The mechanics of Ethereum’s Fusaka scaling

The Fusaka upgrade addresses data availability bottlenecks through PeerDAS. Dankrad Feist and the development team use statistical sampling to allow nodes to verify data without downloading the full blockchain. This mechanism uses the data column sidecar subnet count and the custody requirement to determine necessary peer connections. If the subnet count is 128 and the custody requirement is 16, a node connects to 25.9 peers on average. When the subnet count is 32 and the requirement is 1, the node connects to exactly 32 peers. The node discovery mechanism finds new peers until the node covers all subnets. The node does not keep a peer that adds no more coverage to the subnets. A higher subnet count increases the expected peer count even if the ratio between the subnet count and the custody requirement remains constant. The implementation of PeerDAS allows nodes to verify data availability through statistical sampling methods, which means individual participants process only small data chunks instead of the full datasets that current blockchain architectures require from every single node in the network. Coinbase’s Base and Worldcoin transaction volumes pushed Ethereum to its first six blobs per block target. The implementation will double the current blob capacity. The first fork after the December 3 deployment will raise the maximum blob count from nine to 15. A second fork in January targets 21 blobs per block.

Vitalik Buterin proposes reducing the validator lockup threshold from 32 ETH to 1 ETH. The current 32 ETH threshold balances security, decentralization, and operational costs. Buterin argues that this high threshold is a barrier for smaller participants and solo stakers who want to contribute to network security. A lower threshold encourages more solo stakers to join the network. An increased number of validators could slow down finality or increase node operational costs. Buterin suggests single-slot finality to mitigate these risks. This concept reduces the time to confirm a block from 15 minutes to 12 seconds. Do the increased costs of running more nodes outweigh the benefits of a more decentralized validator set?

Nethermind’s Lukasz Rozmej works on the Grandine client to manage PeerDAS data. Grandine splits blobs into 128 data columns using erasure coding. A node custodies a subset of these columns. The protocol requires nodes with validators to custody more columns. Grandine computes the requirement from the total effective balance of the registered validators. A node with 256 ETH of effective balance custodies 8 groups. A node with 4,096 ETH or more custodies all 128 groups. A semi-supernode needs about half the bandwidth of a supernode. The 68% supernode bandwidth reduction necessitates precise parameter tuning in the PeerDAS simulation.

Grandine automatically reconstructs missing columns if a node receives at least half of the columns. The default reconstruction delay is 250 milliseconds. The node requests columns from peers and then publishes the rebuilt columns to help the network. The node re-checks the requirement once per epoch and only raises the custody group count. When the count increases, the node backfills columns for the data availability window. Grandine keeps sidecars for the data availability window of 4,096 epochs. For Fulu blocks, Grandine builds blobs from stored data columns. This requires at least half of the columns. The execution client provides blobs to Grandine through engine_getBlobsV1 or engine_getBlobsV2.

Regulatory updates in 2026 shape the staking environment. The MiCA transitional period ended on July 1. Only 17% of pre-MiCA licensed entities converted to full CASP authorization. In Australia, the ASIC no-action relief expired on June 30. Any firm offering crypto-related financial services to Australian clients without a license breaches the Corporations Act. The UK FCA will commence its full regime on October 25, 2027. The SEC published its Draft Strategic Plan for 2026 to 2030 on June 2. This plan designates digital assets as the agency’s first regulatory objective. The SEC aims for a rational, coherent approach to oversight. The SEC’s plan targets oversight without duplicative or conflicting requirements. The UK FCA states that validator operators lose their technology-only exemption when they provide user dashboards or reward-compounding tools.