DA Exposé: Commodification of DA

There are many data availability solutions, possibly too many, and for the current demand for applications, L2s, etc, the revenue opportunity for DA providers is just too small.

Of the total list of DA services on the market, only around 4-5 are live. Of these services, we can see a clear pattern: the price for data availability is dropping substantially. Data availability is being commodified in real-time, the costs for DA are dropping quickly, and there is little differentiation between these services. For instance, using Celestia over Ethereum for DA has minor differences. Today, the largest difference between the two chains is that Celestia’s DA capacity allows for more expressive virtual machines, like the SVM, but this difference isn’t meaningful as Ethereum and other competitors will likely catch up in capacity over time. DA capacity matters here as more expressive VMs cannot use Ethereum at this point in time. Eclipse uses Celestia, and Movement will use NEAR DA. 

Given this trend of commoditization, differentiation is becoming more important for DA providers seeking to maintain profitability. DA providers will need to think of new strategies to be profitable – and we’re seeing protocols like Nuffle and Metaplex stand out by essentially bundling data availability within a broader product suite. Nuffle is to offer cross-chain communication for its fast-finality service and multichain staking protocol, which is only possible by using NEAR DA. Nuffle will internalize its costs for NEAR DA to offer its Nuffle Protocol service, so we may see DA as being part of the bundle for some larger service in the future. On a similar note, Metaplex has created Aura, a DA network, a natural complement to it Metaplex developer tooling. For example, when a dev spins up a temporary NFT on Solana via the Metaplex Developer Platform, Metaplex DA could be a natural choice for storage.

Additionally, while Ethereum is the most expensive option currently, this is not expected to last long. Ethereum’s PeerDAS upgrade aims to increase the blob count target from 3 (with a maximum of 6) to 8-16. And this is only the beginning; there are plans for increased scalability, with the medium-term target in the future being 16 MB per slot. Currently, Ethereum’s 3 blob per block target means that the DA capacity is about .032 Mb/s. Without factoring in any changes in block time from the recently presented EIP-7781, just a shift to PeerDAS’ schematic would bring DA capacity to .085 Mb/s. According to the Ethereum foundation, the medium-term target is 16 MB per block, representing a substantial increase from the current target of 384 kB per block, and even the short-term PeerDAS increase to 1-2 MB per block.

This isn’t the only upcoming upgrade, though. Celestia’s recent Mammoth Mini testnet announcement aims to eventually push 1GB blocks. Already today, validators on the Celestia network can vote to increase the DA capacity of the chain, so while .16 Mb/s is the current capacity, this can increase to .6Mb/s per a vote from validators. All this is to say that the data is available and becoming available at incredible speeds. This rapid expansion in blobspace will drive a race to lower blob fees, so situations like the recent blob burn for Ethereum will not continuously occur. Moreover, Ethereum is still a competitor to Celestia, while Celestia is the better provider for DA strictly in terms of DA capacity and cost today, as Ethereum steps into its Surge roadmap, it may be able to match Celestia in capacity. At the same time, it is dubious to say that these will be used over a Data Availability Committee (DAC) structure. While DACs have more centralization concerns (though can be secured by a restaking layer), they offer cheaper services relative to Celestia or Ethereum because of this and may be more suitable to applications that do not care for censorship resistance, like gaming, which is one of the larger consumers of data availability services today.