What Was the Pectra Upgrade?

Pectra was the Ethereum hard fork that went live on 7 May 2025, combining Prague, the execution-layer upgrade, with Electra, the consensus-layer upgrade. Developers name each fork after a Devcon city and a star, then merge the two words, which is how the release got its compound name.

Measured by proposal count, Pectra was the biggest upgrade since the 2022 Merge, carrying eleven EIPs against six in the preceding Dencun fork. Rather than chasing one theme, it spread changes across three areas at once: how accounts behave, how validators stake, and how much data rollups can post.

Getting there was not smooth. The upgrade broke finality on the Holesky testnet in February 2025 and tripped up Sepolia shortly afterwards, so developers launched a new testnet, Hoodi, as a clean rehearsal. Once Hoodi forked without trouble, the Ethereum Foundation confirmed mainnet activation at epoch 364032.

The Eleven Pectra EIPs That Shipped

Pectra's eleven proposals fall into three groups: features that make accounts programmable, changes that improve staking and protocol plumbing, and tweaks that give rollups more room. Reading them by job rather than by number makes the upgrade easier to follow.

Account and Wallet Changes

EIP-7702 was the headline act. It lets a regular externally owned account sign an authorisation that points its address at a smart-contract implementation, borrowing that code for as long as the delegation stands. Batching, gas sponsorship, spending caps and passkey logins all become possible without abandoning the original address.

EIP-2537 is less visible but just as practical. It adds precompiled contracts for BLS12-381 elliptic curve operations, so verifying the signatures and proofs used by zero-knowledge rollups, bridges and staking protocols costs far less gas than earlier hand-rolled implementations. Developers building privacy and proof systems benefit most from it.

Staking and Protocol Plumbing

Six proposals reshaped how validators join, grow, exit and communicate with the execution layer. Together they remove much of the operational friction that made running Ethereum validators tedious for both solo stakers and large staking providers.

The staking and infrastructure EIPs that shipped in Pectra are:

• EIP-7251: Lifts the maximum effective balance from 32 ETH to 2,048 ETH for validators using new 0x02 compounding credentials, while leaving the 32 ETH entry minimum untouched.
• EIP-6110: Brings validator deposits onto the execution layer, shrinking the gap between sending a deposit and seeing it processed from roughly nine hours to about thirteen minutes.
• EIP-7002: Lets the withdrawal address trigger full exits and partial withdrawals directly, so whoever owns the staked ETH no longer depends on the operator holding the signing key.
• EIP-7549: Moves the committee index outside the signed attestation, letting identical votes aggregate more efficiently and trimming the data that every consensus client must gossip and process.
• EIP-7685: Defines a general request bus between the execution and consensus layers, the shared plumbing that deposits, triggered withdrawals and consolidation requests all travel through after Pectra.
• EIP-2935: Stores the last 8,191 block hashes in a system contract, giving stateless clients, light clients and cross-chain provers a cheap, verifiable record of recent chain history.

Blob and Data Changes

EIP-7691 doubled the blob target from three to six per block and raised the maximum from six to nine. Because the blob fee only climbs once usage exceeds the target, the extra headroom gave Layer 2 networks cheaper and more predictable data costs almost overnight.

EIP-7623 pulled in the opposite direction on calldata. It raised the price of data-heavy transactions that do little computation, nudging rollups towards blobs and capping worst-case block size so the extra blob bandwidth would not overload node operators running Ethereum on modest home internet connections.

EIP-7840 looked like housekeeping at the time. It moved blob targets and limits into a schedule inside client configuration files, which quietly paved the way for the blob-parameter-only forks that later raised capacity in Fusaka without requiring an entire hard fork for every single increase.

Pectra 18 Months On: How It Actually Played Out

Judged on delivery, Pectra did its job. Mainnet forked without consensus incidents, a sharp contrast with its troubled testnet run, and every one of the eleven proposals went live as specified. The surprises came later, in how users, attackers, stakers and markets responded to the new primitives.

Smart-account adoption scaled far faster than most developers predicted. Wallets such as MetaMask, Ambire, Trust Wallet and Safe built EIP-7702 upgrade paths, and by September 2026 more than 27 million Ethereum accounts had used a delegation, making it one of the fastest-adopted protocol changes Ethereum has shipped.

Staking changed shape just as visibly. Active validators fell from roughly 975,000 at the start of 2026 to about 863,000 by early October, even as staked ETH climbed to a record near 43.7 million, a pattern explained almost entirely by operators merging validators under EIP-7251.

Two outcomes were less welcome. Attackers adapted to account abstraction within days, turning delegations into an efficient draining tool, and ETH's price completed a full round trip from launch-week euphoria to a deep 2026 drawdown. Both stories, covered below, say more about crypto's wider ecosystem than about Pectra's engineering.

