The Ethereum Foundation has published a detailed update on its blob scaling strategy, outlining incremental improvements to increase data availability throughput for layer-2 networks. Building on the previous Protocol Update 001, the new plan introduces a series of changes spanning the upcoming Fusaka and Glamsterdam network upgrades, as well as intermediate adjustments.

The first major milestone is the delivery of PeerDAS in the Fusaka upgrade, expected later this year. PeerDAS, which stands for peer data availability sampling, allows individual nodes to download only a subset of blob data per block via randomized sampling. This approach keeps computational load bounded even as the total number of blobs increases, enabling a theoretical increase of up to 8x from current throughput levels, moving from about 64 KB/s to 512 KB/s.

Rather than jumping immediately to the maximum capacity, developers have opted for gradual increments through Blob Parameter Only (BPO) hard forks. These pre-programmed steps automatically increase blob limits over time without requiring manual activation after deployment. Monitoring on mainnet will guide the schedule for subsequent increases. Barnabas Busa and the EF PandaOps team are coordinating with client teams to define the correct cadence.

Alongside BPOs, bandwidth optimizations are being developed to reduce redundant communication on the network. Raul Kripalani and Marco Munizaga are leading efforts on cell-level messaging, which allows nodes to query specific parts of blob samples more efficiently. These changes do not require consensus or execution protocol modifications and can be shipped before the Glamsterdam upgrade in early 2026.

Looking further ahead, PeerDAS v2 planned for Glamsterdam will build on the pipelining benefits of EIP-7732 and further refine cell-level messaging and data reconstruction techniques. This next-stage design aims to exceed the limits of the initial PeerDAS implementation. The exact throughput gains will depend on actual performance and mainnet analysis as blob counts are scaled in production.

The foundation also emphasized preserving Ethereum’s core values, particularly censorship resistance, as blob capacity scales. Csaba Kiraly is leading work on scaling the blob mempool to maintain a decentralized inclusion mechanism even as throughput grows. Current implementations support near-term targets, with further research needed for higher levels unlocked by Fusaka.

For the longer term, research under the banner of FullDAS is exploring next-generation data availability sampling. Studies such as Robust Distributed Arrays aim to formalize a highly diverse peer-to-peer network that can shard an increasing number of samples while remaining fault-tolerant. Additional considerations include low-latency inclusion, censorship resistance, and enhancements to the blob fee market. This research is stewarded by Francesco D’Amato.