Ethereum's largest hard fork by proposal count went live on May 7, 2025, as core developers, in the days surrounding activation, confirmed the scope of the next upgrade, shelved a long-debated EVM bytecode redesign, and restructured how protocol decisions get made.

Pectra, a combined execution and consensus layer upgrade, activated on mainnet at epoch 364032 (approximately 10:05 UTC). The fork bundled several significant changes: EIP-7702, which gives standard user wallets temporary smart contract capabilities; EIP-7251, which raised the maximum validator stake from 32 ETH to 2,048 ETH; and EIP-7691, which doubled the target blob count per block from 3 to 6 (with a maximum of 9). Blobs are the data packets Ethereum's base layer provides to Layer 2 networks for cheap transaction settlement. This mechanism was introduced by EIP-4844 in 2024, and EIP-7691's expansion of blob capacity builds directly on that foundation. Within one week of activation, more than 11,000 EIP-7702 authorizations were recorded on mainnet, indicating rapid adoption by wallet developers.

EOF Out, PeerDAS In

On April 28, 2025, nine days before Pectra activated, a developer session produced a significant cut to the subsequent upgrade. All 16 EIPs comprising EOF (Ethereum Object Format) were removed from Fusaka, the fork then targeted for late 2025. Developers specifically preferred an October 2025 deadline to avoid losing momentum during the holiday period, making "late 2025" a deliberate and meaningful target rather than a loose aspiration. EOF had aimed to restructure how EVM bytecode is formatted, introducing new opcodes and restricting certain runtime behaviors, including gas and code introspection, that make the execution environment harder to reason about. Ethereum developer Pascal Caversaccio had argued in a post on Ethereum Magicians that EOF would "introduce new semantics, overhaul opcodes, and significantly expand the attack surface." Developers concluded that unresolved disagreements over which variant of EOF to implement posed too much risk of delay. Tim Beiko noted in an April 28 post on X that developers found "technical uncertainty about [EOF's] impact" and that it "risked delaying the Fusaka rollout." The proposal may return for Glamsterdam, the fork currently scoped for 2026.

With EOF out, Fusaka's headline feature became PeerDAS (EIP-7594), a data availability sampling system already running on its sixth devnet at the time of the Checkpoint #2 update (with devnet-7 described as imminent at the time). PeerDAS allows nodes to verify that blob data is available on the network by checking only a small random sample rather than downloading full payloads. Using erasure coding, blobs are split into 64 columns and extended to 128 column subnets; a standard node downloads only 8 of them, roughly one-eighth of the total. This design theoretically supports up to eight times the current blob throughput while keeping node hardware requirements manageable. Gabriel Trintinalia, a Consensys engineer and Ethereum core developer, put the stakes plainly: "PeerDAS' importance was such that, during the initial development of the Fusaka upgrade, any feature that carried a risk of delaying the fork...was deprioritized and removed from the scope."

Incremental Blob Scaling Between Forks

Developers also reached broad agreement on a mechanism called BPO (Blob Parameter Only) forks. Rather than waiting for a full named hard fork to adjust blob capacity, BPO forks allow the network to raise blob limits through a minimal, targeted coordination process. Following Fusaka's activation, this approach played out in practice: a BPO fork on December 9, 2025 raised blob targets from 6/9 to 10/15 per block, and a second on January 7, 2026 moved them again to 14/21. The longer-term roadmap targets 48 blobs per block once PeerDAS is mature.

Lower Node Storage, Better Developer Coordination

EIP-4444, known as History Expiry, was entering its rollout phase alongside these developments, with Sepolia testnet support scheduled for May 1, 2025 and mainnet support to follow post-Pectra. The change allows execution clients to stop storing pre-Merge block data, which predates September 2022. Full Ethereum archive nodes currently require several terabytes of storage. History Expiry trims that by 300 to 500 GB, with Go-ethereum and Nethermind both supporting the feature by default in recent versions. Historical data remains accessible through the Portal Network and public archival endpoints.

On the process side, Tim Beiko, an Ethereum Foundation developer, proposed splitting All Core Developer calls into separate testing sessions for the current upgrade and scoping sessions for the next one. A new protocol research call series would run in parallel to front-load longer-term direction work. The structural change allows client teams to progress Pectra testing and Fusaka scoping simultaneously rather than alternating between them.

Why This Matters Outside North America and Europe

The cumulative trajectory of these upgrades carries outsized relevance for users in markets such as India, Nigeria, and Kenya. In those regions, most on-chain activity involves transactions in the $1 to $50 range: remittances, mobile payments, and small-scale DeFi activity. After EIP-4844 shipped in 2024, Arbitrum transaction fees dropped from roughly $0.37 to $0.012; Optimism fees moved from $0.32 to $0.009. Analysts expect that every subsequent blob throughput increase follows the same cost curve, applying comparable downward pressure to fees across Layer 2 networks. For most users in these markets, analysts note, $0.37 renders cross-border payments unviable; $0.012 does not, according to CoinLaw and Nasscom analysis of transaction viability thresholds in emerging markets.

History Expiry similarly matters in regions where high-end server infrastructure is expensive or unavailable. A 300 to 500 GB reduction in node storage requirements brings full node operation within reach for community operators running consumer hardware in South Asia and Sub-Saharan Africa. EIP-7702's account abstraction features, including sponsored transactions, batched calls, and session keys, address a persistent onboarding problem in mobile-first markets where seed phrases and manual gas management have historically been barriers.

What Comes Next

Developers have already named the post-Fusaka fork Glamsterdam, with scoping beginning in parallel. Preliminary targets include potential six-second block times and further gas limit increases, with a 2026 timeline. Marius Van Der Wijden, a core developer at the Ethereum Foundation, noted that blob increases following Fusaka would be gradual: "The improvements will take a few months to fully play out, since we will only slowly increase the blobs in order to make sure the network can handle the increased throughput safely." Total L2 TVL reached approximately $47 billion in October 2025, a figure that analysts expect will inform how aggressively developers push the scaling timeline through Glamsterdam.