protocol
EIP-8037 hits Last Call: state-creation gas costs up 5–7x ahead of Glamsterdam
EIP-8037 raises Ethereum's gas cost for new accounts, storage slots, and bytecode, and meters state creation in its own dimension. Buterin defends it as unblocking scaling.
EIP-8037, "State Creation Gas Cost Increase," is in Last Call with a deadline of November 1, 2026. It introduces a new unit — gas per state byte (CPSB = 1,530) — and separates state creation into its own gas dimension, so the cost of growing Ethereum's state stops competing with the single transaction gas limit. The proposal was authored by Maria Silva, Ansgar Dietrichs, Carlos Perez, Jochem Brouwer, Łukasz Rozmej, Anders Elowsson, Francesco D'Amato, Dragan Rakita, Paweł Bylica, Spencer Taylor-Brown, and Gary Rong. Full text at eips.ethereum.org/EIPS/eip-8037. Vitalik Buterin defended the proposal in an article shared Monday.
The numbers
EIP-8037 reprices operations that cause the state trie to grow. Key constants from the EIP:
- CPSB: 1,530 gas per state byte.
- New account: 120 bytes × CPSB = 183,600 gas (today:
GAS_NEW_ACCOUNT= 25,000). - New storage slot: 64 bytes × CPSB = 97,920 gas (today:
GAS_STORAGE_SET= 20,000). - EIP-7702 authorization base: 23 bytes × CPSB = 35,190 gas.
- Contract code deposit: moves from 200 gas per byte to CPSB per byte — a ~7.65x increase.
SYSTEM_MAX_SSTORES_PER_CALL: 16 (cap on new storage slots set per call).TX_MAX_TOTAL_GAS_LIMIT: 2³² − 1 (4,294,967,295).
Buterin's defense uses slightly different illustrative numbers — new-account creation at 197,900 gas, a 24 KiB deployment at roughly 37.8M gas — which track the EIP's formula on a 150M-gas reference block.
What the proposal targets
The design target is ≈120 GiB of state growth per year at a 150M block gas limit. Buterin's argument, summarized from his post:
- Execution capacity scales faster than state. Letting both grow at the block-gas-limit rate would push state growth to roughly 1,100 GiB/year under a 10x limit increase, versus roughly 110 GiB today.
- State growth burdens every node operator who has to carry the trie. Repricing state is the primary lever that lets the block gas limit rise without forcing bigger validators.
- Lowering gas costs on other operations would be more disruptive than raising state costs. The direction of travel is upward on state, flat or down elsewhere.
The second structural change is the new gas dimension. State creation becomes its own meter, so spending on new accounts or storage does not eat into the same 30M or 60M block gas limit used for execution. That is what makes the sharp per-byte increase tolerable in practice — a deployment pays its state cost separately from its execution cost.
Risks called out
Buterin acknowledges breakage risk. Contracts with hardcoded gas limits — including some bridges and relayers that forward .call with fixed gas — can fail when the callee's cost rises above the hardcoded number. The same applies to wallets, gas estimators, and indexers that assume today's GAS_NEW_ACCOUNT and GAS_STORAGE_SET.
The EIP's authors and Ethereum devs have been running contract replays against the new costs in the Glamsterdam repricing testing series (started August 2026) to identify affected contracts. Many are fixable by raising a call's forwarded gas; some will need redeploys.
What to watch
- EIP-8037 status after Last Call. The deadline is November 1, 2026; the Last Call period is where authors lock numbers before Final. Any amendments at this point should be minor.
- Companion EIPs. EIP-8037 is part of the Glamsterdam repricing bundle. EIP-8007 is the parent meta-EIP; related items in the bundle cover auth repricing (EIP-7702 interactions), SSTORE semantics, and the block-gas-limit raise that the repricing is designed to unblock.
- Glamsterdam activation date. EIP-8037 is scheduled to ship with Glamsterdam. No mainnet date is set yet; testnet work is ongoing. Developers with contracts affected by state-cost changes should run their test suites against the Devnet-3 repricing fork.
- Bridge audits. Any bridge forwarding a hardcoded gas value to a destination-chain callback is in scope. Ethereum Foundation's
evm-repricing-cost-analysistooling flags many of them; the long tail of custom integrations will not be flagged automatically.
Context
Glamsterdam is the upgrade after Fusaka and Pectra. Its central bet is that Ethereum's execution capacity can scale several times — 3x throughput is a commonly cited figure in the repricing call notes — without turning the state trie into a bottleneck. EIP-8037 is the price Ethereum pays to get there: a sharp jump in the cost of growing the chain's state, in exchange for a bigger execution envelope on top.
The change is not cosmetic, and it is not a parameter tweak. A deployment that currently costs 5M gas may cost nearly 38M after Glamsterdam, meter in a separate dimension, and still fit under a raised block gas limit. That is the design, and it is what every deployer touching state in 2027 will have to budget for.