On this page
The RealValidator Team Editorial team

The Nakamoto Coefficient is the smallest number of validators whose combined delegated stake exceeds one-third of the total network stake — the threshold at which a coordinated subset could halt consensus on a proof-of-stake network. On Solana, the Nakamoto Coefficient at epoch 971 (2026-05-15) is reported above 20 (per Solana Beach network statistics), though the exact value drifts each epoch as stake moves between validators.

The metric is a single-number summary of stake concentration at the consensus-halting threshold; it is not a complete decentralization score.

Why one-third

Proof-of-stake consensus protocols including Solana's TowerBFT require greater than two-thirds of stake to confirm blocks. A coalition holding more than one-third can therefore prevent block confirmation by refusing to vote, halting the chain. The Nakamoto Coefficient counts the smallest number of validators whose combined stake crosses that one-third mark — the minimum coalition size required to halt liveness.

The figure is computed by sorting validators by active stake, descending, and counting from the top until the cumulative share exceeds 33.34% of total active stake. The count at that point is the coefficient.

What it measures, and what it does not

It measures: stake concentration at the consensus-halting threshold. A higher coefficient means more validators would need to collude to halt the network — a measure of how distributed stake is across operators.

It does not measure:

  • Geographic concentration. Two validators in the same datacenter count as two validators in the coefficient. The metric does not weight by physical location.
  • ASN concentration. Two validators on the same network provider count as two. Network-level concentration (see the ASN entry) is a separate signal.
  • Operator concentration. A single operator running multiple validator identities under separate vote accounts counts as multiple validators. The metric does not detect sybil-style operator concentration.
  • Client diversity. Two validators running the same software (e.g. both on Jito-Solana — see the Jito disambiguation) count as two; a client-bug fault that affected both would still halt the same fraction of stake.
  • Smaller-coalition censorship. Halting requires >33%; selective censorship of specific transactions can be attempted with smaller coalitions and is not captured.

In short, the Nakamoto Coefficient is one dimension of decentralization. Reporting it alongside ASN concentration, city/country distribution, and client diversity is the standard practice for a complete picture.

Solana's value relative to other networks

A higher Nakamoto Coefficient is better — more validators required to halt. Cross-chain comparisons published by Messari and academic decentralization studies place Solana's coefficient in the 20+ range as of 2026, in line with or higher than several major proof-of-stake chains.

The figure is meaningful in absolute terms (a coefficient of 1 would mean a single validator could halt the chain) but the practical interpretation of a value above ~15 is dominated by the other concentration dimensions listed above. A network with a Nakamoto Coefficient of 25 but 80% of stake on a single ASN has a lower effective decentralization than the headline number suggests.

How it changes

The coefficient moves with stake migration each epoch. New large validators entering the network or existing large validators losing stake (commission changes, delinquency, exclusion from stake pools — see the stake pool entry) shift the coefficient. A single high-stake validator losing 1% of network share to a smaller validator can change the coefficient by ±1.

For stakers, the takeaway is that delegation to a validator outside the top-coefficient cohort directly improves the network's coefficient. A staker delegating to a mid-stake validator instead of a top-stake validator pushes some stake out of the halting coalition, raising the coefficient over time.

Sources

Coefficient value cited: Solana Beach network statistics, epoch 971, 2026-05-15. The figure changes each epoch as stake migrates; verify on a fresh snapshot when citing.