Claude on Windows eats 2 GB of RAM, a problem for budget laptops

Claude wastes about 2 GB of RAM on Windows laptops, leaving minimal resources for other applications.

Why Claude needs so much memory

Users of Windows laptops complain that Claude reserves about 2 GB of RAM even when it is not doing any active work. This is pure overhead: the memory is wasted, taken away from the operating system and other applications.

For comparison, on Linux and macOS the problem is less critical or does not appear at all. On Windows the behavior is especially greedy, which makes working on devices with 8 GB of RAM uncomfortable.

From the available information, it is unclear what exactly causes this behavior: it may be a quirk of how the app runs on Windows, accumulated bugs in the code, or a problem with the libraries Claude uses. Complaints suggest the problem is not new and has long been on the developers' agenda, but no fix has appeared yet.

Why this matters now

Laptops with 8 GB of memory have become popular again. A few years ago the market shifted toward 16+ GB configurations as the standard, but high prices for modern devices have brought back demand for more modest, affordable options. Many professionals and students have no choice but to work on older hardware that was considered normal just a few years ago.

On such machines every gigabyte counts. If Claude wastes 25% of RAM in the background, very little is left for normal work:

• Browser (Firefox, Chrome) — 1–2 GB with open tabs
• IDE (VS Code, JetBrains) — 0.5–1.5 GB with a project
• Other work applications — 0.5–1 GB
• System processes and updates — 0.5–1 GB

The result: the laptop starts swapping heavily to disk, freezes, and everything crawls.

Current options

There is no convenient way out, and users face a difficult choice:

1. Give up Claude — switch to other AI tools with lower memory use 2. Upgrade the laptop — buy a new device with 16+ GB of memory (expensive) 3. Add RAM — if the design allows it (often impossible on modern models) 4. Wait for a patch — hope that Anthropic optimizes the app (no timeline)

None of these options looks attractive to an ordinary user.

What this means

AI assistants are quickly becoming more resource-hungry. Where they once lived in the browser or ran in the cloud, they increasingly require local installation with heavy memory and compute use. For the mass market of consumer laptops, this means access to modern AI tools is narrowing precisely for users on budget hardware, the people who most need help with automation.