Huge divergence in bandwidth allotment between app platform and regular droplets

In evaluating app platform, I am struck by the huge divergence between the reasonably generous outgoing bandwidth allotted cumulatively based on the number of “regular” droplets one pays for, vs. the very small and inelastic bandwidth allotment on the app platform.

In summary:

• if we pay for e.g. 5 of the cheapest droplets, we get 5,000 GB of outgoing bandwidth (and this applies if the droplets are part of a k8s cluster as well).
• if we have a single app platform deployment, it could utilize any number of droplets; its outgoing allotment would only ever be 100 GB outgoing

Can someone explain the rationale for this, especially in light of app platform nodes’ much higher cost vs. equivalent regular droplets?