According to a recent social media post by Etherscan, the Ethereum network processed a record-breaking 1.91 million transactions on Layer 1 (L1) in a single day. At the same time, the fees are incredibly low at $0.16.
This shows that the network can now handle massive traffic without pricing out normal users.
This combination of high throughput and low costs is the direct result of two major network upgrades executed in 2025: Pectra and Fusaka.
Fusaka, which went live earlier in the month, is the most immediate cause of the record was the upgrade. This upgrade directly expanded the capacity of the Ethereum L1 blockchain.
The upgrade has managed to increase the size of each block by roughly 33%. This allowed the L1 network to fit significantly more transactions into every block.
Previously, all nodes had to download all data, which created a bottleneck. PeerDAS, a new feature introduced in Fusaka, has made it possible for nodes to verify data “blobs”, large chunks of transaction data, by sampling just tiny parts of them. Blobs, which were introduced in an earlier update called Dencun but expanded here, are like sidecars attached to the main block. They carry data cheaply and don’t compete with standard transactions.
The Pectra upgrade, which took place in May, laid the groundwork for scaling by optimizing how Layer 2 networks of the likes of Arbitrum, Optimism, and Base interact with the main chain.
Pectra doubled the number of these “sidecars” from 3 to 6 per block. Because there was suddenly double the supply of space for Layer 2 data, the cost for L2s to “settle” on Ethereum dropped. This kept the overall network uncongested.
Ethereum scaling is still not “finished” despite the massive success of the 2025 upgrades.
The Ethereum ecosystem is still fractured. Users tend to find it challenging to use L2 funds without complicated bridges. Hence, the fragmentation remains a major issue.
The database of all accounts, balances, and smart contracts (the “State”) grows larger and larger. Eventually, the State becomes terabytes or petabytes in size. If it gets too big, a normal person can’t buy a hard drive big enough to run a node.
