation without moving outside approved software delivery workflows. (ir.gitlab.com) The partnership lands as software vendors try to move buyers from one-off code assistants to broader “agentic” systems that can take multi-step actions across development and security work. GitLab’s January launch argued that faster code writing alone leaves bottlenecks in code review, compliance checks, bug fixing, and vulnerability handling. (about.gitlab.com) Investors treated the Google Cloud tie-up as material, at least initially. MarketWatch reported GitLab shares rose about 5% in after-hours trading on April 14 after the announcement. (marketwatch.com) For GitLab customers already locked into Google Cloud budgets, the pitch is simple: use the same cloud contract, the same infrastructure estate, and the same compliance controls to add artificial intelligence agents to software delivery. GitLab and Google Cloud are betting that billing simplicity will help sell the next layer of coding automation. (about.gitlab.com)