LearnBitcoin

Glossary

Block Propagation

The process by which newly mined blocks are transmitted among Bitcoin nodes worldwide.

Block propagation is how a newly-mined block travels from the miner who found it to every other node on the network. Speed matters: until a peer hears about the new block, they're potentially still mining or building on the previous tip, and the longer that takes, the higher the chance that a second miner finds a competing block and one of the two ends up a stale block.

The protocol-level optimizations:

  • Compact Blocks (BIP 152). Instead of sending the full block, the sender transmits the 80-byte header plus a 6-byte short ID for each transaction. The receiver rebuilds the block from transactions it already has in its mempool and requests only the ones it is missing. Bandwidth drops to a small fraction of a full block, and blocks arrive sooner as a side effect. In Bitcoin Core since 0.13 (August 2016).
  • High-bandwidth mode. A node asks up to three peers to push compact blocks to it unsolicited, as soon as the header checks out and before the block has been fully validated. Lower latency, at the cost of some extra bandwidth from those peers.
  • FIBRE and similar relay networks. FIBRE is a patchset on Bitcoin Core that combines UDP transport, forward error correction and compact blocks to move blocks between its nodes at close to the speed of light in fiber. Matt Corallo's original network, the successor to his Bitcoin Relay Network, sat dormant for most of a decade before Localhost Research revived a public FIBRE network and invited others to run their own.
  • Bitcoin Satellite. Blockstream's satellite broadcast (covered in Bitcoin Satellite) broadcasts the blockchain around the world for free, a redundant non-internet path for block data.

Why all the engineering: every additional second of propagation delay adds to the chance of a stale block, roughly in proportion, for whichever miner is downstream. That is lost revenue, and in stale-block races the largest mining operations hold an asymmetric advantage. The protocol level keeps improving so the playing field doesn't tilt entirely toward the best-connected.

How fast it is: in 2013, before compact blocks, a measurement covering the reachable network found a median of 6.5 seconds for a node to receive a new block, with a mean of 12.6 seconds. The KIT network monitor in Germany, which records when each of its peers announces a new block, saw half of the announcements arrive within about 0.8 seconds of the first and 90% within about 2.5 seconds in October 2026; the slowest 1% took more than ten seconds.

Key takeaways

  • Miners broadcast newly found blocks to peers
  • Efficient relay methods cut bandwidth and stale-block rates
  • Rapid propagation maintains global network consensus

Corrected

  • Was: The entry said that in 2026 a new block typically reaches about 80% of reachable nodes within 5-8 seconds and substantially everyone within 30 seconds, and listed Falcon alongside FIBRE as dedicated relay networks run by miners that push blocks across the globe in tens of milliseconds.

    Now: No measurement supports those figures. The KIT network monitor, which records when its peers in Germany announce a new block, saw half of the announcements within about 0.8 seconds of the first and 90% within about 2.5 seconds in October 2026; the 2013 measurement across the reachable network (median 6.5 seconds) is given for comparison. Falcon was removed, since its site no longer exists, and the FIBRE description now follows the FIBRE project's own page, which says the original network sat in stasis for almost a decade before a public network was revived. Details

Sources (6)

Related terms (12)

Published . Revised . One correction logged for this page.

Edit this page on GitHub