There is a particular irony in a company that sells software development tools deciding it needs fewer software developers. That is, in effect, what GitLab announced this month, and it is not alone. A run of cuts across the developer economy suggests the firms closest to artificial intelligence are the first to bet their own headcount on it.
GitLab bets the company on agents
GitLab is cutting about 14 percent of its full time staff, roughly 350 people, and exiting 22 countries, a 37 percent reduction in its geographic footprint. Chief executive Bill Staples framed the move around what he calls the agentic era, saying it is creating structural tailwinds for the company and pointing to accelerating platform activity in the first quarter. The reorganisation also strips out up to three layers of management in some functions. Crucially, GitLab says the savings will be reinvested into research and AI products rather than used to pad margins, with a pre tax restructuring charge of 30 to 35 million dollars.
The agents need somewhere to run
The cuts come as GitLab pushes its Duo Agent Platform, deepening an integration with Anthropic's Claude models and announcing tie ups to run agentic features on Amazon Web Services Bedrock and Google Cloud Vertex AI. Staples has described the effort as a generational rebuild of the underlying technology to handle a step change in automated workloads. This is the central tension of the moment in miniature: the same quarter that revenue rose to 264.2 million dollars, up from 214.5 million a year earlier, the company decided it could grow with a smaller, flatter workforce.
A pattern across the developer economy
GitLab is part of a wider trend. Sea, the parent of the e commerce platform Shopee, is cutting hundreds of developer roles globally, about 8 percent of Shopee's developer workforce, as it pivots toward AI. Salesforce has trimmed staff across its Agentforce, MuleSoft and Marketing Cloud products even while marketing those tools as labour saving. In each case the message to investors is the same: AI lets us do more with fewer engineers, and the proof is that we are now doing exactly that to ourselves.
What it means for software careers
For years software engineering looked like the safest possible bet against automation, the discipline that built the machines. That assumption is being tested in public. None of these announcements proves AI can replace skilled developers wholesale, and revenue at firms like GitLab is still climbing. But the signal to anyone entering the field is unmistakable: the work is being reorganised around tools that did not exist three years ago, and the companies selling those tools are reorganising first.
