Ethereum's Glamsterdam hard fork — the network's largest protocol change since the 2022 Merge — entered public testnet on August 20, 2026, when the Platåberget network activated the full rule set. The upgrade ships ten Ethereum Improvement Proposals (EIPs) under Meta EIP-7773, headlined by enshri...
"In Glamsterdam, Ethereum is getting ePBS, which lets proposers outsource to a free permissionless market of block builders. This ensures that block builder centralization does not creep into staking centralization." — Vitalik Buterin, Ethereum Co-Founder
Ethereum's Glamsterdam hard fork — the network's largest protocol change since the 2022 Merge — entered public testnet on August 20, 2026, when the Platåberget network activated the full rule set. The upgrade ships ten Ethereum Improvement Proposals (EIPs) under Meta EIP-7773, headlined by enshrined Proposer-Builder Separation (EIP-7732) and Block-Level Access Lists (EIP-7928). Together, the changes target a path to 200 million gas per block, roughly 3.3x the current 60 million ceiling, and lay groundwork for approximately 10,000 TPS-equivalent throughput on Layer 1.
Sepolia testnet activation is scheduled for September 28, 2026, at epoch 351232. A mainnet date has not been set, though developers now target Q4 2026 after the original H1 timeline slipped due to ePBS implementation complexity. The Ethereum Foundation's April 2026 Checkpoint #9 acknowledged that ePBS was "proving to be trickier than anticipated," as the mechanism requires every part of the consensus stack to reason about partial blocks and two-party coordination. A stable multi-client devnet has been running since the Soldøgn interop concluded on May 2, 2026.
The economic implications are significant. Approximately 90% of Ethereum validators currently run MEV-Boost, a third-party relay system that two builders — Titan and Quasar — dominate with roughly 73% of all block construction. ePBS replaces this middleware dependency with a protocol-native mechanism, while the gas repricing EIPs restructure how the network prices state growth — a prerequisite for raising throughput without making state bloat unsustainable.
Glamsterdam bundles ten EIPs under the Meta EIP-7773 umbrella. Two are structural overhauls; the remaining eight are gas repricings, opcode additions, and protocol-layer improvements. The full list:
| EIP | Description | |-----|-------------| | EIP-7732 | Enshrined Proposer-Builder Separation (ePBS) | | EIP-7928 | Block-Level Access Lists (BALs) | | EIP-8037 | Cost Per State Byte pricing; separate state_gas_reservoir | | EIP-8038 | Recalibrated state-access gas for EXTCODESIZE, EXTCODECOPY | | EIP-8159 | eth/71 Block Access List network exchange | | EIP-7954 | Maximum contract size raised from ~24 KiB to 32 KiB | | EIP-7708 | Standard logs for non-zero ETH transfers and burns | | EIP-8024 | Backward-compatible SWAPN, DUPN, EXCHANGE opcodes | | EIP-7997 | Deterministic factory predeploy at address 0x12 | | EIP-2780 | Intrinsic gas cost reduction for ETH transfers (up to 71%) |
The package touches block production, execution throughput, state pricing, contract deployment limits, and the peer-to-peer protocol simultaneously. According to CoinDesk, developers described the scope as "probably the largest fork we've had since the Merge."
EIP-7732 formalizes the separation between proposers (who select the consensus block) and builders (who assemble the execution payload) directly into the protocol. The mechanism extends the execution payload propagation window from approximately 2 seconds to approximately 9 seconds, giving builders more time to construct optimized blocks while removing trust assumptions from third-party relays.
The current state of block production illustrates why the change matters. Approximately 90% of validators run MEV-Boost, which routes block construction through external relays. According to mid-2026 data, the top two builders (Titan and Quasar) construct roughly 73% of all blocks. Including the third and fourth largest builders (Eureka and BuilderNet) brings that figure above 95%. MEV-Boost adds 10–30% on top of base APR for validators, creating strong economic incentives against local block building.
ePBS does not solve builder concentration. Buterin has acknowledged this directly, stating that "ePBS only prevents builder centralization from spilling over into the staking layer, but the issue of builder centralization itself remains." The mechanism is a firewall, not a fix. Buterin's post-Glamsterdam roadmap includes FOCIL (Fork-choice enforced inclusion lists, confirmed as the lead candidate for the Hegotá upgrade) and encrypted mempools as subsequent countermeasures.
