Its 3.3 million $ADA fee total amounted to about one $ADA for every 150 $ADA in staking rewards. Reserve emissions supplied the dominant share of the reward economy during the period.
Cardano’s epoch 655 supply page offers a shorter official snapshot. It attributed 108,500 transactions and 33,855 $ADA in fees to completed epoch 654. Spread across five days, that was about 21,700 transactions per day, or 0.251 transactions per second. Fees equaled about 0.339% of roughly 9.998 million $ADA in distributed rewards for the epoch.
The one-epoch snapshot uses a different reward denominator from Bitquery’s 12-month staker total, so the percentages are not directly interchangeable. Each window nevertheless places transaction fees at well below 1% of its respective reward measure.
The long-window transaction count also points to weaker demand. Cardano averaged 90,294 transactions per day in 2022. From January through August 2026, the average was 24,869, a decline of 72.46%.
Activity composition changed at the same time. Bitquery classified wallets sending at least 3,000 transactions in a month as bots unless their behavior resembled an exchange. Under that method, bots’ share of transactions rose from 11.5% in 2022 to 32.8% in 2026, with batchers forming the largest identified bot subgroup.
Those counts describe on-chain actions rather than unique people. One bot can submit thousands of transactions, while a decentralized exchange batcher can process orders for many customers. A modern Cardano wallet can use multiple addresses tied to one stake key, and a holder who stakes $ADA without moving it remains absent from a count of sending wallets.
The transaction decline still matters for fee revenue because every action creates an opportunity to pay a fee. It cannot reveal how many people left, remained active or delegated their coins.
Bitquery also measured a decline in the share of circulating $ADA staked, from 75.6% at the end of 2022 to 58.3% in the last epoch of its study period. That is a participation measure. The cited evidence contains no direct security-outcome metric, so it cannot support a claim that network security has already deteriorated.
The reserve creates runway as its contribution declines
Cardano’s epoch 655 data showed 6,126,859,027 $ADA left in reserves, equal to 13.62% of the network’s 45 billion $ADA maximum supply.
The reserve is designed to shrink. Under Cardano’s monetary policy, transaction fees and 0.3% of the remaining reserve enter a virtual pot each epoch. The treasury receives 20% of that pot, and the balance is available for stake rewards, subject to pool performance. Unclaimed rewards remain in the reserve.
Applying a fixed percentage to the remaining balance produces exponential decay. Cardano’s documentation describes a reserve half-life of roughly four to five years, without setting a definitive exhaustion date.
That declining reserve contribution changes the reward side of the equation. Nominal reward outlays can fall as emissions decline, which would reduce the fee revenue needed to match them. A lower target would still leave the network dependent on real economic activity if fees are to replace a larger share of rewards.
Cardano’s fee structure provides another variable. Current minimum fees combine a fixed component with a charge based on transaction size, and protocol governance can change those parameters. Higher revenue per transaction would narrow the gap with less traffic, although the price of block space can also influence demand.
Holding average fees constant shifts the calculation toward activity. Scaling the 2026 average of 24,869 daily transactions by the current 149.6 reward-to-fee ratio produces about 3.72 million transactions per day, or roughly 43.1 transactions per second.
That is a simplified gross scenario calculated before the 20% treasury allocation. It is a translation of the measured gap, rather than a forecast or a precise break-even point.
The 43.1 TPS illustration broadly aligns with CryptoSlate’s Sept. 8 model, which estimated that replacing reserve-funded rewards could require roughly 36 to 50 sustained transactions per second. That model placed a central estimate near 45 TPS after the treasury cut.
Linear Leios is designed for throughput above that range. The proposed CIP-164 specification models sustained capacity above the simplified 43 TPS scenario, giving Cardano a plausible technical path to process the necessary volume.
Leios test results leave the demand question open
Test results address only the capacity side. Cardano reported roughly sixfold Leios performance in an August public testnet update, using synthetic traffic. The result showed that the design could process more load under test conditions. It supplied no evidence that mainnet users would generate enough activity to multiply fee revenue by roughly 150.
Deployment also remains ahead. Intersect’s Dijkstra planning document targets code completion in the fourth quarter of 2026, excluding Preview and pre-production testing and governance time from that schedule. The mainnet hard-fork date remains undetermined.
The economic equation has several moving parts. More transactions and higher average fees increase the revenue side. Declining reserve emissions and lower reward outlays reduce the amount that fees would need to replace. Leios enlarges Cardano’s processing capacity, but adoption decides how much of that capacity generates revenue.
For now, the measured distance is stark: 3.3 million $ADA in fees against 493.7 million $ADA in rewards. Leios may remove a technical ceiling, while Cardano’s larger test is attracting enough paid activity to turn capacity into durable network income.
