GitLab Integrates Agentic AI with AWS Bedrock, Streamlining Enterprise DevSecOps

Event summary

  • GitLab and AWS have deepened their integration, allowing GitLab Duo Agent Platform inference to run through Amazon Bedrock.
  • The integration enables enterprises to use existing AWS Bedrock accounts, IAM policies, and spending commitments for agentic DevSecOps.
  • GitLab’s orchestration layer adds workflow-level governance on top of Bedrock’s model-level controls.
  • The move builds on GitLab’s Bring Your Own Model (BYOM) capability, allowing teams to use self-hosted or GitLab-managed AI models.
  • GitLab Duo Agent Platform is billed per request, with costs drawn down from existing AWS spend commitments for customers using AWS Marketplace.

The big picture

This collaboration addresses a critical challenge for enterprises embracing AI: integrating generative AI capabilities into existing workflows without creating parallel infrastructure or increasing vendor risk. By leveraging existing AWS investments and GitLab’s orchestration layer, the partnership aims to accelerate AI adoption within established cloud environments, a trend that will likely see increased competition and further integration across the DevSecOps landscape.

What we're watching

Governance Dynamics
The success of this integration hinges on GitLab’s ability to effectively govern AI agent workflows within existing AWS environments, preventing model drift and ensuring compliance.
Adoption Rate
The pace at which enterprise customers adopt this integrated solution will depend on their existing AWS Bedrock commitments and comfort level with agentic AI.
Competitive Response
Other DevSecOps platforms will likely respond with similar integrations, intensifying competition and potentially driving down pricing for agentic AI services.