Vitalik Buterin, co-founder of Ethereum, boldly risk-controlled the network’s next scaling step, saying “Fusaka will fix this,” “safety is paramount.” In a detailed post, Buterin described Fusaka’s core feature, Peerdas, as “it’s trying to do something quite unprecedented: it has a live blockchain that doesn’t require a single node to download full data.”
He added that Peerdas relies on stochastic sampling of “chunks” of data, and if more than half of those chunks are available, the nodes can retrieve them and reconstruct the rest via erasure coding. “This is all new technology, and it’s wise for core developers to be very careful about testing,” wrote Buterin, noting that Blob Supply will increase “conservatively at first” and then more aggressively if conditions allow.
Buterin calls Fusaka the key to Ethereum L2 scaling
Buterin’s remarks arrive just as Ethereum’s new blob market shows signs of tension. Data Director at Dragonfly Hildebert Moulié reported that the chain “has hit six blobs/block for the first time”, with recent surges being largely due to rollups and projects that include bases and World Coins. According to the same thread, the base filed about 35% of the Blob and used about 42% of the Blobpace, and WorldCoin contributed about 20% of the submission and 25% of the usage. Arbitrum, Op Mainnet, Soneium, Scroll and more made up most of the rest.
Analysts added that L2S now accounts for around $200,000 per week in mainnet fees submitted. Validators require more than 70 GB of storage on their blobs, and many blobs are used more frequently, especially by posting more frequently than they can fulfill a payload of 128 kB. Hildobby warned that “Blob Price Discovery” still requires long-term saturation of demand, but the first sustained basefly spike has been observed since the Pectra hard fork was also observed.
Peerdas is an architectural response. Buterin explained that each node only requires a small number of chunks to ensure that more than half of the data is available. In that case, the node “theoretically you can download these chunks and recover the rest using erase coding.”
In the first incarnation, the two non-obligatory roles must exist in a location on the network. This requires early broadcasts and emergency reconstruction when publishers reveal only a small portion of the block. Endgame is to unlock persistent L2 scaling and ultimately route more L1 execution data to the BLOB as L1 block gas limits rise.
This pivot lands in the rapidly evolving blob marketplace. After Pectra, Ethereum increased the blob target and maximum per block, increasing daily data capacity, paving the way for higher throughput from rollups. The research desk links the shift to the complex interaction between L1 base fees, BLOB fees, and L2 submission behaviors.
Fusaka timetables add urgency. The core developers have shown the mainnet activation on December 3, 2025, and will clearly save the emphasis on butaline’s “safety” following a gradual testnet rollout. Peerdas debuted under strict restrictions, with BLOB counts increasing “conservatively at first.” This is an attitude designed to avoid charge whiplashes and observe how the L2 actually consumes additional capacity.
Outside of protocol notes, empirical work accumulates on how networks can use Blobspace more efficiently. A 2024 survey on “blob sharing” claimed that small rollups don’t fill blobs frequently, so they could reduce posting costs by more than 85% by cooperatively packing shared blobs, smoothing out the base fee and reducing the total number of blobs submitted.
Ethereum researchers then expand on that debate, modeling how sharing reduces more blocks than blobs with target numbers, thereby attenuating the exponential blob-fee adjustments that use kicks when overshooting the target. These findings hamper Mourier’s observation that “many chunks are not full.” This means that as the market matures, large savings can be taken advantage of through better adjustments.
Peerdas’ conceptual roots go back through Ethereum research notes on sampling data availability and his own writings on Buterin’s own “The Surge.” Peerdas itself implements one-dimensional sampling using erase coding and concise per-cell proofs, allowing availability to be verified without naive downloading everything. That’s why live, high-value blockchains make the approach “pretty unprecedented.” It aims to decentralize and tune throughput by reducing bandwidth and storage requirements per node, while maintaining a strong assurance that data actually exists.
Still, shifts are not without risk. Butaline’s claims to be careful development reflect the reality that Ethereum’s blob economy is young, unstable and sensitive to sudden changes in demand. As Jostle due to L2S capacity, the fee dynamics can be reversed quickly, with incomplete chunks, spiny usage, and side effects of MEV complicating predictions. Fusaka’s promise is that Perdas can bending these dynamics towards sustainable growth by allowing network-scale data availability without forcing a single node across the chain.
At the time of pressing, ETH traded for $4,028.
Featured images created with dall.e, charts on tradingview.com
Editing process Bitconists focus on delivering thorough research, accurate and unbiased content. We support strict sourcing standards, and each page receives a hard-working review by a team of top technology experts and veteran editors. This process ensures the integrity, relevance and value of your readers’ content.
