macOS: Cowork VM process loads on app launch without ever opening Cowork tab

Resolved 💬 4 comments Opened Mar 5, 2026 by prophetbrianoblivion Closed Apr 3, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

On macOS, the Cowork virtual machine process initializes automatically when Claude Desktop launches, consuming ~1.87 GB of RAM, even when the user never clicks on or interacts with the Cowork tab at all.

The Activity Monitor screenshot was captured immediately after launch — no Cowork tab was clicked, no Cowork session was initiated. The VM process was already running and consuming ~1.87 GB, sandwiched between Adobe Illustrator and WindowServer in memory usage.

What Should Happen?

The VM process should remain dormant until the user explicitly opens or interacts with the Cowork tab. Users who only use Claude Desktop for chat should not incur ~1.87 GB of RAM overhead for a feature they never use.

Error Messages/Logs

Steps to Reproduce

  1. Install Claude Desktop on macOS
  2. Launch the application
  3. Do NOT click the Cowork tab or initiate any Cowork session — no Cowork interaction whatsoever
  4. Open Activity Monitor
  5. Observe: VM process is already running and consuming ~1.87 GB of RAM

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

N/A - Claude Desktop app

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

This is the macOS equivalent of issue #29045 (Windows: Vmmem process consuming ~1.8 GB on every launch). The root cause appears to be VM infrastructure initializing eagerly at startup rather than on-demand.

Note: This is a Claude Desktop bug, not a Claude Code CLI bug — filing here as it appears to be the appropriate repository.

View original on GitHub ↗

This issue has 4 comments on GitHub. Read the full discussion on GitHub ↗