EIP-7702 and the Smart-Account Era

EIP-7702 changed what an Ethereum wallet can be. Instead of forcing users onto a separate contract wallet, it lets an existing address borrow contract logic, then revoke or replace it, keeping the account's history, tokens and DeFi positions intact.

What Builders Did With It

Payments were the first obvious win. Circle used EIP-7702 to enable gasless USDC transfers, letting users move stablecoins without first acquiring ETH for fees. Wallets bundled approve-and-swap flows into a single signature, and session keys made repeated interactions with games and trading apps far less tedious.

The bigger story is convergence with ERC-4337. Most production wallets now delegate to 4337-compatible implementations, so upgraded EOAs plug straight into existing bundler and paymaster networks. MetaMask made its Smart Account upgrade the primary path for existing users, while Safe released modules that bring multisig logic to delegated addresses.

The Security Fallout

Attackers moved faster than most builders. Within weeks of launch, Wintermute found that over 97% of mainnet delegations pointed to near-identical sweeper contracts, the most common nicknamed CrimeEnjoyor. These bots watch wallets whose keys are already leaked and drain any incoming funds almost instantly.

Peer-reviewed research has since put a number on it. A USENIX Security 2026 study of 3.66 million early authorisation transactions across seven chains linked 63% of them to attacker contracts, confirmed $2.36 million in losses, and flagged roughly $10 million sitting in older contracts that assumed every caller was a plain wallet.

The nuance is that most losses did not stem from a flaw in EIP-7702 itself. They still began with a phished signature or leaked seed phrase; what changed is how quickly a compromised wallet can be emptied, and how one malicious authorisation can bundle several drains together.

How to Stay Safe With EIP-7702 Delegations

Smart accounts are only as safe as the contract they delegate to, so a little discipline goes a long way. These habits cut most of the risk that the 2025 sweeper wave exposed.

Practical steps for anyone holding ETH in an upgradeable wallet:

1. Read prompts: Treat any request to sign a delegation or "upgrade your account" with the same suspicion as a seed-phrase request, and reject prompts arriving from unfamiliar sites.
2. Native upgrades: Only upgrade through your wallet's built-in smart-account flow, which delegates to audited implementations, rather than approving delegation targets suggested by decentralised apps or airdrop pages.
3. Check status: Look up your address on a block explorer to see whether it carries a delegation and which contract it points to, especially after connecting to unfamiliar dapps.
4. Revoke quickly: If a delegation looks wrong, reset it by delegating to the null address, then move funds to a fresh address if the key might be exposed.
5. Prune approvals: Old token approvals remain dangerous after an upgrade, so review and revoke unused allowances regularly; batched transactions make a single stale approval far easier to exploit.
6. Hardware signing: Keep long-term holdings on a hardware wallet that shows delegation details on its screen, so a compromised browser cannot silently swap the contract you approve.
7. Chain scope: Prefer authorisations bound to a single chain ID; a delegation signed with chain ID zero is valid on every EVM network that supports EIP-7702.

What Pectra Changed for Stakers

For validators, EIP-7251 carried the most practical weight. It ended the rigid 32 ETH ceiling that left rewards idle, and it gave large operators a way to shrink their validator fleets without exiting the beacon chain.

Compounding and Consolidation

Before Pectra, any balance above 32 ETH was swept to the withdrawal address and earned nothing until it could fund another full validator. Switching to 0x02 credentials keeps rewards inside the validator, where they compound automatically up to 2,048 ETH, and consolidation requests merge validators without passing through the exit queue.

The benefit is real but uneven. Research published in mid-2026 estimated that native compounding adds roughly 5% in relative consensus-layer yield for small balances, but less than 1% for the largest providers, which helps explain why some institutions consolidated slowly despite the obvious operational savings.

Lido, the largest liquid staking protocol, became the biggest test case. The protocol began folding more than 265,000 validators into roughly 4,000 larger ones, a move it projects will cut the network's validator count towards 628,000 and reduce attestation messages by about 29% per epoch.

Trade-Offs Stakers Should Weigh

Consolidation changes the risk profile of a validator as much as its yield. Anyone deciding whether to merge balances should weigh the operational savings against the concentration and liquidity trade-offs below.

