What EIP-4844 and Blobs Actually Changed

EIP-4844, which went live in March 2024 as part of the Dencun upgrade, is the most significant practical change to Ethereum's scaling story in years. The effects were immediate and measurable. L2 transaction fees dropped by 80-95% almost overnight.

Understanding why requires a quick technical background. Before Dencun, rollups posted their transaction data to Ethereum L1 as "calldata" — regular transaction data that the entire Ethereum network processes and stores permanently. This is expensive because you're paying for permanent storage and execution capacity on a globally distributed network. Even though rollups batch hundreds of transactions together, the per-transaction cost was still significant enough to create a noticeable user experience problem: Ethereum L2 transactions were cheaper than L1 but not dramatically so for small transfers.

EIP-4844 introduced a new transaction type that carries "blobs" — large chunks of data attached to transactions but not processed by the EVM and not stored permanently on chain. Blobs are only available for a short window (about 18 days), after which the data can be pruned. The key insight is that rollups don't need L1 to store their transaction data forever — they just need it available long enough that anyone can download it and verify it if they want to challenge the rollup. After that, the rollup's own state and proofs are enough.

This dramatically reduced the cost of what rollups were actually paying for. Blob space is priced with its own separate fee market, which starts cheap and only becomes expensive when there's high demand for it. In practice, for the first several months after Dencun, blob fees were essentially zero because demand wasn't high enough to drive up the price.

The immediate effects: Base (Coinbase's L2) went from averaging over a dollar per transaction to well under a cent. Arbitrum and Optimism saw similar reductions. This opened up use cases — microtransactions, gaming, payments — that weren't economically viable before.

The limitations are also worth being clear about. Blobs gave rollups cheap data availability today, but the long-term goal requires data availability that scales with demand. Proto-Danksharding (what 4844 is officially called) is a stepping stone to full Danksharding, which would scale blob capacity dramatically by having different validators responsible for different portions of the data. Full Danksharding is years away and involves significant engineering work.

There's also a question about whether cheap L2 transactions actually benefit L1 ETH holders. In the bull case, lower L2 fees drive more users, which drives more L2 activity, which generates more blob fees, which means more ETH is burned (through the EIP-1559 mechanism). In practice, the relationship is more complicated and the net effect on ETH supply and staking yields depends on a lot of factors that are still playing out.

The honest takeaway: EIP-4844 was a clean technical win that delivered real user-facing improvements ahead of schedule. It doesn't solve every scaling challenge, but it made a significant concrete improvement to the experience of using Ethereum-based applications right now.

What EIP-4844 and Blobs Actually Changed

// COMMENTS

ON THIS PAGE