Decentralization Index
3.6out of 10
| Measure | Weight | Score |
|---|---|---|
| Nakamoto coefficient | 25% | 6.3 |
| Full nodes | 15% | – |
| Client diversity | 10% | 0.5 |
| Hardware to run a node | 10% | 2.0 |
| Token concentration | 15% | – |
| Insider allocation at launch | 15% | 0.8 |
| Control over upgrades | 10% | 6.0 |
How the index works · method v1.0
Insider allocation at launch
0.8Very high insider share
Insiders 49% Insider-controlled 49% Public 1.6%
| Group | Counted as | Share |
|---|---|---|
| Seed Sale | Insider | 16% |
| Founding Sale | Insider | 13% |
| Validator Sale | Insider | 5.2% |
| Strategic Sale | Insider | 1.9% |
| CoinList Auction Sale | Public | 1.6% |
| Team | Insider | 13% |
| Foundation | Insider-controlled | 10% |
| Community | Insider-controlled | 39% |
Shares of: Launch supply of 488,634,933 SOL: the 500 million SOL created at genesis on March 16, 2020, less 11,365,067 SOL that the Solana Foundation burned in May 2020 after community concerns about tokens lent to market makers (Binance Research; Helius)
Insider lockups: Under two years
Shares come from Binance Research's report of April 7, 2020, which lists the same eight groups and percentages most sources repeat. Its SOL amounts per round (for example, 79,290,466 SOL in the seed sale at 16.23%) match a base of about 488.6 million SOL. So the shares are of supply after the Foundation's 11.365 million SOL removal, which we leave out as burned. Totals: insiders 49.00%, insider-controlled 49.35%, public 1.64% (99.99% because of rounding). Classification: Helius describes the seed, founding, strategic and validator rounds as private SAFT sales (simple agreements for future tokens, contracts that deliver tokens later) held before the public auction, so all four count as insider. This includes the validator round, under our rule for sales open only to selected buyers. The CoinList auction counts as public. It was a Dutch auction (the price drops until all tokens are bid for, and every winner pays the same final price) of 8 million SOL. It was open to buyers outside the US and other jurisdictions CoinList didn't serve. Community (38.89%) counts as insider-controlled: Binance Research says these tokens 'are held by the Swiss Foundation, which is run by an independent board' and are used for 'bounties, incentives programs, marketing, and grants.' The Foundation's own 10.46% also counts as insider-controlled. Lockups (per Helius): all SAFT investor tokens unlocked about nine months after the March 2020 auction. Founders received 50% then and the rest gradually over the following 24 months. Non-founding employees received 100% then, with undisclosed selling restrictions. Most insider tokens were free to sell within a year, so we class vesting as short. A small amount from Binance airdrop campaigns also entered early circulation and isn't shown separately. Coins created later through inflation aren't part of the launch supply.
All measures
| Measure | Value | As of | How we got it |
|---|---|---|---|
| Nakamoto coefficient Fewest block producers who could together disrupt the network. | 14 14 operators together hold just over one third of active stake (434.7 million SOL), per validators.solutions (updated November 3, 2026), after merging validators run by the same company. We merged #1 Figment and #7 Ledger by Figment, which carries Figment's name and which Ledger offers through its collaboration with Figment (https://www.ledger.com/stake-your-solana-sol-with-complete-confidence-using-ledger-validator). Without that merge the figure would be 15. Counted by validator, the figure is 18. We also merged validators that carry the same company's name in Solana Compass's validator table (https://solanacompass.com/validators): Bitwise Onchain Solutions (five, including Ledger by Bitwise), Coinbase (six), P2P.org (three), Kiln (two) and Staking Facilities (two). Stake for validators outside validators.solutions's top 10 comes from that table, whose amounts match validators.solutions for the top 10. Together the 14 operators hold about 146.4 million SOL (33.7%). The largest, Figment, held 6.0% across its two validators (3.9% and 2.1%). Some large validators publish no name, including #5 and #6 (2.2% each), so we couldn't match them to an operator. If any are run by the same company as each other or as a listed operator, the figure would be lower. Conflicts: Chainspect showed 17 the same day (https://chainspect.app/chain/solana), and nakaflow.io showed 6 without explaining its method (https://nakaflow.io/). Since Alpenglow, finalization needs 60% of stake to respond (80% for the fast path). Anza says the design tolerates 20% adversarial plus 20% non-responsive stake (https://www.anza.xyz/blog/alpenglow-a-new-consensus-for-solana). So finality stalls if more than 40% of stake goes offline, and safety assumes less than 20% adversarial stake. Our method uses one third of stake for every proof-of-stake network, whatever its protocol's own thresholds, so figures compare like for like. Grouped the same way, it would take 21 operators to hold more than 40%. | November 3, 2026 | Researched |
