Nethermind 2.0 Brings Flat State to Ethereum
Ethereum infrastructure provider Nethermind has released version 2.0.0 of its execution client, introducing a flat state database as the default storage backend and significantly changing the storage requirements for operators running historical Ethereum nodes.
According to the official Nethermind 2.0.0 release notes, a full Ethereum mainnet archive now requires approximately 2.1 TB for the execution database, with total node disk requirements of about 2.3 TB in the measured configuration. Nethermind recommends provisioning roughly 3 TB for a full mainnet archive to allow room for the consensus layer and future growth.
The release also introduces configurable historical-data retention. Operators can maintain the complete archive, keep a rolling block window, retain history from a selected block height, or preserve deep history for specific addresses. This gives node operators more control over storage instead of requiring every node to retain the same amount of historical state.
Flat State Reduces Archive Storage Requirements
The main change in Nethermind 2.0 is the move toward a flat state database for new nodes and resyncs. Rather than relying on the previous archive-storage architecture, the new system stores archive-grade history directly through the flat database.
Nethermind says the historical data remains header-verifiable, meaning the state root at a retained historical height can be reconstructed from the stored data and checked against the corresponding block header. The system also fails closed when a query requests history below the configured retention floor instead of silently returning information from the current state.
The release supports four main approaches to historical storage: full archive retention, a rolling window, retention from a fixed block height and per-address history slices. This flexibility can be particularly useful for operators that need deep historical data for only certain contracts rather than the entire Ethereum state.
Performance Gains Extend Beyond Storage
Nethermind 2.0 is also designed to improve execution performance. In Nethermind's 12-hour mainnet benchmark, the flat configuration processed blocks at the chain tip at 1.80 times the throughput on 8-vCPU standard machines and 2.55 times the throughput on 16-vCPU premium machines compared with the 1.39.2 half-path baseline.
The client also showed improvements in RPC performance. Across Nethermind's 497-record eth_call benchmark, version 2.0.0 answered requests 15% to 27% faster than version 1.39.3, while producing byte-identical responses across the measured configurations. Flat storage delivered up to 29.6% more throughput on amd64 and 36.4% more on arm64 in the benchmark.
There is a trade-off during initial synchronization. Nethermind reported that flat storage took longer to sync to the chain tip than the previous baseline in its measurements. Existing Patricia-based nodes also remain on their current database after upgrading unless operators take additional steps to migrate or resync.
Nethermind 2.0 Includes Breaking Changes
The upgrade is not purely a storage optimization. Nethermind warns operators about several breaking configuration changes, particularly those running archive nodes or using custom settings.
Among the changes, Db.FlushOnExit is now an enum rather than a Boolean, Blocks.PreWarmStateOnBlockProcessing has been replaced by Blocks.PreWarming, and several numeric configuration properties now require unsigned values. The default Receipt.MaxBlockDepth was also reduced from 10,000 to 1,000, meaning operators relying on wider eth_getLogs queries may need to adjust their configuration.
Because of those changes, Nethermind specifically recommends reviewing the breaking-change section before upgrading. The release notes state that most nodes can upgrade in place without a resync, but archive-node operators and users with custom configurations need to pay particular attention to the new storage and retention settings.
Safe Contracts 1.5.0 Adds Module Guards
Ethereum account infrastructure is also receiving an update. Safe has released Safe contracts version 1.5.0, which becomes the default contract version for newly created Safe smart accounts.
The update introduces Module Guards, designed to make transactions initiated by third-party modules subject to the same security rules enforced by the Safe account's guard system. Safe's official version 1.5.0 announcement says the update also tightens ERC-1271 contract-signature validation and requires smart-contract owners to return the appropriate signature magic values defined by the standard.
Safe also says failed calls now return more specific revert reasons rather than generic execution errors. Existing Safe accounts using versions 1.3.0 and 1.4.1 can upgrade through a single transaction from the account settings, while their addresses, owners, signing thresholds, modules and existing transaction guards remain unchanged.
Google Cloud Joins Puffer's Preconfirmation Infrastructure
Ethereum scaling infrastructure is also expanding through a new partnership between Puffer Finance and Google Cloud.
Under the arrangement, Google Cloud joins the Puffer Preconf gateway registry as a registered operator. According to Puffer's announcement, the infrastructure is designed to route transactions, enforce ordering policies and provide sub-50-millisecond execution preconfirmations before transactions ultimately settle on Ethereum Layer 1.
The system is being used with Puffer UniFi, a rollup architecture designed around multiple gateway providers. Puffer says the multi-gateway structure can allow rollups to use parallel providers or fallbacks rather than relying on a single infrastructure operator. The arrangement is aimed at applications where low latency is important, including decentralized finance, autonomous-agent commerce and institutional trading.
NYMIS Launches an Ethereum Wallet Privacy Scanner
The infrastructure developments are accompanied by a different type of tool focused on wallet privacy.
NYMIS has introduced a scanner designed to show how much information can potentially be inferred from a public Ethereum wallet address. The tool uses transaction and wallet-clustering intelligence attributed to Nansen to examine an address's visible onchain activity.
The scanner can identify potential links involving KYC services, historical funding paths, token-transfer activity and Safe smart accounts. It also examines frequently used protocols, interaction patterns and connected onchain social or identity profiles. The goal is to show how an apparently pseudonymous wallet can develop a recognizable operational footprint when its transactions are analyzed together.
For users, the tool highlights an important characteristic of public blockchains: transactions may not reveal a person's name directly, but repeated funding patterns, contract interactions and links between wallets can potentially connect separate addresses to the same broader activity.
Ethereum Infrastructure Keeps Evolving
The four developments highlight different layers of Ethereum's expanding infrastructure.
Nethermind 2.0 focuses on making node storage and execution more efficient, with a full archive measured at roughly 2.1 TB of execution data. Safe 1.5.0 strengthens smart-account controls through Module Guards and tighter signature validation. Puffer and Google Cloud are working on low-latency preconfirmation infrastructure, while NYMIS focuses on the privacy implications of publicly visible wallet activity.
Together, these releases show that Ethereum development is extending beyond transaction throughput alone. Storage efficiency, smart-account security, low-latency transaction infrastructure and wallet privacy are all becoming increasingly important parts of the ecosystem as developers build more complex applications on top of the network.
