Pectra is an upgrade to the Ethereum blockchain that was activated on mainnet on 7 May 2025. Its name joins Prague, the changes to the execution layer that processes transactions and smart contracts, and Electra, the changes to the consensus layer that runs proof of stake. Both parts went live together as a single hard fork, so every node operator had to update its client software. The execution-layer half on its own is referred to as the Prague hard fork, following the convention of naming execution-layer upgrades after cities that have hosted Devcon.
Pectra followed the Dencun upgrade of March 2024 and bundled a larger set of changes than most earlier forks. Its main themes were making ordinary user accounts more flexible, simplifying life for large staking operators and increasing capacity for rollups. Each change was specified as an Ethereum Improvement Proposal (EIP).
The naming follows a convention used since 2023: execution-layer upgrades are named after cities that have hosted Devcon, the Ethereum developer conference, and consensus-layer upgrades after stars. Earlier pairs were Shanghai and Capella, which enabled staking withdrawals in 2023, and Cancun and Deneb, which together formed Dencun. The contents of each upgrade are agreed by client developers in regular public calls, then tested on short-lived developer networks and public test networks before a mainnet date is set.
Smart accounts: EIP-7702
The most visible change for users was EIP-7702. It lets an externally owned account, the standard type controlled by a single private key, point to the code of a smart contract and run it as if the account were a contract. It introduced a new transaction type carrying a signed authorisation for this delegation. The result is that an existing address can gain features associated with a smart wallet, such as batching several actions into one transaction, letting a third party pay gas, or using spending limits and session keys, without moving funds to a new address.
EIP-7702 builds on earlier account abstraction work, including ERC-4337, and the two can be used together. It also created a new risk: a user tricked into signing a delegation to a malicious contract can hand over control of the whole account, so phishing campaigns quickly began targeting such signatures and wallets added warnings.
Staking changes
Several EIPs changed how validators work:
- EIP-7251 raised the maximum effective balance of a validator from 32 ETH to 2,048 ETH. The minimum to activate stays at 32 ETH, but large operators can consolidate many validators into fewer, and rewards above 32 ETH can compound within one validator rather than being skimmed off.
- EIP-7002 allows exits and partial withdrawals to be triggered from the execution layer using the validator's withdrawal credentials, which matters for staking arrangements where the withdrawal address is a smart contract.
- EIP-6110 moved the processing of new validator deposits onto the execution layer, shortening the delay between a deposit and its recognition by the consensus layer.
Together these changes reduced the operational load on large staking pools and made staking contracts easier to design, while leaving solo staking with 32 ETH unchanged.
Data and blob capacity
EIP-7691 increased the number of blobs each block can carry, raising the target from three to six and the maximum from six to nine. Blobs were introduced by proto-danksharding in Dencun to give rollups cheap, temporary data space, and higher capacity was meant to ease rising rollup demand. At the same time, EIP-7623 raised the cost of calldata for transactions that are mostly data, nudging bulk data into blobs and limiting the worst-case size of blocks.
Other changes
EIP-2537 added precompiled contracts for operations on the BLS12-381 curve, making it cheaper for smart contracts to verify BLS signatures and some zero-knowledge proofs. EIP-2935 stores recent historical block hashes in a system contract so they can be read from state, which helps light clients and future stateless designs. Further EIPs dealt with how attestations and validator requests are structured.
Preparing Pectra took longer than planned. When it was activated on the Holesky and Sepolia test networks in early 2025, configuration problems in some clients disrupted both networks, and developers launched a new test network, Hoodi, to rehearse the upgrade again before scheduling mainnet activation. Some proposals discussed for the same upgrade, including a new contract code format, were moved to later forks. The next upgrade on the roadmap focused on data availability sampling to increase blob capacity further. That upgrade, Fusaka, activated on mainnet on 3 December 2025 and introduced peer data availability sampling (PeerDAS).
External links
Ref 1070001351