Researchers at the Ethereum Foundation (EF) have warned that the ever-increasing data load on the network (“state bloat”) is becoming a serious problem for Ethereum nodes.
The Foundation’s Stateless Consensus team noted that the storage space required for accounts, smart contract data, and application code maintained on the network is increasing day by day, making it harder for nodes to run.
Ethereum’s “state” includes all information about the current state of the network, including account balances, contract storage, and the code that powers applications. According to EF, Ethereum is currently a global infrastructure that processes billions of dollars in value and runs thousands of applications. But this scale comes with significant challenges. Nations are always growing and never shrinking.
The researchers note that as states grow, running full nodes becomes more expensive and more vulnerable. A blog post shared by EF states that Ethereum’s decentralized nature could be compromised if the state is too large or can only be controlled by a limited number of powerful operators.
Widespread adoption of layer 2 solutions, along with scaling measures such as EIP-4844 (proto-dunk sharding) and increased gas limits, has increased network throughput. But these advances are also accelerating the state’s growth, according to EF researchers.
The team noted that allowing full state storage and services only to a small number of large, technologically advanced operators poses risks to Ethereum’s censorship resistance, neutrality, and network security. Therefore, researchers are actively testing when state size becomes a barrier to scaling and how node software behaves under excessive data loads.
Ethereum’s long-term plans include stateless validation, which aims to allow validators to validate blocks without storing the full state. This approach reduces the burden on validators while shifting data storage responsibility to a narrower group.
According to EF, in a stateless architecture, most of the state is maintained by specialized operators such as block producers, RPC service providers, MEV seekers, and block explorers. Researchers say this could create new challenges in terms of synchronization issues, censorship risks and resilience to external pressures.
The Stateless Consensus team proposed three different approaches to reduce state overhead and make nodes more sustainable.
State expiration: The goal is to remove long-term unused data from active state and restore it with evidence if necessary. According to EF, about 80% of the current equipment has not been used for more than a year.
State archiving: The plan is to separate frequently used “active” data from rarely accessed “archived” data. This is intended to prevent node performance degradation due to chain aging and allow the system to remain more stable over time.
Partial stateless architecture: Nodes maintain only part of the state, and wallets and light clients are expected to cache the necessary data themselves. This approach aims to reduce storage costs, allow more users to run nodes, and reduce dependence on large RPC providers.
*This is not investment advice.
