[1/2] GitLab Act 2Why GitLab is Rebuilding Development for Agents (Category AI4SDLC)
Dealing with GitLab Act 2. Formally, this CEO letter is about the new stage of the company: focus, restructuring, AI and DevSecOps. But more importantly, GitLab does not describe a set of AI features, but a change in the production model of development. In short, GitLab no longer wants to be a DevSecOps platform to which AI has been pinned. He wants to become a control plane for SDLC, where not only people work, but also agents.
This document shows several product shifts.
1Git and platform must withstand the load of machines If agents open MR in parallel (merge request)If you run pipelines and come back with edits faster than any human team, the development infrastructure becomes a bottleneck. Therefore, GitLab speaks not only about AI assistant, but about API-first services and Git for machine loads.
2You need to orchestrate everything. cycle The agent who wrote the code does not solve the enterprise problem by itself. Link issue, code, review, security, CI/CD, deployment, incidents, approvals and policies. Otherwise, there will be a lot of quick action, but not a managed delivery of change.
3The main asset is the context. Code generation is quickly becoming a commodity: everyone will have good models. The difference will be what context the agent gets: tasks, architecture, MR history, vulnerabilities, ownership, deploy history, incidents, past team decisions. GitLab calls it the connected data model across the lifecycle, which is the graph of the engineering context.
4Governance should be built into the core, not added sideways In agent development, speed without audit, identity, policies, and deployment controls quickly turns into risk. You need to understand who or what changed the code, by what rules, with what rights and where a person is obliged to remain in the loop.
5GitLab honestly describes hybrid mode Not “everything is autonomous tomorrow,” but the coexistence of the work of humans, assistant agents, and autonomous agents. This is a sobering picture: big companies will move toward agency not in a single leap, but through layers of control and trust.
The organizational part is no less important than the product part. GitLab makes management more flat, says something like 60 Autonomous R&D teams, requires daily use of AI and changes the old CREDIT framework (Collaboration, Results, Efficiency, Diversity & Inclusion, Iteration, and Transparency) Speed with Quality, Ownership Mindset and Customer Outcomes. In report 2 June 2026 The company also described the restructuring: a reduction of 14% of the state, about 350 man out 22 The decrease in footprint is approximately 37%. It seems that this is an attempt to rebuild the company itself for a new speed of development.
It's not just what GitLab says about agents that matters to me. Importantly, it links them to Git, CI/CD, governance, data models, pricing, and organizational structure. This is no longer a productivity tool for a developer, but a new production system for software development. In next Let's see how lonely GitLab is. ||Spoiler: hardly alone. GitHub, Atlassian, Cursor, JetBrains, and OpenAI/Codex all go the same way, but each goes from its own point of power.||
#AI #AI4SDLC #Engineering #DevSecOps #Management #Architecture #DevOps #PlatformEngineering #Agents #DevEx #Strategy