Development workloads have a characteristic shape: long idle stretches punctuated by a compile or a container stack coming up, both of which want memory and sustained processor power rather than peak burst speed. A modest Docker Compose stack with a web service, a database, a cache and a worker claims four to eight gigabytes and several cores before the editor has opened a file.

That makes memory the first ceiling most developers meet. Sixteen gigabytes suits front-end and scripting work; thirty-two is the realistic baseline once containers, a local database, a browser full of tabs and a full IDE run together. The processor question comes second, and the architecture question, x86 or ARM, arrives with conditions attached.