As of today, Gemini CLI stops working for Google AI Pro, Ultra, and free-tier users. If you have pipelines, scripts, or workflows built on Gemini CLI, they broke this morning unless you migrated.

Google announced at its I/O developer conference on May 19, 2026 that it was transitioning the Gemini CLI to a new tool called Antigravity CLI, giving developers a 30-day window to migrate. That window closes today, June 18.

What Changed

Gemini CLI launched roughly a year ago as Google’s command-line interface for accessing Gemini models. It picked up a following among developers and data professionals who wanted to integrate Gemini into their local workflows, scripts, and CI pipelines.

Antigravity CLI replaces it entirely. According to Google’s developer blog, the new tool is built in Go rather than Python, making it faster and more responsive in use. The bigger change is architectural: Antigravity CLI is designed as an agent-first platform rather than a model access tool. Instead of running one request at a time, it can orchestrate multiple agents handling different tasks in the background, letting developers run large-scale code refactors or concurrent research tasks without locking the terminal session.

Google confirmed that the most critical capabilities from Gemini CLI carry over: Agent Skills, Hooks, Subagents, and Extensions all work in Antigravity. The company acknowledged there will not be full feature parity from day one, which has drawn criticism from developers who built tools around specific Gemini CLI behaviors.

The Enterprise Exception

Not everyone is affected equally. Teams on Google’s paid tiers have a grace period that personal and free users do not.

Organizations with Gemini Code Assist Standard or Enterprise licenses keep uninterrupted Gemini CLI access. The same applies to developers accessing Gemini via paid Gemini Enterprise Agent Platform API keys. Enterprise customers will not find their workflows cut off today.

For individual developers, students, and teams on free or Pro plans, today is the hard cutoff.

Why Google Made This Move Now

The shift is not just a rebrand. Google has been repositioning its AI tools around agent workflows since Google I/O 2025, and Antigravity represents the cleaner architectural foundation that work requires.

Gemini CLI was built when the primary interaction model was: send a prompt, get a response. Antigravity CLI is built for orchestration, where a developer might kick off five parallel agent tasks and monitor them asynchronously. That difference matters for data teams running multi-step pipelines, for developers building tools that call models as part of larger workflows, and for enterprises wiring AI into their operational systems.

The forced migration is blunt, but the underlying logic is sound. Running two parallel developer tools with overlapping capabilities is expensive to maintain and confusing for users. Google is consolidating on the architecture it believes will define the next three years of AI tooling.

What This Means for Business

If your team uses Google AI developer tools, check your stack today. Automated scripts, CI/CD pipelines, internal tools, and any workflow that calls the Gemini CLI will need updating.

For data teams specifically, this is a good moment to audit how AI model access is wired into your infrastructure. The shift to Antigravity CLI is happening now, but the broader pattern, where AI tools become orchestration platforms rather than single-call utilities, is one every data organization will need to navigate over the next year.

The practical steps are straightforward. Install Antigravity CLI, update your authentication, and test existing workflows. Google has published a migration guide on its developers blog. Enterprise teams on paid plans have more runway, but the architecture is changing regardless of tier.

The deeper implication is worth noting: Google is signaling that its AI developer tools are no longer about model access alone. They are infrastructure for running agents at scale. For anyone building data workflows, internal automation, or AI-powered operations, that shift in framing matters more than the specific CLI change happening today.

Enterprise DNA works with businesses at exactly this inflection point, where the question shifts from “how do we access AI” to “how do we build and govern the agents that run our operations.” If your team is rethinking its AI tooling stack, that is a conversation worth having.