How Mina keeps a blockchain under 22KB

Web3September 2, 20263 min read

This is the first piece in a deep dive series on Mina Protocol. The original version of this article is published on Medium — head there for the full piece. The rest of the series, starting with recursive zk-SNARKs, continues here on the blog.

How does Mina keep its blockchain so small?

Instead of storing the full chain, a Mina node holds a constant-sized cryptographic proof — historically described as 22KB, though Mina's own updated technical breakdown puts the core figure closer to 11KB, depending on what a given node needs to do with it (a light client verifying state has different requirements than a block producer).

That number stays the same size whether the network is a day old or ten years old.

What's the problem with normal blockchains?

One characteristic of blockchain technology is decentralization: the network runs on independent nodes rather than a single central authority. Normally, that means every node has to store and verify the entire transaction history to trust the current state — and that history only grows.

How does Mina solve this?

Mina takes a different approach. It uses recursive composition of zk-SNARKs to represent its blockchain in a constant, succinct form. Instead of storing the full chain, a node holds a constant-sized proof that stands in for the validity of that entire history.

The next article in this series explains the mechanism that makes that possible: recursive zk-SNARKs.

What does the 22KB number actually mean?

  • It's not the blockchain data itself — it's a cryptographic proof that the blockchain is valid
  • It stays constant — whether the chain has 1,000 blocks or 1,000,000 blocks
  • It varies by node type — light clients need less (~11KB), block producers need more
  • It's a proof, not a copy — verifying it is equivalent to having verified the entire history

Why does this matter?

  1. Lower hardware requirements — anyone can run a node on modest hardware
  2. True decentralization — no need for trusted third parties or archive nodes
  3. Fast sync — new nodes sync in seconds, not days
  4. Mobile-friendly — light clients can run on phones

What's next?

Continue to Article 2: Recursive zk-SNARKs for the mechanism that makes this possible.