The numbers are seductive. Since the Dencun upgrade went live on March 13, 2024, over 2.5 million blobs have been posted to Ethereum by various rollups. The average fee per blob has remained under 0.001 ETH, a fraction of what calldata cost before the upgrade. Developers celebrate. L2s are thriving. The narrative is clear: Ethereum is scaling.
But I've been auditing cryptographic systems long enough to know that low fees in a low-utilization environment are not a proof of success—they are a proof of nothing. The real test comes when demand returns. And based on my experience analyzing the fee dynamics of the 2017 ICO era and the 2021 DeFi summer, I can tell you exactly what's coming: within two years, blob data will be saturated, and every rollup gas fee will double again.
Context: The Dencun Promise
Dencun introduced EIP-4844, which created a separate data layer for blobs—temporary, large chunks of data that rollups can post to Ethereum without competing with regular transaction calldata. The idea was to decouple data availability costs from execution costs, giving rollups a cheap highway for their proof data. And it worked. Optimism and Arbitrum saw their per-transaction fees drop by over 90% overnight. Base, Coinbase's L2, exploded in usage, processing over 1 million transactions per day within weeks.
But here's the uncomfortable truth that no one in the marketing departments wants to admit: the blob market is a fixed-bandwidth resource. Ethereum can only process a limited number of blobs per block—currently targeted at 3, with a maximum of 6. The system is designed to be stable, but it is not designed to be elastic. When demand exceeds supply, the fee market kicks in, and prices rise.
Core Analysis: The Saturation Curve
Let me take you through the math. At the current rate of approximately 8,000 blobs per day, we are using about 40% of the theoretical capacity (assuming 6 blobs per block at 7,200 blocks per day = 43,200 blobs max). That seems comfortable. But consider the growth trajectory. Every major rollup is planning to scale its user base. Arbitrum is targeting 10x growth in the next year. Optimism is building its Superchain. zkSync, Scroll, Linea—they all have aggressive onboarding plans.
During my time auditing whitepapers for the Paris Protocol Defense in 2017, I learned that teams always underestimate the demand side. They model linear growth, but adoption follows a logistic curve. When the network effects kick in, usage explodes. I saw it with ICOs, I saw it with DeFi, and I will see it with rollups.
Assume a conservative 3x growth in blob demand over the next 18 months. That puts us at 24,000 blobs per day, or 55% of maximum capacity. But the real issue is peak congestion. Rollups batch their proofs on a schedule—often every 15 minutes or every hour. That means blob demand is not uniform; it spikes. During high-activity periods like a DeFi liquidation event or a popular NFT mint, multiple rollups will try to post blobs simultaneously. The fee market will react not to the average, but to the peak.
I've run the numbers using a simple model based on historical Ethereum gas fee spikes. When blob utilization reaches 70% of peak capacity, the average blob fee could increase by 10x. At 90%, it could be 100x. That means a transaction that costs $0.01 today could cost $0.10 or $1.00—not catastrophic, but enough to kill the user experience for micro-transactions, gaming, and social applications. The very use cases that L2s are supposed to enable.
And here is the contrarian angle: the rollup teams know this, but they are not incentivized to solve it collectively. Each rollup is a separate entity competing for users and TVL. Blob space is a shared resource, but there is no governance mechanism to allocate it efficiently. The Ethereum community, led by the core developers, has focused on the technical layer—making blobs work—but has ignored the social layer: how do we govern access to this scarce resource?
Contrarian: The Governance Blind Spot
During the DeFi Community Bridge workshops I ran in Paris in 2020, I saw firsthand how a lack of inclusive governance leads to capture by the most vocal, not the most needy. The same is happening here. The blob fee market is a pure price mechanism, which favors the wealthiest rollups—those with the deepest treasuries and the highest transaction volumes. Smaller, community-driven L2s will be priced out. The result is a centralization of data availability, which undermines the very decentralization that Ethereum is supposed to protect.
Code is law, but people are the soul. The code of EIP-4844 is elegant, but the soul of the Ethereum network is its community. If we do not create a governance framework that ensures fair access to blob space, we will repeat the mistakes of the 2021 gas wars, where only the rich could transact on the base layer. The difference is that this time, we are building the infrastructure for billions of users—and we are already building in inequality.
I propose a different path: a DAO-based blob allocation system that combines market pricing with a community-weighted allocation for non-profit and public-good rollups. This is not a radical idea. We already have similar mechanisms in other shared resource environments, like the Ethereum Foundation's allocation of MEV rewards to public goods. The technology exists, but the will does not.
Takeaway: The Weavers Must Act
We are still early. The blob market is not yet saturated, the fees are still low, and the narrative is still positive. But the ethical guarddog in me sees the pattern. I have been through three market cycles, and I have watched good projects fail not because of bad code, but because of bad governance. The Ethereum community has a unique opportunity to design the rules of the blob economy before the crisis hits. If we wait until the fees double, it will be too late—the powerful rollups will have already captured the resource, and the smaller players will have no choice but to leave.
"The future of Ethereum is not a technical problem. It is a governance problem. And we, the community weavers, must start the conversation now. Listen more than you code. And when you do code, code for the collective, not just for the protocol.