Memory is the spec that most often decides whether a laptop still feels fast in year three — and on many modern laptops it can’t be upgraded after purchase. Buy too little and you’ll live with slowdowns; buy too much and you’ve wasted money that could have gone to storage or a better screen. This guide explains what RAM does, how unified memory on Apple silicon changes the picture and how much you need for your work.
Short answer
- 8 GB: only for very light use (a few browser tabs, documents). Avoid on a new laptop if you can.
- 16 GB: the sensible minimum for most people in 2026.
- 24–32 GB: developers, heavy multitaskers, photo and 4K video editors.
- 48–64 GB+: virtual machines, large datasets, 8K video, 3D and running large AI models locally.
What RAM actually does
RAM is your computer’s short-term workspace. Every open app, browser tab and background process uses some of it. When RAM fills up, the operating system moves data to the SSD (called swap or page file). SSDs are fast, but far slower than RAM, so the system starts to stutter: switching apps takes a moment, tabs reload, and exports slow down. More RAM doesn’t make a computer faster when it isn’t full — it prevents it from becoming slow when you push it.
Why needs have grown
- Browsers isolate each site for security, so dozens of tabs can use several gigabytes.
- Desktop apps built on web technology (chat, collaboration and note-taking apps) each carry their own browser engine.
- Video calls with background effects are memory- and processor-hungry.
- On-device AI features in modern operating systems and apps reserve memory for models.
Unified memory on Apple silicon
On Macs with M-series chips, the CPU and GPU share one pool of “unified memory”. That’s efficient — graphics don’t need a separate copy of data — but it also means the GPU uses part of the same memory your apps need. macOS manages memory well, so a Mac often feels smoother than a Windows laptop with the same amount, but it doesn’t make 8 GB equal 16 GB. Apple now ships its Macs with 16 GB as the starting point, which tells you where the baseline is.
Unified memory is also why high-memory Macs are popular for running local AI models: the model can use most of the memory pool as graphics memory, something laptops with smaller dedicated GPUs can’t match.
How much for your work
| Use | Recommended |
|---|---|
| Browsing, email, streaming, documents | 16 GB (8 GB works but ages badly) |
| Office work with many tabs and video calls | 16 GB |
| Photo editing (Lightroom, Photoshop) | 16–32 GB |
| Web or mobile development with Docker and simulators | 24–32 GB |
| 4K video editing | 32 GB |
| Virtual machines, large codebases, data science | 32–64 GB |
| Local LLMs, 8K video, large 3D scenes | 64 GB or more |
| Gaming (Windows) | 16 GB system RAM, 32 GB for heavy modern titles plus streaming — GPU VRAM matters too |
How to check if you need more
- macOS: open Activity Monitor → Memory. If “Memory Pressure” is often yellow or red, you’re short on memory.
- Windows: open Task Manager → Performance → Memory. If “In use” is regularly above about 85–90% and “Committed” exceeds your installed RAM, more memory will help.
Can you upgrade later?
All current Macs and most thin laptops have soldered memory — what you buy is what you keep. Some Windows laptops, many gaming laptops and designs like the Framework Laptop still use removable memory modules (SO-DIMM or CAMM). Desktops are almost always upgradeable. If a laptop allows upgrades, buying a lower configuration and adding memory later can save money; check the maximum supported capacity and module type first.
RAM or storage — where to spend?
If you can only upgrade one on a non-upgradeable laptop, choose memory. Storage can be expanded with a fast external SSD or cloud storage; memory can’t. Choosing a laptop? See our guides to laptops for programming, laptops for students and the MacBook Pro M4.