Fusaka improves blob scaling

Ethereum’s scaling problem was never just “make blocks bigger.” The real bottleneck was blob data — the temporary payloads rollups post to Ethereum so anyone can verify what happened. Before Fusaka, every full node had to download every blob. That worked at low volume, but it made higher throughput expensive for nodes and risky for decentralization. Fusaka changed that on December 3, 2025 by activating PeerDAS, a new way to check blob availability without forcing every node to carry the whole load. ### What are blobs, exactly? Blobs are a special data type Ethereum added in the Dencun upgrade for layer-2 rollups. They are cheaper than regular calldata because the EVM does not execute them, and nodes only keep them for a limited time. Rollups use blobs to post the data needed for security while doing most transaction processing off the main chain. ### Why did blobs become the choke point? Because pre-Fusaka Ethereum treated blob availability in the simplest possible way — every node downloaded every blob. (ethereum.org) If rollups wanted to post more data, each node’s bandwidth and storage burden climbed with them. That is the part Ethereum has been careful about, because if node requirements rise too far, running a node starts looking like a data-center job instead of something normal operators can do. (ethereum.org) ### So what does PeerDAS do? PeerDAS stands for Peer Data Availability Sampling. Instead of every node storing the full blob set, nodes sample and hold only part of it. Ethereum’s own roadmap page describes full nodes as holding roughly 1/8 of the blob data, randomly distributed across the network. The trick that makes this safe is erasure coding — basically, the data is encoded so the full blob can be rebuilt from enough pieces rather than every piece. (ethereum.org) ### Why is that enough for security? Because the question is not “does my node personally have every byte?” The question is “can the network prove the full data exists and can be recovered?” PeerDAS answers that by combining sampling with reconstruction guarantees. Ethereum’s docs say any portion of the data can be reconstructed from any existing 50% of the whole, with the chance of bad or missing data pushed down to a cryptographically negligible range. That is the whole unlock — less duplication per node without giving up the availability check. (ethereum.org) ### Why does this help rollups? Rollups live and die by data-availability costs. If Ethereum can support more blobs per block, rollups can batch more transactions into each posting window. That spreads fixed costs across more users and pushes fees down. Fusaka’s mainnet announcement was explicit about the point here: PeerDAS is the step that lets Ethereum raise blob throughput beyond the old limits while keeping security and decentralization intact. (ethereum.org) ### Did Fusaka raise blob capacity immediately? Yes, but the bigger story is that it created a safer path for repeated increases. Ethereum paired PeerDAS with Blob Parameter Only forks — small follow-on upgrades that adjust blob target and max values without bundling in a giant named fork. The first two planned moves after mainnet were from 6 target and 9 max blobs per block to 10 and 15, then to 14 and 21. That tells you the strategy: unlock the architecture first, then turn the dial upward in stages. (blog.ethereum.org) ### Is this just an L2 fee story? Mostly, but not only. Lower DA costs change rollup design itself. If posting data to Ethereum gets cheaper and more abundant, teams can be less aggressive about offloading data elsewhere or compressing every last byte at the expense of complexity. In plain English — PeerDAS does not just make today’s rollups cheaper, it changes the tradeoffs for how future rollups split work between storage, proving, and settlement. That last part is an inference from the throughput and cost mechanics Ethereum lays out. (blog.ethereum.org) ### Bottom line? Fusaka matters because it solved the part of blob scaling that was about node burden, not just demand. PeerDAS did not magically finish Ethereum scaling. But it removed a hard ceiling. Now blob capacity can grow without requiring every node to swallow every blob — and that is the difference between a one-off fee dip and a real scaling roadmap. (ethereum.org)