Direct answer

Minecraft bandwidth per player varies with chunk sends, movement, entity updates, world changes, voice chat and proxy layout. A constant per-player estimate is unreliable.

How to check it

  1. Check whether the uplink is measured in Mbps or MB/s, whether the host charges for transfer, and whether a monthly egress cap applies.
  2. At the hosting interface or Linux level, record sent and received bytes before and after a representative busy period using a provider graph or ip -s link.
  3. Compare new joins, rapid exploration, idle players and events separately: chunk delivery can be much heavier than a steady idle session.
  4. Use spark to verify that a network bottleneck is not actually tick lag. Test proxies and voice-chat traffic as separate components where applicable.
ip -s link
/spark tps

What to change

Budget for peak bursts as well as monthly transfer, leaving room for backups and network overhead. If bandwidth is constrained, reduce avoidable chunk-send pressure only after measuring its effect. Do not publish a universal per-player bandwidth figure as a requirement.

Software and version notes

The same player count can use very different bandwidth across versions and server configurations. Network compression settings have CPU trade-offs and should be adjusted only with a measured reason.

Verify the result

Measure peak outgoing Mbps and cumulative GB transferred through multiple typical sessions; confirm players can explore and join without stalls.

References