AI Assistant Plug Causes High CPU Utilization

Hi Scotcurry4 ! Thank you for your question. Tested this with 2025.2.3 version and had no issue. Could you please try upgrading and see if this helps? If not, I'd kindly ask you to answer the following:

1. Does the issue happen in any project? What if you create a small one, simple “Hello, world!”? Will it happen there as well
2. Can you please reproduce the issue and share IDE logs (Help | Collect Logs and Diagnostic Data) with our team? 
3. Also, please follow the steps from this article to generate the CPU snapshot and send it our way https://www.jetbrains.com/help/go/performance-issues-high-cpu-usage.html (for logs and snapshot, you can use this link to securely upload the data - https://uploads.jetbrains.com ).

Thank you!

Hi Vadym, I have the same problem with GoLand and WebStorm 2025.2.3.

The issue occurs regardless of the size of the project. I have uploaded the diagnostic files for WebStorm here: 2025_10_10_eV1i6iwEWLg4cGF69utw9e.

Thank you!

Here's another interesting observation.  My CPU only spikes when the code example dialog box is visible.  As soon as I scroll up to a location where the code example dialog box is off the screen my CPU utilization goes back to “normal.”  I've included screenshots to hopefully clarify what I mean.

I am seeing something very similar in CLion 2025.2.3 with AI assistant 252.26830.83. When I open the AI window with a new chat, everything is fine. After there is a bunch of content, or I bring up an old conversation, the CPU utilization goes up. The display manager (xorg or wayland, I have tried both) hits ~130%, and clion hits 30% just sitting there. As soon as I close the AI assistant window the CPU utilization drops.

This, same exact symptoms and workaround, except the cpu hogs in my case are clion and gnome-shell. Scrolling the code content box out of the view makes both clion and gnome-shell cpu utilization drop to nearly zero.

I experience the same issue with PyCharm 2025.2.3 and Idea 2025.2.3.

The problem is not present in 2025.1.5 version.

The process which takes 100% of a core is Xwayland.

I second Scotcurry's observation that it disappears when a code segment is out of view (or AI Assistant is closed)

I opened an issue in Fedora but the conclusion was that this is a probem on Idea side

My spec:

Operating System: Fedora Linux 42
KDE Plasma Version: 6.4.5
KDE Frameworks Version: 6.18.0
Qt Version: 6.9.2
Kernel Version: 6.16.10-200.fc42.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 24 × AMD Ryzen 9 7900X 12-Core Processor
Memory: 64 GiB of RAM (62.5 GiB usable)
Graphics Processor: NVIDIA GeForce RTX 4070 Ti
Manufacturer: ASUS

Hey everyone,
Thanks a lot for bringing this issue to our attention and for sharing the detailed information!
Based on your description, I believe you’ve encountered this known issue:
LLM-20308 [2025.2.2 - 2025.2.3] High GPU usage when AI chat is open and contains a conversation
We’ve identified the root cause, and the fix will be included in the 2025.2.4 release (expected in about two weeks)
 

The issue mentions 100% usage of GPU. This thread is about CPU. How come it's the same?

I logged in to report the same issue - looks like they just released an update to AI Assistant to fix it (available now).

It was fixed in last release, latest release again introduced the same issue, this time is not going to static 100% but it is almost constant 80% and the whole IDE freezes. @Scotcurry4  observations hold true for me as well. The issue is worsens with increase in context of a single chat. 

The issue is worsens with increase in context of a single chat. 

The issue is worsens with increase in context of a single chat. 

Thank you!

Thank you!

Thank you!

Thank you!