I'm due for a major upgrade and want something that will stay responsive for several years. My workload includes large spreadsheets, running local AI models, video editing, and an excessive number of browser tabs. I care more about reliability and performance than aesthetics or brand names.
I've been comparing custom PCs, prebuilt workstations, gaming systems, and Macs, but the advice is all over the place. Some people recommend prioritizing CPU cores, others emphasize single-core speed, and some say a high-end gaming PC is the best value. I'm especially unsure how much RAM and GPU VRAM I should target for local AI and editing.
What are people actually buying for this kind of workload, and what specifications would help prevent the system from feeling outdated within six months?
5 Answers
A modern high-end desktop CPU with roughly 12–16 cores should handle this mix well. More cores help with rendering and other production tasks, while strong single-core performance keeps spreadsheets and everyday work responsive. A gaming-oriented PC can be a good creator machine if it has enough RAM, storage, cooling, and a GPU suited to your AI models.
A high-memory Mac Studio or MacBook Pro is also worth considering if you’re comfortable with macOS. Apple silicon can be very capable for video work and general productivity, but local AI compatibility and memory behavior differ from a Windows/Linux PC. Check whether your specific AI tools support the platform before choosing it.
Start with benchmarks for the exact applications you use rather than general workstation rankings. Video-editing software, spreadsheet workloads, and AI tools can favor different hardware, so application-specific testing is much more useful than simply counting cores or choosing the newest part.
A single high-end GPU is usually simpler than trying to run two cards. Choose based on the model sizes and software you actually plan to use: GPU memory matters more for local AI than headline gaming performance. For a single powerful GPU, plan on a quality power supply around 850W or higher, along with fast NVMe storage and plenty of cooling.
The main balancing act is GPU VRAM. If you want to run anything beyond small local AI models, less than 32GB of VRAM can become restrictive quickly. Video editing may also favor particular CPU features depending on the software, so check its hardware recommendations. For everything else, 64GB of system RAM is a sensible target, especially with large spreadsheets and many browser tabs.
Related Questions
Lenovo Thinkpad Stuck In Update Loop Install FilterDriverU2_Reload