From an economic value perspective, ePBS shifts MEV relay infrastructure costs from an off-protocol expense — currently dominated by Flashbots' MEV-Boost — to a protocol-native function. Validators no longer need to trust relays to honestly report bid values or deliver payloads. The economic surplus that relay operators currently capture becomes contestable within the protocol itself.
EIP-7928 mandates that every block include a complete map of all accounts and storage locations touched by its transactions. This metadata enables execution clients to process unrelated transactions in parallel rather than sequentially.
The performance implication is direct. According to Thirdweb's technical analysis, a block that currently takes 12 seconds to validate sequentially could complete in approximately 3 seconds across four parallel execution lanes. This parallelism is what makes a higher gas limit feasible without proportionally increasing hardware requirements for node operators.
The 200 million gas limit is a design target, not a value the fork enforces. Validators set the limit through standard gas-vote signaling, currently coordinated around 60 million. The community raised the limit from 30 million to 60 million during 2025, and Fusaka's EIP-7935 standardized 60 million as the default. Glamsterdam's BALs and state pricing changes are designed to make a further increase to 200 million technically sustainable.
At 200 million gas, proponents estimate approximately 10,000 TPS-equivalent throughput under realistic workloads, up from a current effective rate of roughly 1,000 TPS on the base layer. For context, rollups currently process about 1,270 user operations per second versus 20.4 on mainnet, according to Q2 2026 data.
On August 17, 2026, the Ethereum Foundation issued a developer advisory: any wallet, tool, or dApp that hardcodes the 21,000 gas limit for ETH transfers will malfunction after Glamsterdam activates.
EIP-2780 replaces the flat 21,000-gas charge with separate costs based on actual resource consumption. Under the new model:
The repricing reflects actual resource costs. Creating a new account writes new state to disk, which is substantially more expensive for the network than transferring between existing accounts. Under the current flat pricing, new-account creation is subsidized by all other transactions.
EIP-2780 also reduces intrinsic gas costs for transfers between existing accounts by up to 71%, meaning common transactions become cheaper while state-expanding transactions become more expensive. The shift aligns incentives: applications that create sustainable economic activity on existing infrastructure pay less; applications that expand state footprint pay proportionally to the burden they impose.
EIP-8037 introduces Cost Per State Byte (CPSB) pricing and creates a separate state_gas_reservoir, independent of the standard gas limit. The mechanism targets a sustainable state growth rate of 120 GiB per year. Without this constraint, a 200 million gas limit would produce approximately 380 GiB per year of state growth — a rate that would push node hardware requirements beyond commodity levels within 2–3 years.
The May 2026 finalization of EIP-8037 was, according to multiple developer accounts, the final piece that gave client teams confidence to support a 200 million gas limit. By capping state growth independently of execution throughput, the protocol can increase transaction capacity without proportionally expanding storage requirements.
EIP-8038 complements this by recalibrating gas costs for state-access opcodes (EXTCODESIZE, EXTCODECOPY) to reflect actual disk I/O performance. The combined effect: gas pricing becomes a more accurate signal of real computational and storage costs, reducing cross-subsidy between transaction types.
This economic rebalancing addresses a structural problem. In Q2 2026, applications on the Ethereum network produced $1.79 billion in fees, but the base layer settled only $88.4 million — 4.9% of the total, according to COINOTAG. Meanwhile, blob-fee burning over a seven-day period in late July 2026 totaled approximately 0.22 ETH, negligible against a 0.85% annual supply growth rate. Glamsterdam's repricing does not directly solve the L1 revenue question, but it does ensure that the state costs L1 bears are at least priced correctly.
The development timeline has proceeded through three stages:
Soldøgn Interop Devnet — Concluded May 2, 2026. Multi-client devnet with all ten EIPs running. Validated end-to-end ePBS execution across "almost all clients," according to the Ethereum Foundation.
Platåberget Public Testnet — Launched August 17, 2026; Glamsterdam fork activated August 20. Approximately 50,000 validators across roughly 50 nodes with different client combinations. Open to developers, validators, and general users for testing.