| Block producers Validators, miners or pools that produce blocks. | – Left blank: we found no published 30-day count of validators that produced at least one block. validators.solutions listed 647 validators scheduled for leader (block production) slots in epoch 1,076 (https://validators.solutions/en/validators/leader-slots/, November 3, 2026). That covers one epoch, about 26 hours since slots shortened to 200 ms (https://solanacompass.com/news/solanas-first-full-epoch-at-the-200ms-slot-time-target-averaged-2183ms-and-lasted-262-hours), and a scheduled leader can still skip its slots. Solana Compass counted 694 validators with active stake in the same epoch, and Chainspect showed 665 (https://chainspect.app/chain/solana). Some operators run more than one validator, so this overstates the number of independent producers. | – | – |
| Full nodes Independent computers that check every rule. | – Left blank: we found no public crawler that counts reachable Solana full nodes. Solana Compass counted 694 validators with active stake in epoch 1,076 (November 3, 2026). Every validator is a full node, but non-voting RPC nodes (servers that answer apps' requests but don't vote) aren't counted anywhere we could open, so 694 would only be a lower bound. | – | – |
| Largest client share Nodes running the most popular software. Lower is safer. | 86% Counted from Solana Compass's validator table, read November 3, 2026, around 15:00 UTC (the table is live, so later counts differ slightly). Of the 670 ranked validators, 89 showed a Firedancer-based client (Firedancer, Harmonic Firedancer or Firedancer BAM) and 18 showed no client. The other 563 ran Agave or builds made from Agave's code: Agave, Agave BAM, Harmonic Agave, Jito-Solana and Rakurai. That is 86% of the 652 validators with a named client (84% if the 18 with no client are kept in the count). We count these as one implementation because they share Agave's consensus and runtime code, so one bug could hit all of them. Counted separately, the largest single build would have a lower share. Conflicts: wenfiredancer.com showed only 7 of 693 validators on Firedancer (0.65% of stake) the same day (https://wenfiredancer.com/). But its version labels (0.1.x) don't match Firedancer's current v26.10.x releases, so it looks out of date. By stake, SOL Strategies put Firedancer and Frankendancer combined at about 14% in April 2026 (https://solstrategies.io/blog/firedancer-what-sol-strategies-is-building). Context: Firedancer wasn't supported during the October 21, 2026 switch to Alpenglow, so operators used Agave-based clients for the migration (https://solanacompass.com/news/solana-alpenglow-is-live-on-mainnet-as-anza-retires-towerbft-at-slot-459653000). Firedancer v26.10.2 (October 14, 2026) added Alpenglow support, enabled in its default testnet and devnet configurations, and later mainnet releases (v26.10.3 to v26.10.5, October 20 to 27, 2026) improved it (https://github.com/firedancer-io/firedancer/releases). Galaxy wrote that Frankendancer would no longer be supported once Alpenglow was active (Solana Compass, October 21, 2026). | November 3, 2026 | Researched |
| Hardware to run a node What it takes to verify the network yourself. | Data-center server Anza's validator requirements: a CPU with a 2.8 GHz base clock or faster and SHA extensions (AMD Gen 3 or Intel Ice Lake or newer) and NVMe SSDs for accounts (1 TB or more), ledger (1 TB or more) and snapshots (500 GB or more), with accounts and ledger on separate disks recommended. ECC memory and a motherboard that takes 512 GB are suggested. Every validator needs a stable connection with a public IPv4 address, and a staked node needs 'at least 2 GBit/s symmetric connection', with 10 GBit/s recommended (an unstaked node needs 1 GBit/s). That bandwidth and the suggested server-class memory point to data-center hosting rather than a home computer. | November 3, 2026 | Researched |
| Who can change the rules Whether one party can upgrade or pause it. | Core team coordinates Client teams such as Anza (Agave) and Firedancer build new features, and each feature gate (a switch that turns a new feature on across the network) is controlled by a key held by the core contributor responsible for it (Helius). On mainnet, activations begin 'two full epochs after 95% of stake adopts the new minor version' (Anza's feature-gate tracker, edited November 2, 2026). Helius: 'A private key unique to each feature gate is owned by the core contributor associated with that feature gate', and 'validators retain full control over what runs on their infrastructure' (https://www.helius.dev/blog/solana-governance--a-comprehensive-analysis). Validators vote on major changes, weighted by the stake delegated to them. The Alpenglow vote closed on September 2, 2025, with 98.27% voting yes, per The Cryptonomist (https://en.cryptonomist.ch/2025/09/03/solana-approves-alpenglow-over-98-yes-votes-and-finality-reduced-to-approximately-150-ms-a-significant-technical-upgrade/). In August 2026, a stake-weighted vote by validators, in which holders of JitoSOL (a liquid staking token) could override their validator's vote, adopted a constitution (SGP-0001) and approved doubling the yearly decline in issuance from 15% to 30% (SGP-0002), in what Solana Compass called the network's first binding governance vote. Even so, the change only takes effect once client teams ship it and validators activate it (https://solanacompass.com/news/sgp-0002-passes-solana). We found no key that can pause the network or force an upgrade. Classed as coordinated. | November 3, 2026 | Researched |
