Ethereum's Fusaka: Buterin Reveals Radical Scaling Shift

Ethereum co-founder Vitalik Buterin has introduced the network’s next significant scaling endeavor, terming it both ambitious and meticulously risk-controlled. He stated, “Fusaka will fix this,” while consistently emphasizing that “safety first is of the utmost importance.” In a comprehensive post, Buterin elaborated on Fusaka’s core innovation, PeerDAS, describing it as an attempt to achieve something truly unprecedented: establishing a live blockchain that does not necessitate any single node to download the entirety of the data.

He further explained that PeerDAS operates by probabilistically sampling data in “chunks.” If more than half of these data chunks are verified as available, nodes possess the capability to retrieve them and reconstruct the remaining data through erasure coding. Buterin acknowledged, “This is all new technology, and the core devs are wise to be super cautious on testing,” indicating that the supply of blobs would initially increase “conservatively,” with more aggressive expansion only if conditions prove favorable and stable.

Buterin Identifies Fusaka as Key to Ethereum L2 Scaling

Buterin’s statements come at a crucial time, as Ethereum’s nascent blob market begins to show signs of stress. Hildebert Moulié, Head of Data at Dragonfly, reported that the chain “hit 6 blobs/block for the first time,” attributing this recent surge primarily to various rollups and projects such as Base and Worldcoin. According to Moulié’s analysis, Base was responsible for approximately 35% of blob submissions and consumed about 42% of the available blobspace. Worldcoin’s contributions were significant as well, accounting for around 20% of submissions and 25% of usage. Other platforms including Arbitrum, OP Mainnet, Soneium, and Scroll made up the majority of the remaining usage.

The analyst also highlighted that Layer 2 (L2) solutions are now generating roughly $200,000 per week in mainnet fees for their data submissions. Furthermore, validators are now faced with a requirement for more than 70 GB of storage specifically for blobs, a figure that escalates to over 1.2 TB if the data remains unpruned. It was also noted that many blobs are not fully utilized, particularly by smaller rollups that post data more frequently than they can fill their 128 KB payloads. The first sustained base-fee spike since the Pectra hard fork was observed, although Moulié cautioned that true “blob price discovery” would require a more prolonged period of demand saturation.

PeerDAS represents the architectural solution to these emerging challenges. Buterin clarified that each node will only request a small, limited number of data chunks to verify that over half of the block’s data is indeed available. If this condition is met, the node “theoretically can download those chunks, and use erasure coding to recover the rest.” This mechanism drastically reduces the data burden on individual nodes.

In its initial implementation, two non-custodial functions will still demand the full-block data to be present somewhere within the network: the initial broadcast of data and emergency reconstruction in scenarios where a publisher only reveals a partial block. However, Buterin assured that “we only need one honest actor” for these specific tasks. He also projected that future advancements, such as “cell-level messaging and distributed block building,” will enable even these two functions to be fully distributed. The ultimate goal, he suggested, is to unlock sustained scaling for L2s and, as Layer 1 (L1) block gas limits eventually increase, to route a greater portion of L1 execution data into blobs as well.

This strategic pivot arrives amidst a rapidly changing blob marketplace. Following the Pectra upgrade, Ethereum significantly increased both the blob target and the maximum number of blobs per block. This expansion boosted the network’s daily data capacity, thereby paving the way for higher throughput from rollups. Research institutions have attributed this shift to a complex interplay between L1 base fees, blob fees, and the behavior of L2 submissions.

The timeline for Fusaka adds a sense of urgency to these developments. Core developers have indicated a target mainnet activation date of December 3, 2025, which will be preceded by carefully staged testnet rollouts. This timetable underscores Buterin’s “safety first” principle. PeerDAS will initially be deployed with strict limitations, and blob counts will be increased “conservatively at first.” This cautious stance is designed to prevent sudden fee volatility and to thoroughly observe how diverse L2 solutions genuinely utilize the enhanced capacity.

Beyond the protocol specifics, a growing body of empirical research is exploring how networks might optimize blobspace utilization more efficiently. A 2024 study on “blob sharing” put forth the argument that smaller rollups frequently under-fill their blobs. By cooperatively packing data into shared blobs, these rollups could potentially reduce their posting costs by over 85%, leading to a smoother base fee and a lower overall number of submitted blobs.

Ethereum researchers have further expanded on this concept, developing models that illustrate how data sharing can decrease the occurrence of blocks exceeding the target number of blobs. This, in turn, helps to dampen the exponential blob-fee adjustments that are triggered when usage surpasses predefined targets. These findings align perfectly with Moulié’s observation that “many blobs aren’t full,” suggesting that substantial cost savings are achievable through improved coordination as the market for blobs matures.

The theoretical foundations of PeerDAS can be traced back through extensive Ethereum research notes on data-availability sampling and Vitalik Buterin’s own writings concerning “The Surge.” PeerDAS itself implements one-dimensional sampling alongside erasure coding and provides succinct per-cell proofs. This allows nodes to validate data availability effectively without the need to download everything indiscriminately. This is precisely what makes the approach “pretty unprecedented” within the context of a live, high-value blockchain: it seeks to harmonize the principles of decentralization and throughput by minimizing the bandwidth and storage requirements for each node, while simultaneously upholding robust guarantees that the data genuinely exists and is accessible.

Despite its promise, this significant shift is not without its inherent risks. Buterin’s unwavering insistence on a meticulously careful rollout reflects the reality that Ethereum’s blob economy remains young, inherently volatile, and highly susceptible to abrupt shifts in demand. As L2s compete for precious capacity, fee dynamics can reverse direction rapidly. Furthermore, factors such as incomplete blobs, erratic usage patterns, and the side effects of Miner Extractable Value (MEV) further complicate accurate forecasting. The overarching promise of Fusaka is that PeerDAS possesses the capability to steer these dynamics towards sustainable growth, by enabling the network to scale data availability without burdening any single node with the responsibility of the entire chain, and crucially, by accomplishing this in a manner that maintains explicit and rigorously testable security assumptions.

At the time of writing, ETH was trading at $4,028.