Transaction Fees Are Rising Up to 4x. FLOW Issuance Is Falling With Them, Marking a Major Self-Sustainability Step
Roughly 171,000 fewer FLOW will be minted per month at current network volume as Flow transaction fees rise. For most transactions, fees will go up 2–4x as part of recalibrating transaction fees to track real execution cost, so every unit of revenue offsets issuance directly.
Networks get more self-sufficient as they grow when costs are better covered by usage. On Flow, staking rewards are paid from transaction fees before any new FLOW is minted, so as activity increases, more of the network is funded by the applications and users transacting on it and less by new issuance. That’s an important step toward a more sustainable network, where growth in usage strengthens the network itself.
Every transaction on Flow pays a fee based on how much work it asks the network to do. This is often called ‘gas’ and is priced in compute units: the runtime counts each operation a transaction performs, and each operation carries a weight that converts that count into execution effort.
Those weights were last calibrated in FLIP-346 against synthetic testnet load, which was the most stable dataset available at the time. FLIP 370: Execution Effort Calibration III is its successor, fit against real mainnet data instead of simulations.
Measured against 3.39 million real mainnet transactions, the old model accounted for only ~15% of what actually made transactions cheap or expensive to execute, and under-predicted real execution cost by roughly 5x. The new model, fit on live mainnet data, accounts for 80%. Fees rise to close that gap: roughly 2–4x for most transactions. The average transaction will pay roughly 0.01 FLOW more, a step towards a more sustainable network, funded by usage instead of new issuance.
Cost-effective gas
Even at 2–4x, the average Flow transaction costs well under a tenth of a cent at current FLOW prices. Nothing changes about how you transact.
Where the extra fees go
Staking rewards on Flow are paid from collected transaction fees first. New FLOW is minted only to cover whatever fees can’t. Fees that reflect real cost therefore flow straight into the reward pool: at current volume, ~171,000 fewer FLOW minted every month. The network mints less, and real usage pays for more of what it does.
Consumer and financial apps generate millions of small transactions that make fee-funded network rewards work. On most networks, consumer-scale activity is a cost problem. Under this model, it’s the opposite: the busier Flow gets, the less FLOW it mints. Growth in usage and sustainability of the token are now the same trend.
What Flow builders should know
A small share of today’s transactions (~0.5%) would hit the network’s compute limit under the new pricing. Computation-heavy workloads such as long loops, number parsing, and signature aggregation will see the biggest increases. Once the new weights reach testnet: re-test your heavy transactions, check any hardcoded compute limits, and if you sponsor fees for your users, re-budget for the new per-transaction cost.
Read the proposal
Want to contribute to the discussion? The full FLIP, with all the data behind it, is on GitHub: https://github.com/onflow/flips/pull/369.
