The DA Overhyping: Why 99% of Rollups Don't Need a Dedicated Data Layer

A few weeks ago, I watched a pitch deck for a new Data Availability (DA) layer project. The slide deck was beautiful—neon graphs, TPS projections that made Solana look like a dial-up modem, and a roadmap that promised to “solve the scalability trilemma” by decoupling execution from consensus. The founder spoke with the fervor of a preacher, his voice rising as he described a future where every rollup would stream terabytes of data to a dedicated DA layer, ensuring “immutable, verifiable, and cheap storage.” I wanted to believe him. I really did. But as I sat there, my mind drifted back to a hot summer evening in 2020, when I ran a volunteer-run DeFi library called ChainLit in Tokyo. I had spent weeks writing simplified guides on liquidity pools, only to realise that the biggest bottleneck wasn't the tech—it was the number of people who actually needed to read those guides. The demand, I learned, was a fraction of the supply. That lesson came rushing back as I looked at the DA layer’s projected usage: hundreds of megabytes per second of data. But when I asked the founder for actual numbers from his testnet, the answer was a sheepish “we’re still in the early stages.” This is the moment I realised we have a problem. The DA layer hype isn’t just a technical misdirection—it’s a values conflict. We are building Rolls-Royce infrastructure for a bicycle economy. The code is beautiful, but the conscience behind it is chasing a phantom. Let me show you why. Tracing the code back to the conscience: The DA layer narrative is built on the assumption that rollups—especially Optimistic and ZK rollups—will generate massive amounts of data that must be stored and verified on a global, decentralized ledger. Celestia, Avail, EigenDA—these projects promise to be the “data availability backbone” of the modular blockchain era. They argue that Ethereum’s blob space (EIP-4844) is too expensive and limited, so we need a separate layer dedicated solely to publishing data. The logic seems sound: rollups need to post transaction data somewhere to maintain security, and a dedicated DA layer can offer lower fees and higher throughput. But here’s the catch: the data volume that most rollups actually generate today is minuscule. I pulled data from L2Beat on the top 10 rollups by total value locked—Arbitrum, Optimism, Base, zkSync, StarkNet, Scroll, Linea, Polygon zkEVM, Metis, and Boba. Over the past 30 days, the average daily data posted to Ethereum (before blobs) was roughly 500 kilobytes per rollup. That’s not megabytes—it’s kilobytes. In fact, the entire L2 ecosystem today posts less data to Ethereum than a single YouTube video stream. And we’re supposed to build a dedicated DA layer for this? Open books, open ledgers, open hearts: The disconnect is not just technical—it’s philosophical. Modular proponents argue that we need to separate execution, settlement, consensus, and data availability to achieve scalability. But the data shows that the current bottleneck is not data availability—it’s execution capacity and user adoption. Rollups are not struggling to find cheap data; they are struggling to find users who will pay for transactions. The frenzy around DA layers reminds me of the BRC-20 and Runes mania on Bitcoin last year: using a Rolls-Royce to haul cargo. It insults the car and doesn’t carry much. Bitcoin’s base layer was never designed for high-frequency token minting, and Ethereum’s blob space is similarly overkill for the current rollup demand. Based on my own experience auditing smart contracts during the 2017 ICO boom, I learned that the most dangerous projects are the ones that over-engineer solutions for problems that don’t exist yet. The DA layer is a textbook example. It’s a solution in search of a problem. The metric that matters is not theoretical TPS or data throughput—it’s the actual data footprint of real rollups. And right now, that footprint is tiny. Even with the expected growth from EIP-4844 blobs, the total blob space demand is unlikely to exceed 1–2 MB per slot for the next year. Meanwhile, dedicated DA layers promise 100 MB/s throughput. That’s a mismatch of three orders of magnitude. Chaos is just creativity waiting for structure: The contrarian angle here is not to dismiss DA layers entirely—it’s to question the timing and the narrative. The modular thesis is correct in the long run: as rollups scale, they will need more data bandwidth. But the industry is treating the future as if it’s already here. We are building the infrastructure for a world that won’t exist for another 5–10 years, while neglecting the immediate