We have an aging server in the closet. Do we still need it, or can everything move to the cloud?

The server question deserves a specific answer, not an ideology. Start with what the machine actually does, because 'the server' is usually several jobs in one box: file shares, user logins, an application database, printing, maybe old email remnants. Each job has its own best home. Files and email have mature cloud destinations (SharePoint/OneDrive or Google Drive) that eliminate hardware, enable remote access, and remove the single point of failure. User management can move to cloud identity. Many applications now run as web services by subscription.

What keeps servers alive legitimately: line-of-business applications that still require a local database (plenty of practice-management, manufacturing, and industry software does), imaging systems with heavy local storage needs, and situations where internet dependence is unacceptable. If that describes one of the server's jobs, the answer is often a smaller, newer, properly backed-up server (or a hosted instance) carrying just that workload, with everything else migrated off. The all-cloud office is real for some businesses and premature for others; the mixed model is the honest mainstream.

What you should not do is wait for the aging box to decide. An old server carrying business-critical jobs is a countdown clock: parts, patches, and support all get worse while the eventual failure gets more expensive. A migration assessment (inventory the jobs, price the cloud paths, plan the sequence) turns a future emergency into a scheduled project, and it is one of the most common first projects a new IT provider runs for exactly this reason.