Ethereum Users the Most Sticky, BNB Chain Leads in Absolute Numbers

Blockchain User Retention Rates

Across 11 major blockchains, Ethereum recorded the highest on-chain user retention rate of 26.2% in our Q1 2025 to Q1 2026 cohort study. This means that roughly 1 in 4 Ethereum users who were active in Q1 2025 (Jan 1, 2025 – Mar 31, 2025) were still transacting on the network a year later in Q1 2026 (Jan 1, 2026 – Mar 31, 2026). This makes Ethereum the most “sticky” chain in our study, outpacing every other blockchain measured, including newer chains that significantly outpace it in raw user numbers.

BNB Chain followed in second place with a 20.5% retention rate, retaining over 1.49 million users in absolute terms, the highest absolute retained user count of any chain we studied. Ronin, the gaming-focused blockchain behind Axie Infinity and Pixels, placed third at 19.1%, a notable result for a chain with a narrower use case, likely reflecting the habitual daily activity that on-chain gaming creates compared to more speculative DeFi usage.

At first glance, Solana appears to have low user retention rates at 7.9% losing more than 16M users in just a year. However this can be explained by the comparison periods of Q1 2025, where memecoins were at their peak, making this an unfair comparison. Nevertheless, Solana is 2nd in terms of the total number of users retained which we will cover below.

BNB Chain and Solana Lead on Absolute Number of Users Kept

Despite Ethereum’s superior retention rate, it does not top the leaderboard when measuring raw retained users. BNB Chain retained 1,494,233 users in absolute terms, followed closely by Solana at 1,394,873. Both blockchains are significantly ahead of Ethereum’s 682,240 retained wallets.

Base, Coinbase’s Ethereum Layer 2, retained 732,539 users built rapidly through memecoin activity and Coinbase’s consumer onboarding funnel, translated into a sizable retained user base. Aptos, despite being a newer non-EVM chain, also performed creditably, retaining 715,657 users from a cohort of 5.45 million, despite its lower 13.1% retention rate.

Ronin’s Gaming Ecosystem Punches Above Its Weight

Ronin’s 19.1% retention rate (#3) is arguably the most contextually interesting finding in the dataset. With a Q1 2025 cohort of 1.63 million and 311,604 retained users, Ronin outperforms every Ethereum Layer 2 in our study on retention rate despite having a fraction of their marketing budgets and ecosystem liquidity. This likely reflects a structural advantage: on-chain gaming creates daily behavioural loops that generate consistent transaction activity independent of market speculation. Where DeFi users return when prices move, gaming users return when they want to play.

Methodology

User retention rates were calculated using a cohort methodology applied to on-chain transaction data sourced from our Dune Analytics dashboard. A wallet was classified as “active” in Q1 2025 if it made 5 or more successful transactions between January 1 and March 31, 2025, on a given blockchain. Retention was measured as the share of those wallets that made at least one transaction on the same chain between January 1 and March 31, 2026.

Data sources used: gas.fees spell for EVM-compatible chains (Ethereum, BNB Chain, Base, Arbitrum, Optimism, Avalanche, Polygon, Ronin); solana.transactions for Solana; aptos.user_transactions for Aptos; sui.transactions for Sui. Bot activity is not filtered from any chain. Retention is measured on a same-chain basis — a wallet that migrated to a different chain in Q1 2026 is counted as churned from its origin chain.

The 5-transaction minimum threshold was applied consistently across all chains to exclude one-off visitors, airdrop claimers, and single-session users, while remaining low enough to preserve a representative cohort across chains with varying transaction costs and activity norms.

This study is for illustrative and informational purposes only, and is not financial advice.

Excluded Blockchains

Tron was excluded after data validation confirmed that top addresses exhibited automated transaction patterns inconsistent with human wallet activity, with individual addresses recording over 10 million transactions per quarter. Including Tron would measure infrastructure uptime rather than user retention.

TON was excluded for a structural reason specific to its architecture. TON’s actor model means the ton.transactions account field records every contract involved in a transaction, not just the human initiator. A single user action can generate dozens of internal messages across contracts, making it impossible to isolate genuine user wallets from the noise.

Mantle was excluded as the 5-transaction minimum threshold produced an unrepresentative cohort. Data validation showed the median Mantle wallet made only 1 transaction in Q1 2025, meaning our threshold captured a narrow slice of outlier behaviour rather than a representative sample of the chain’s user base.

HyperEVM is not yet indexed in Dune’s gas.fees spell and could not be included in this study.

Cardano, and other non-EVM chains outside of Solana, Aptos, and Sui either lack comprehensive transaction-level coverage on Dune or use account models that are not directly comparable to the wallet-based methodology used here.

Onchain Retention Rates Across 11 Blockchains: Q1 2025 Cohort