Source: Avail; Compiled by: AIMan@黄金财经
This is the biggest upgrade for Ethereum since Dencun went live 420 days ago. These changes will improve how Ethereum handles publishing data from L2, as well as other changes included in the upgrade.
The Pectra upgrade introduces two key changes to L2, which uses Ethereum data availability. It increases blob throughput and increases the cost of call data. We analyzed the usage of Ethereum blobs since Dencun (EIP-4844) to understand the impact that the Pectra upgrade may have on the blob market.
Blob Utilization
Since November 2024, the number of blobs has hovered around the target of 3 blobs per block (100% utilization). While the cap is 6, the price of submitting blobs increases when more than 3 blobs are submitted, just like Ethereum fees increase during peak load. The purpose of this is to provide additional capacity when needed while maintaining the target of 3 blobs per block.

Source: https://dune.com/hildobby/blobs
With the Pectra mainnet live, we can expect utilization to increase, moving towards the new target of 6 blobs per block, with the cap rising to 9.
After the Dencun upgrade, it took about 8 months for Ethereum blobs to reach full capacity. Vitalik Buterin warned in September 2024 (just six months after the Dencun upgrade) that blob space was about 75% full and “close to the cap.” He also said that “no doubt there were multiple layer 2 nodes that considered migrating to blobs but ultimately decided not to,” mainly because there was not enough space on Ethereum at current capacity to accommodate them.

Source: https ://github.com/ethereum/pm/issues/1153
Under peak load, Blob fees occasionally soared to higher than call data fees. A month after the Dencun upgrade, with the rise of the Blob subscription boom, the Blob base fee also soared. Then, with the Arbitrum LayerZero airdrop and Arbitrum’s increased volume, Blob base fees spiked again in June. This time, L2 nodes submitted blobs instead of call data, and therefore paid about $550,000 more.
While L2s can implement strategies such as switching to call data when fees rise, slowing down batch submissions, paying more to have their blobs included, or simply stop publishing data until fees stabilize, the fact is that L2s still need more capacity when demand surges.
So, which protocols are consuming all the blobs?
The chart below shows the number of blobs published per week, broken down by blob submitter (L2). Note that each line in the chart below represents the number of blobs published in a week.

Source: https://dune.com/hildobby/blobs
Doing a quick napkin math we found that there were about 150k blobs available per week at a target rate of 3 blobs per block, which has now increased to about 300k blobs per week at a target rate of 6 blobs per block.
~7,150 Ethereum blocks per day x 6 blobs per block x 7 days = 300,300 blobs available per week
The Base blockchain has been consuming about a third of available capacity, consuming about 50,000-60,000 blobs per week. Taiko is another large blob consumer, regularly consuming 20,000-25,000 blobs per week. In addition, World Chain and Arbitrum are both consuming between 10,000 and 20,000 blobs per week.
Assuming these four L2s don’t start publishing more data, their current blob consumption will take up about 30-40% of Ethereum’s blob capacity after Pectra.
What happens next?
It is undeniable that the introduction of Rollup can significantly improve the Ethereum ecosystem, leading to higher throughput, more teams, more applications, and more users. However, as we all know, Pectra is just a stopgap measure. Currently, these incremental improvements are undergoing necessary adjustments to increase throughput and support more Rollups, but the market seems ready to digest all the Blobs that Ethereum can produce.
Full Danksharding will once again provide more capacity improvements, including the introduction of Data Availability Sampling (DAS), which enables end users to verify data availability from edge devices.
However, if we look at the bigger picture a little bit, such as the blobscriptions craze, or a popular airdrop on a certain L2 layer causing unrelated network outages and congestion, then we are still a long way from reaching the throughput required to support a new global financial system.