The Problem: Nostr’s resilience comes at a high cost in network resources. The strategy of replicating posts across multiple relays to ensure availability generates massive data overhead, making the network inefficient and expensive to maintain.

The Data: A large-scale study measured 17.8 million posts replicated 616 million times. This means that, on average, each single post is stored on 34.6 different relays. To retrieve posts, clients often query multiple relays in parallel, causing huge bandwidth waste: 98.2% of retrieval operations are redundant (they fetch identical copies of a post already obtained from another relay).

Practical Implication: This model is unsustainable at very large scale. It consumes resources (storage, bandwidth, CPU) disproportionate to the information transmitted, driving up operational costs for relay operators. This inefficiency is a structural driver pushing towards recentralization, as only large, well-funded relays can afford to handle such a load.