The Fusaka upgrade on the Ethereum network, activated in December 2025, has successfully expanded data availability capacity but failed to drive increased usage, while revealing network reliability issues at higher loads. An analysis by MigaLabs of over 750,000 slots since the upgrade's activation shows that median blob usage per block actually declined after the first parameter adjustment, despite the network's capacity increasing.
The upgrade's goal was to reduce layer-2 rollup costs by increasing throughput for blob data—the compressed transaction bundles rollups post to Ethereum for security. Through a mechanism called Blob Parameter Overrides, the target blob count per block was raised incrementally from a baseline of 6 (with a maximum of 9) to a target of 10, and then to 14, with a maximum ceiling of 21 blobs. These changes did not require hard forks.
However, the data reveals a significant gap between capacity and utilization. The median blob count per block fell from 6 before the first override to 4 afterward. Blocks containing 16 or more blobs remain extremely rare. More critically, the analysis shows network reliability degrades at the edges of the new capacity. Miss rates for blocks—indicating failures in propagation or attestation—climb from a baseline of around 0.5% to between 0.77% and 1.79% for blocks with 16 or more blobs. At the maximum capacity of 21 blobs, the miss rate hits 1.79%, more than triple the baseline.
"The report's conclusion is direct: no further increases in the blob parameter until high-blob miss rates normalize and demand materializes for the headroom already created," the analysis states. This suggests the network's infrastructure, including validator hardware and bandwidth, struggles to handle blocks at the upper end of the new capacity.
On a positive note, the upgrade's other key component, EIP-7918, which introduces a reserve price floor for blob fees, appears to be functioning. This prevents blob auction fees from collapsing to near zero, ensuring they remain a meaningful economic signal of demand.
The findings indicate that the pre-upgrade concern—that limited blob capacity was a bottleneck for Layer 2 scaling—may have been misplaced. The bottleneck appears to have shifted, with rollup demand not yet filling the available space. This raises questions about whether factors like sequencer economics, user activity, or cross-rollup fragmentation are now more limiting than raw data availability. The results suggest Ethereum has headroom for growth within its current parameters, but further capacity expansions should wait until utilization increases and network stability at high blob counts improves.