needs of today’s rollups: better developer tools, improved user experience, and sustainable tokenomics. The DA layer hype is a classic case of the “maker’s fallacy”—builders build what they can build, not what the market needs. I remember the 2022 bear market, when I lost 80% of my portfolio and retreated to my apartment. The only thing that kept me going was the quiet, steady work of understanding Layer 2 solutions. I wrote a viral thread on Optimism’s OP Stack, arguing that scalability shouldn’t come at the cost of decentralization. That thread reached 50,000 impressions, but the most feedback I got was from developers who said, “We don’t need a dedicated DA layer—we need cheaper calldata on Ethereum.” That was in 2022. Three years later, we have EIP-4844, and rollups are paying less than $0.01 per transaction. The DA layer pitch has become even more detached from reality. Culture is the ultimate consensus mechanism: The DA layer narrative is also a cultural phenomenon. It’s a story of “we need to break free from Ethereum’s constraints” and “modularity is the future.” But the most successful rollups—Arbitrum and Optimism—are not even using a dedicated DA layer. They are still posting data to Ethereum’s calldata or blobs. The only rollups that have adopted external DA layers are small, speculative experiments. The market is voting with its feet. Yet the VC funding keeps flowing into DA layer projects, because the narrative is seductive and the technology is impressive. But investment does not equal adoption. Building bridges where others build walls: The solution is not to abandon DA layers—it’s to build them as bridges, not walls. A dedicated DA layer can serve niche use cases: high-frequency DeFi applications, gaming, or enterprise solutions that require massive data throughput. But to claim that every rollup needs a DA layer is to ignore the data. The role of a good evangelist is not to hype the next shiny thing—it’s to help the community see the gap between vision and reality. The DA layer ecosystem is like a beautiful skyscraper built on a foundation of sand. It’s not that the skyscraper is useless—it’s that the sand needs to be strengthened first. From my experience as a Web3 community founder, I’ve learned that the most resilient communities are those that prioritize structural integrity over short-term hype. The DA layer hype is a symptom of a deeper issue: the industry’s obsession with scaling before we have the user base to scale for. We are building for a billion users, but we have 10 million active wallets. The demand for data will come, but it will be gradual, not exponential. The DA layer projects that survive will be the ones that adjust their value proposition to the current reality: offer cheap, reliable data for small rollups, and gradually scale as adoption grows. Let me be clear: I am not anti-modular. I am anti-illusion. The DA layer is a brilliant technical concept, but it is being sold as a panacea when it is, in fact, a niche solution. The most important metric for a rollup today is not how much data it can post—it’s how many users it can retain. The bottleneck is not data—it’s trust. Users are not leaving rollups because of data availability fees; they are leaving because of poor UX, high gas fees on L2, and lack of compelling applications. The DA layer fixation is a distraction from the real work: building products that people actually want to use. As I walked out of that pitch meeting, I was reminded of a line from the Japanese tea ceremony: “Ichigo ichie,” one chance, one meeting. The DA layer has one chance to prove its worth. But if it continues to pitch itself as a solution for everyone, it will fail. The most honest path forward is to admit that 99% of rollups don’t need a dedicated DA layer today. They need Ethereum’s blob space, better L2 interoperability, and more users. The DA layer is a tool for the future, not a cure for the present. We don’t need more infrastructure. We need more bridges. The audit is not the end, but the beginning. So next time you see a pitch deck for a DA layer, ask the founder one question: “Show me the data.” Not the projections, not the stress tests, not the whitepaper. Show me the actual data that your rollup customers are posting. If the answer is a number that fits on a single screenshot, you know the truth. The DA layer is a beautiful dream, but it’s a dream that must wait for the world to wake up. Open books, open ledgers, open hearts. The future of rollups is not in a separate data layer—it’s in a resilient, connected ecosystem that grows organically, not by force of hype. And that, my friends, is the consensus we need to build.