Losing a seed phrase means losing everything. A transaction requires the sending account to have ETH for gas. One wrong click approves a malicious contract. These are not edge cases. They are the normal failure modes of the wallet system most crypto users interact with every day.
Account abstraction is the attempt to fix that at the protocol level.
The Two Types of Ethereum Account
To understand account abstraction, you need to understand Ethereum’s account model.
Ethereum has two types of accounts. An externally owned account (EOA) is the standard wallet: a private key controls it, it can send transactions, and it has no logic of its own. A contract account is a smart contract: it has code, it can execute logic, but it cannot initiate transactions on its own. It can only respond to being called.
This split creates the problem. Because only EOAs can initiate transactions, users are stuck with whatever behavior an EOA allows. No programmable spending limits. No alternative signing methods. No way to pay gas with tokens other than ETH. No recovery option if you lose your key.
Account abstraction collapses this distinction. The goal is to let any account, including your wallet, be governed by arbitrary code.
ERC-4337: Abstraction Without Changing the Protocol
The first major implementation of account abstraction on Ethereum did not require a protocol change. ERC-4337, finalized in March 2023, introduced a parallel transaction system that runs on top of Ethereum rather than inside it.
Instead of submitting a regular transaction, users submit a “UserOperation” to a shared mempool. A new class of infrastructure providers called “bundlers” collect these operations, wrap them into a regular transaction, and submit them to the network. A singleton smart contract called the “EntryPoint” validates and executes everything.
This architecture enables several things that were previously impossible:
Gas sponsorship. A “paymaster” contract can cover gas fees on behalf of users. This means a dapp can pay the gas for its users, or users can pay gas in stablecoins, or gas can be abstracted away entirely from the user experience. Coinbase has used this on Base to let new users transact without holding ETH first.
Alternative signature schemes. Instead of being locked to ECDSA (the algorithm behind standard Ethereum keys), a smart account can validate transactions using passkeys, biometrics, or multisig setups. Argent and Safe both offer social recovery built on this.
Batch transactions. A single UserOperation can execute multiple contract calls atomically. Approving and swapping in a DEX becomes one action, not two separate transactions with two gas costs.
By early 2026, the ERC-4337 ecosystem had deployed tens of millions of smart accounts across mainnet and layer-2 networks, according to data tracked by ERC-4337 infrastructure providers including Alchemy and Pimlico.
EIP-7702: The Protocol-Level Upgrade
ERC-4337 works, but it is layered on top of Ethereum’s architecture in a way that adds complexity and cost. EIP-7702, included in Ethereum’s Pectra upgrade (which shipped on mainnet on May 7, 2025), takes a different approach.
EIP-7702 lets an EOA temporarily delegate its authority to a smart contract. When a user signs a special authorization message, their regular wallet gains the capabilities of whichever contract they’ve delegated to, just for that transaction. This means existing wallets can get smart account features without migrating to a new account type.
The practical effect is that users do not have to abandon their existing addresses or go through a migration process. A MetaMask wallet or a hardware wallet address can opt into batch transactions and gas sponsorship without moving funds anywhere.
EIP-7702 and ERC-4337 are not competing standards. They address slightly different layers of the same problem and are designed to work alongside each other.
What This Enables in Practice
Account abstraction is infrastructure, so its value shows up in products rather than in the mechanism itself.
Safe (formerly Gnosis Safe) has long been the standard multisig for DAOs and institutional crypto. Smart account infrastructure lets similar multi-approver logic extend to individual consumer wallets without the complexity overhead.
Coinbase’s Smart Wallet, built on Base, lets users create a wallet with a passkey (the same technology used by password managers and face ID logins) rather than a seed phrase. No seed phrase to write down. No risk of losing a 12-word phrase.
Session keys are another use case that has emerged. A session key lets a user pre-authorize a game or dapp to send limited transactions on their behalf, without giving full wallet access. In practice this means a blockchain game can execute moves without a signature popup for every action.
The Gap That Remains
Account abstraction solves the programmability problem. It does not solve the key management problem entirely. Users still need to protect whatever root credential governs their smart account, whether that is a private key, a passkey device, or a social recovery configuration. If all three social recovery contacts are compromised, recovery is not possible.
The more realistic remaining gap is adoption. Most crypto users still interact with EOA wallets. Most dapps are not yet built to take advantage of paymasters or session keys. The infrastructure exists; the product layer is catching up.
EIP-7702 has been live since Pectra shipped in May 2025. The ability to upgrade existing wallets without migration removes the largest friction point for the wallets that already have users, and adoption has been building since activation.