Hi, thanks for reporting! We have an open bug ticket for this issue on Windows (I know you’re on Mac). Would you mind sharing the cursor version info ? You can grab it by going to Cursor → About Cursor and copying the version information you see there. This will help us clarify the impacted versions.
Ok cool, thank you.
We shipped out an update for this ticket that should alleviate the issue on Windows. Are you still seeing the same behavior currently?
hey @gengjiawen just to confirm, on what version of cursor are you now? some major and non major update have come up since (tho idk if the bug where indeed fixed or not)
Thanks for keeping the version info coming, @gengjiawen and sorry this has dragged on for so long.
To clear up where things stand: the fix that shipped earlier was for a Windows-specific version of this problem. The macOS idle polling you’re capturing - the repeated GetTeams / full_stripe_profile / metrics calls firing every second or two while the app just sits idle - is a separate issue on our side, and your latest capture on 3.7.19 (plus @bbur’s) confirms it’s still happening. This isn’t anything on your end.
I’ve raised this with our team to track separately from the Windows fix. No ETA to share yet, but the Process Explorer captures you’ve both posted are exactly what’s needed to chase it down.
One thing that would help narrow it: when you grab these captures, do you have multiple Cursor windows open, or are you connected to any Remote (SSH / WSL / Dev Container)? Multiple windows or a remote connection each add their own background syncing, so knowing your setup helps us tell whether it’s one loop misbehaving or several stacking up.
Following up — still seeing this on a much later version than originally reported here.
Version: 3.9.8 (Commit 4aa8ff1b7877ed7bd01bcba308698f71a6735380, Electron 40.10.3, Darwin arm64 25.5.0)
Battery dropped 100%→37% in 2.5 hours. Activity Monitor Energy tab (12hr Power) showed Cursor at 7,782 while every other app was in the double/triple digits. Did a full update + restart expecting it to resolve — instead it got worse on the next session: 9,202 (12hr Power), still far above everything else running.
@kevinn — given this thread says the issue was “addressed” back in May (v3.3.8) but I’m still hitting it hard on 3.9.8 in June, is this a regression, or a separate issue that looks the same from Activity Monitor’s perspective? Happy to provide a Sample Process capture if useful.
Hi @Moody Thanks for following up, and sorry this is still affecting you.
A Sample Process capture would be useful. Feel free to capture the Cursor Helper (Renderer) process showing the highest Energy Impact while Cursor is idle and attach the output. Could you also confirm:
How many Cursor windows were open
Whether any used Remote SSH or Dev Containers
Whether it persists with one local window and no remote sessions