Sepolia Testnet — Scheduled for September 28, 2026, at 14:44:48 UTC (epoch 351232, slot 11,239,424). Six client teams have agreed on the date. During All Core Devs Consensus (ACDC) calls, discussions included potentially adding an extra week of testing, with final confirmation pending a subsequent call.
Mainnet — No date set. Developers target Q4 2026. The upgrade slipped from its original H1 2026 timeline due to ePBS complexity.
EIP-7954 raises the maximum smart contract size from approximately 24 KiB to 32 KiB, and maximum initcode size from 48 KiB to 128 KiB, changes that will take effect on testnet before mainnet validation.
While Glamsterdam occupies current development resources, the scoping for its successor — Hegotá — is already underway. As of August 17, 2026, 66 EIPs are under consideration, with developers planning to narrow the list over the next several core dev calls.
Only one Hegotá change has been approved so far: FOCIL (Fork-choice enforced inclusion lists), which allows a committee of validators to force pending transactions into blocks, boosting censorship resistance. FOCIL is the direct follow-up to ePBS — where ePBS prevents builder centralization from affecting staking, FOCIL prevents builders from selectively excluding transactions.
The privacy proposals represent a structural shift. EIP-8141 (Frame Transactions), along with EIP-8272 (Keyed Nonces), would reduce infrastructure needed for private payments on Ethereum while leaving ordinary ETH transfers transparent. The proposals aim to make privacy a protocol-level feature rather than an application-layer add-on.
Hegotá is expected to ship in 2027. The combination of FOCIL and Frame Transactions would address the two largest structural gaps that Glamsterdam leaves open: censorship resistance at the transaction level and user privacy.
Glamsterdam ships 10 EIPs under Meta EIP-7773, headlined by ePBS (EIP-7732) and BALs (EIP-7928). It is described as Ethereum's largest protocol change since the Merge.
ePBS replaces MEV-Boost middleware with a protocol-native mechanism. Currently, 90% of validators use MEV-Boost, and two builders construct 73% of all blocks. ePBS removes the relay trust assumption but does not solve builder concentration.
The 200M gas limit target is enabled by BALs (parallel execution) and state repricing (EIP-8037). The current 60M limit would approximately triple, targeting 10,000 TPS-equivalent throughput.
The 21,000-gas rule breaks. Transfers to new addresses will cost up to 204,600 gas. Wallets, tools, and dApps that hardcode the current flat fee will malfunction.
State growth is capped at 120 GiB/year via a separate state_gas_reservoir, preventing unconstrained storage expansion at higher gas limits.
Sepolia testnet activation is scheduled for September 28, 2026. Mainnet is targeted for Q4 2026 but has no confirmed date.
Hegotá scoping has begun with 66 proposals under review. FOCIL (censorship resistance) is the only confirmed inclusion. Privacy features including Frame Transactions (EIP-8141) are leading candidates for 2027.
Glamsterdam attempts to resolve a fundamental tension in Ethereum's architecture: the network cannot meaningfully increase throughput without simultaneously restructuring how it prices state, builds blocks, and processes transactions. The ten EIPs are interdependent — ePBS without BALs does not justify a gas increase; BALs without state repricing creates unsustainable storage growth; state repricing without the 21,000-gas rule change creates internal inconsistency.
The economic stakes are substantial. In Q2 2026, Ethereum's base layer captured only 4.9% of the $1.79 billion in application fees generated on its network. Rollups process 62x more user operations per second than mainnet. Glamsterdam does not directly address this value leakage, but by tripling L1 capacity and correcting gas pricing, it changes the cost calculus for applications deciding between L1 and L2 deployment.
The risk profile is also elevated. Breaking the 21,000-gas assumption affects every wallet, exchange, and smart contract that interacts with ETH transfers. The ePBS propagation window extension from 2 to 9 seconds changes timing assumptions across the entire MEV supply chain. These are not incremental changes.
Whether Glamsterdam delivers on its throughput targets will depend on validator signaling after mainnet activation. The protocol provides the ceiling; the community decides whether to reach it.