Will Glamsterdam Actually Reduce Ethereum Fees?
Published 6/17/2026, 8:03:30 PM
Short answer: No — not in the way most users expect, and not across the board.
Glamsterdam is a throughput and efficiency upgrade, not a fee-cutting upgrade. The fee impact is bifurcated: it raises the ceiling on what the network can handle, which suppresses spikes during congestion, but it does not structurally lower the floor of fees at all times.
What Makes Fees Lower (or Less Volatile)
| Mechanism | Effect | Source |
|---|---|---|
| Gas limit increase to 200M (from ~60M, a 3.33× jump) | Directly suppresses base fees during high-demand periods by increasing per-block throughput capacity | ethereum.org/roadmap/glamsterdam |
| EIP-7904 repricing | Reduces costs for computation-heavy operations (e.g., KECCAK256 per-word drops from 6 to 1 gas) | EIP-7904 implementation notes |
| Fee volatility decreases | With a 3× larger ceiling, abrupt base fee spikes during congestion events should become less common | ethereum.org/roadmap/glamsterdam |
| L2s benefit indirectly | Cheaper L1 data blob posting (estimated ~40% reduction) lowers L2 settlement costs, feeding through to user-facing L2 fees | ethereum.org/roadmap/glamsterdam |
| Parallel execution (BALs) | Enables higher safe gas limits without overloading validators, improving sustained throughput | ethereum.org/roadmap/glamsterdam |
What Makes Fees Higher or Unchanged
| Mechanism | Effect | Source |
|---|---|---|
| EIP-8037 | Increases new account creation costs ~8–10× to target state DB growth of ~120 GiB/year | ethereum.org/roadmap/glamsterdam |
| EIP-7976 | Raises calldata floor costs, increasing costs for on-chain data-heavy applications | ethereum.org/roadmap/glamsterdam |
| Blob fees & priority fees | Determined by separate mechanisms (blob market + EIP-1559 competition) — not directly affected | ethereum.org/roadmap/glamsterdam |
| Base fees | Remain governed by EIP-1559 dynamics — demand-driven, not structurally reduced | ethereum.org/roadmap/glamsterdam |
Claim Verdict Table
| Claim | Verdict |
|---|---|
| "Fees will drop dramatically for everyone" | Incorrect |
| "Fee spikes during congestion will be less severe" | Supported |
| "State-write-heavy apps will pay more" | Supported |
| "L2s will benefit from cheaper L1 data posting" | Supported |
| "Capacity growth will outpace demand growth, suppressing fees long-term" | Uncertain — depends on demand trajectory |
Timeline Context
Glamsterdam targets H2 2026 for mainnet, with an internal working target of end of August 2026 (aspirational). The primary bottleneck is ePBS implementation complexity — cross-client parity across five execution-layer clients and five consensus-layer clients must be confirmed before a fork date is called. The June 2026 date originally circulating has already slipped to Q3 2026.
Bottom Line
Glamsterdam raises the ceiling of L1 throughput, not the floor. Users should expect fewer fee shocks during high demand rather than lower fees at all times. The upgrade is most directly beneficial to L2s (cheaper blob posting) and to applications that rely on computation-heavy operations (cheaper KECCAK256). Applications that write heavily to state (certain NFT minting patterns, new account creation) will face higher costs due to EIP-8037.
What remains open: No quantitative fee reduction metrics have been published for actual pre/post-upgrade fee levels. The long-term fee trajectory depends on whether demand growth outpaces the 3.33× capacity increase — a question that cannot be answered until mainnet activation and beyond.