| Held by top 100 addresses Excluding known exchanges and contracts where possible. | – Left blank: we found no top-100 figure that excludes exchanges and stake accounts. Most staked SOL sits in separate stake accounts, which makes a simple rich list hard to read. | – | – |
| Transactions per second 30-day average of user transactions, not consensus votes. | 1,813 4.70 billion non-vote transactions in September 2026, per Solana Compass, divided by 2,592,000 seconds. That total includes 1.52 billion transactions that reverted. Counting only completed transactions (3.18 billion) gives 1,227 TPS. Validator votes are excluded. Since Alpenglow went live on October 21, 2026, vote transactions no longer appear in blocks at all (https://solanacompass.com/news/solana-alpenglow-is-live-on-mainnet-as-anza-retires-towerbft-at-slot-459653000). September is the latest 30-day period with a published count; we found no total for October 4 to November 2, 2026. For comparison, Chainspect's last-hour figure read between 1,848 and 2,109 TPS across our readings on November 3, 2026 (https://chainspect.app/chain/solana). We don't use Chainspect's 65,000 TPS theoretical figure. | September 30, 2026 | Researched |
| Daily active addresses 30-day average. Addresses are not people. | – Left blank: we found no 30-day average from a primary dashboard we could open (Artemis, Blockworks and Dune load their numbers with scripts). Crypto Briefing (October 9, 2026) reported 1.88 million unique active addresses (SolanaFloor and Blockworks data) and separately about 4.27 million daily active addresses from an unnamed tracker (https://cryptobriefing.com/solana-active-addresses-13-month-high/). It warned the two aren't the same measurement. | – | – |
| Typical fee Median for a simple transfer over 30 days. | – Left blank: we found no 30-day median fee for simple transfers. Known figures: the minimum fee is 0.000005 SOL (5,000 lamports), per Solana Compass's fee tracker. Solana Compass's fee tracker (https://solanacompass.com/statistics/fees) showed an average non-vote fee of about 0.000048 to 0.000049 SOL over the 24 hours before each of our readings on November 3, 2026, with about 68% to 69% of user transactions paying a priority fee (an optional extra fee to be processed sooner). Chainspect showed a 1-day average fee of $0.0032 (https://chainspect.app/chain/solana). The tracker says trading bots pull the average up by paying extra to get their trades in on time, while most human users pay low or no priority fees. | – | – |
| Time to finality Until a payment practically cannot be reversed. | – Left blank: Solana switched consensus to Alpenglow on October 21, 2026, and we found no independent measurement of mainnet finality since then. Anza expects 'about 150 ms (median)', based on simulations that leave out computation overhead (https://www.anza.xyz/blog/alpenglow-a-new-consensus-for-solana). Cointelegraph reported that Solana's official Alpenglow dashboard showed 6.6 seconds at press time on the day of the switch (https://cointelegraph.com/news/solanas-alpenglow-launches-mainnet-99-faster-finality). Chainspect lists 400 ms with no date or method (https://chainspect.app/chain/solana). Anza puts TowerBFT's finality at about 12.8 seconds, a figure that no longer applies. Re-measure once dashboards or RPC providers publish mainnet finality data. | – | – |
| Outages, last 24 months Times the network halted or stopped producing blocks. | 0 Solana's official status page lists no incidents after February 6, 2024 (its feed was built on November 3, 2026). That outage stopped block production for about 5 hours, but it falls outside the window of November 3, 2024 to November 3, 2026. We found no reports of a halt or a stop in finalization in 2025 or 2026, including during the switch to Alpenglow on October 21, 2026. | November 3, 2026 | Researched |
| New coins per year Yearly issuance as a share of supply. | 3.6% Annual inflation rate shown by Solana Compass in epoch 1,076. The schedule started at 8% in February 2021, falls 15% a year and stops falling at 1.5% (Helius). In August 2026 validators approved doubling the yearly decline to 30%, but it wasn't active when we checked. This is gross issuance: it doesn't subtract the fees that are burned. For comparison, Crypto Briefing reported about 3.82% as of June 2026 (https://cryptobriefing.com/solana-fees-record-inflation-cuts/), in line with a 15% yearly decline. | November 3, 2026 | Researched |
See how Solana stacks up
Compare it side by side with up to three other networks.
Comparing networks isn't advice to buy their coins. Spotted an error? Email us with a source and we'll correct it.