Buy Plex headroom for the workload that will actually grow: concurrent transcodes, library expansion, backup windows, or new clients—not an arbitrary CPU tier.
Start with a measured baseline
Count peak simultaneous viewers, then separate direct play from video and audio transcodes. Record codec, resolution, subtitles, remote bitrate, and whether the client mix is stable. A community hardware discussion shows why a stream count without the direct-play/transcode split can mislead buyers (multi-stream workload example).
Keep the reused computer as the zero-new-hardware baseline if it passes the real test. Upgrade only when a measured bottleneck repeats during the busiest hour.
Choose the minimum compute and media path
Direct play makes storage and network reliability more important than raw CPU. Transcoding makes hardware-video support, driver compatibility, and thermal headroom the first gate. Choose a CPU or iGPU that leaves room for the host OS, database activity, and one unexpected transcode rather than sizing to a marketing stream number.
Buy the minimum platform that clears today’s peak with a documented margin. If remote users routinely force 4K-to-1080p conversion, use a transcoding-capable path or maintain an intentionally compatible remote library instead of assuming more storage solves compute.
Price storage and backup growth separately
Estimate media growth per month, retention for downloads or recordings, and the size of the Plex database and artwork cache. Add a backup destination and its write window; mirrored disks improve availability but do not create an independent backup. Headroom is useful only when the storage role, network path, and recovery path can grow with it.
Use one upgrade trigger and one stop rule
Upgrade when two conditions occur together: peak transcoding or backup work exceeds the acceptance margin, and the next six-to-twelve-month workload is known. Do not upgrade for a speculative codec or an unused drive bay. Stop shopping when the baseline clears peak use, the backup finishes inside its window, and the next expansion can be named as an added component rather than a full rebuild.
Who should buy more—and who should not
Buy more headroom if household users, remote clients, or a planned second service make the peak workload predictable. Stay with the baseline if most clients direct play and the current server has verified thermal and backup margins. A compact option such as ZimaBoard 2 fits only after those thresholds are measured; it is not a substitute for a missing media or backup plan.
Buying Guide
More to Read

How to Choose a Home Server for Jellyfin and Kodi
Kodi can reduce Jellyfin transcode demand when clients Direct Play well, so size the server from fallback conversion, storage, network, and shared services.

How to Choose SSD, HDD, and Backup Capacity for Jellyfin
Size Jellyfin storage by role: SSD for active app data and scratch, HDD for media capacity, and independent backup space for retained recovery points.

Before Buying a Jellyfin Server: Can Your Old PC Pass the Workload?
Reuse an old PC only after it passes the real Jellyfin workload, power, noise, storage, and recovery checks a new server would need to...

