A workstation can have a fast processor and a capable graphics card yet still feel slow when it runs out of RAM. Applications pause, preview playback stutters, renders compete with other tasks, and the system begins using much slower storage as temporary memory. Knowing how to size workstation memory starts with a simple question: what must this computer keep open, active, and responsive at the same time?
For professional work, memory should be sized around the entire workflow, not one application’s published minimum requirement. A video editor may have an editing application, media cache, browser, audio tools, proxy generator, and collaboration software running together. An architect may have a CAD model, rendering engine, reference files, and dozens of browser tabs open. The right answer depends on that real-world workload.
What workstation memory actually does
RAM is the system’s active workspace. It holds the operating system, open applications, project data, previews, textures, simulations, and temporary working files that need fast access. When there is enough RAM available, your processor and GPU can spend more time doing useful work instead of waiting for data.
When RAM fills up, Windows moves less-active data to the page file on the storage drive. Fast NVMe storage makes this less painful than it was years ago, but it is still dramatically slower than memory. This is why a system can appear responsive while opening a project, then become sluggish once editing, rendering, or multitasking begins.
More RAM does not automatically make every task faster. If a workflow is limited by CPU cores, GPU performance, storage speed, or network bandwidth, adding memory beyond what the workload uses will not change much. The goal is not to buy the largest number on a spec sheet. It is to avoid memory becoming the constraint that holds back the rest of the workstation.
How to size workstation memory by workload
Start with the largest project you work on regularly, not the occasional small file. Then account for concurrent applications and future project growth. For many professional users, these ranges provide a sensible planning baseline:
| Workflow | Practical starting point | When to move higher | |—|—:|—| | Office, business apps, light photo work | 32GB | Large spreadsheets, many virtual meetings, extensive multitasking | | Photography and 2D design | 32GB to 64GB | High-resolution panoramas, layered PSD files, batch exports | | HD and 4K video editing | 64GB | Multicam timelines, After Effects, large codecs, background exports | | 6K, 8K, motion graphics, live production | 128GB | Complex compositions, multiple applications, virtual production workflows | | CAD, BIM, architecture, and GIS | 64GB to 128GB | Large assemblies, point clouds, detailed maps, rendering alongside design work | | 3D animation, VFX, and rendering | 128GB | Heavy scenes, high-resolution textures, simulations, local rendering | | Scientific computing, photogrammetry, AI, and virtualization | 128GB to 512GB+ | Large datasets, multiple VMs, in-memory processing, large model development |
These are planning ranges, not universal rules. A photographer working with 24-megapixel RAW files may be well served by 32GB. A photographer stitching hundreds of high-resolution drone images into a panorama or processing extensive batches may need 64GB or more. Two people can use the same software and have completely different memory requirements because their source files, project scale, and working habits are different.
Video, motion graphics, and live production
Video work frequently needs more memory than customers expect because several memory-hungry processes often run together. The editing timeline is only part of the load. Footage decoding, frame caching, color tools, audio processing, media management, graphics applications, and web-based review tools all take their share.
For straightforward 4K editing, 64GB is generally the comfortable professional baseline. It provides breathing room for an NLE, a motion graphics application, and normal multitasking. Moving to 128GB makes sense for long-form productions, multicamera projects, 6K or 8K media, larger After Effects compositions, or workflows where editors routinely render while continuing to work.
Live production has its own pattern. Switching software, graphics playback, recording, streaming, remote guests, and browser-based control tools can all be active at once. A system built for a show that cannot pause deserves capacity margin, not a configuration designed around bare minimum use.
CAD, 3D, GIS, and visualization
Design and visualization workloads often grow gradually, which can hide a memory problem until a project becomes urgent. A small CAD assembly may run well on 32GB, but a detailed building model, complex mechanical assembly, GIS dataset, or point cloud can change the picture quickly.
For professional CAD and BIM work, 64GB is a strong starting point. For large assemblies, detailed Revit models, GIS analysis, photogrammetry, or visualization work with rendering tools open, 128GB is often the more dependable choice. 3D artists should also consider the difference between viewport work and final production scenes. Large texture sets, particle simulations, fluid caches, and CPU rendering can require substantially more system memory than basic modeling.
GPU memory matters here too, but it is not a substitute for system RAM. The GPU uses its own VRAM for many display, rendering, and AI tasks. A workstation needs both sized correctly for the applications involved.
AI, research, virtual machines, and data-heavy work
Technical workloads can use memory in ways that are less visible than media applications. A research tool may load a large dataset into memory. A development system may run containers, databases, local models, and IDEs together. A virtual machine receives an assigned portion of the host’s RAM, so several VMs can exhaust a 64GB workstation quickly.
For these jobs, 128GB is often the practical entry point, with 256GB, 512GB, or more appropriate for larger datasets and multi-user or virtualized environments. Capacity planning should include the data size, how much must remain in memory, whether processes run in parallel, and whether the workload is expected to scale over the system’s useful life.
Check your current system before choosing a number
If you are replacing or upgrading a workstation, measure actual memory pressure during normal work. In Windows Task Manager, review memory usage while your largest project is open and you are working the way you normally work. Do not test only when the system is idle.
Sustained usage near the installed capacity, frequent paging activity, or a noticeable slowdown when switching applications are signs that additional RAM can help. A brief spike is not always a problem. The concern is a machine that repeatedly has little available memory during the work that matters.
It also helps to ask a few workflow questions before setting a target: What is the largest file or project? Which applications stay open together? Do you render, export, or process data while continuing other work? Will project resolution, camera formats, dataset sizes, or team demands increase over the next few years? These answers are more useful than a generic recommendation from a software vendor.
Capacity comes first, then configuration
Once you have determined the capacity, the memory configuration matters. Workstation platforms commonly perform best with matched memory modules installed in the channels supported by the motherboard and processor. Mixing different capacities, speeds, or kits can create compatibility issues or cause the system to run at a lower speed.
Memory speed and timings affect performance, but they are secondary to having enough capacity. For a production workstation, 64GB of compatible, properly configured RAM is usually far more valuable than 32GB of faster memory that forces the system to page to storage. Reliability should take priority over chasing small benchmark gains.
Error-correcting code, or ECC, memory is worth considering for systems where data integrity and uptime carry extra weight. Scientific computing, engineering, servers, virtualization, research, and mission-critical business work can benefit from ECC-capable platforms. It is not required for every creative workstation, but it is a design decision that should be made alongside processor and motherboard selection because support is platform-specific.
Leave room for growth, but do it intelligently
A workstation should not be built so tightly that a modest project increase requires replacing everything. At the same time, filling every memory slot on day one is not always the best approach. A configuration with open slots can provide a cleaner upgrade path, provided the initial capacity is still sufficient for the work now.
The trade-off is platform dependent. Some systems need particular module arrangements to achieve their intended memory-channel performance. Others have limited slots or are better served by installing the final capacity from the beginning. This is one reason generic online memory calculators can miss the mark: they do not know your application mix, motherboard limits, processor memory channels, or upgrade plans.
For a workstation that supports billable creative work, production schedules, or critical technical tasks, memory planning is insurance against avoidable delays. Sandia Computers approaches that decision by looking at the software, project files, storage, GPU needs, and how the system will be used together. No guessing, and no forcing a professional workflow into a one-size-fits-all configuration.
The best memory choice is the one that lets you open the next project with confidence, keep working while background tasks run, and spend your day creating, designing, analyzing, or delivering instead of waiting on a progress bar.