Simplified Payment Verification (SPV)

Rather than trusting a third party, an SPV client proves inclusion cryptographically: it downloads only the chain of 80-byte block headers, roughly 50 to 60 megabytes for the entire history of Bitcoin, and asks a connected peer for a Merkle branch, the short path of hashes linking a specific transaction to the Merkle root stored in a header. If the hashes recompute correctly, the transaction is proven to sit inside that block without the client ever seeing the block's other contents.

This design trades some trust for a large reduction in resources. A full node validates every protocol rule against the complete blockchain; an SPV client instead assumes the longest valid header chain reflects honest consensus, leaning on proof of work rather than independent rule-checking. Early mobile wallets used BIP 37 bloom filters so peers could return relevant transactions without an exact address list, but researchers showed the filters leaked enough data to de-anonymize users, and Bitcoin Core stopped serving them by default in 2019. Newer light clients favor compact block filters (BIP 157/158), where full nodes publish one filter per block and clients check it locally instead of querying a server.

SPV remains attractive wherever storage or bandwidth is scarce, but it carries real trade-offs: a client connected only to dishonest peers can be fed a fabricated but internally consistent chain, and inclusion in a block is not the same as finality, since a transaction can still be reversed by a deep chain reorganization. Waiting for multiple confirmations reduces, though never eliminates, that risk.