The main trade-offs of consolidating validators after Pectra:

• Slashing penalty: Pectra cut the initial slashing penalty to 1/4096 of effective balance, roughly 0.5 ETH even for a full 2,048 ETH validator, making large balances less punishing.
• Correlated risk: Correlation penalties still scale with stake, so a single misconfigured key controlling thousands of ETH can now cause far larger losses during a mass slashing event.
• Network load: Fewer validators mean fewer signatures to aggregate each epoch, freeing bandwidth that developers can redirect towards faster finality and higher gas limits in future forks.
• Operator concentration: Consolidation does not by itself reduce the number of independent operators, but concentrating stake behind fewer keys raises the impact of any single client or infrastructure failure.
• Exit liquidity: Partial withdrawals under 2,048 ETH need a manual EIP-7002 request, and churn limits mean big consolidated validators can wait longer to exit during congested queues.
• Solo stakers: Small home validators gain the most relative yield from compounding, while flexible stake sizes above 32 ETH let them add capital without spinning up another machine.

How Pectra Reshaped Rollup and Blob Economics

Pectra's blob increase landed in a market that was already saturated. Rollups had been hitting the three-blob target for months, so doubling supply collapsed prices: Coin Metrics data showed daily blob spending falling from around $16,000 to fractions of a cent within a week, while usage rose about a third.

Cheap data was great for rollup margins and user fees on networks such as Base and Arbitrum, but it created a value-capture debate. With blob fees near the one-wei floor, very little ETH was burned from Layer 2 activity, feeding criticism that rollups were free-riding on Ethereum security.

Demand eventually caught up, with blocks regularly filling the new six-blob target by September 2025. Fusaka then answered both problems at once, adding PeerDAS for capacity, EIP-7918 to tie the blob fee floor to execution costs, and blob-parameter-only forks that lifted the target to 14 by January 2026.

In hindsight, EIP-7691 worked less as a final fix and more as a bridge. It bought rollups roughly seven months of cheap headroom while PeerDAS matured, and EIP-7840's blob schedule is what made the faster, incremental capacity increases of the post-Fusaka era technically far more straightforward.

How ETH Price Reacted to Pectra

Pectra arrived at a low point for sentiment. ETH had slid to a two-year low near $1,417 in early April 2025 and spent the first quarter badly trailing Bitcoin, with many traders and fund managers questioning whether upgrades still mattered to the asset's valuation at all.

The upgrade flipped that mood quickly. ETH rose more than 20% in the two days around activation, its biggest daily move since 2021, and then extended the rally through summer, reaching an all-time high near $5,000 in August 2025 as spot ETF inflows and treasury-company buying accelerated.

The round trip since then has been sharp. Recession fears and heavy selling pushed ETH below $2,000 by spring 2026, and despite a late-summer recovery it traded near $2,400 in late October 2026, roughly 38% lower than a year earlier, according to Fortune's daily price tracker.

The lesson is that upgrades shape capability far more reliably than price. Pectra's launch rally coincided with ETF demand and a broad risk-on rotation, and those same macro forces later reversed. Anyone trading around future forks such as Glamsterdam should treat activation dates as sentiment catalysts rather than reliable fundamental signals.

Where Pectra Sits in the Roadmap

Pectra was one stop on a longer route rather than a destination. It cleared technical debt and shipped user-facing wins, leaving deeper data and execution scaling to the forks scheduled after it.

How Pectra fits into Ethereum's upgrade timeline, from Dencun to Hegotá:

1. Dencun: Activated on 13 March 2024 and introduced blobs through EIP-4844, creating the separate data market for rollups that Pectra later expanded with more capacity.
2. Pectra: Went live on 7 May 2025 with eleven EIPs, delivering EIP-7702 smart accounts, 2,048 ETH validators, faster deposits and a doubled blob target of six.
3. Fusaka: Activated on 3 December 2025, adding PeerDAS so nodes sample blob data instead of downloading all of it, plus a dynamic blob fee floor under EIP-7918.
4. BPO forks: Two blob-parameter-only forks on 9 December 2025 and 7 January 2026 lifted the blob target and maximum to 10/15, then 14/21, without full hard forks.
5. Glamsterdam: Bundles enshrined proposer-builder separation and block-level access lists for parallel execution; it forked Sepolia on 6 October 2026, but mainnet remains undated after the Hoodi activation was postponed.
6. Hegotá: Expected around Q2 2027, it is headlined by FOCIL inclusion lists for censorship resistance and also schedules EIP-8141 frame transactions, extending account abstraction beyond EIP-7702.

Final Thoughts

Nearly eighteen months later, Pectra reads as a disciplined, foundational upgrade that delivered what it promised. Ethereum accounts became programmable without migrations, stakers gained compounding and simpler operations, and rollups received breathing room that carried them until Fusaka's PeerDAS arrived, all without a mainnet incident at the fork.

Its rough edges are just as instructive. The sweeper wave showed that powerful primitives attract abuse the day they ship, and the price round trip showed how weakly forks predict markets. For builders and stakers, though, Pectra's legacy is structural, and Glamsterdam and Hegotá are building directly on it.