The term “Ethereum Layer-2” is generally thought to mean that the chain inherits Ethereum security. However, the lines are always blurry and remain intact, especially with the increase in OP stack chains using alternative data availability (DA) layers. These so-called “optims” settle for Ethereum, but post data elsewhere and introduce some additional trust.
Togluru Maharamov, a roll-up researcher and former scroll contributor, is a critic of the voice of loose L2 definition, declared on X on Wednesday.
“The best thing isn’t L2. Anyone who says otherwise is trying to make you a PSYOP.”
A good example of this is CELO, where BlockWorks is reported as a case study in which L1 becomes L2. In most cases, the Ethereum rollup posts the data to Ethereum itself. Originally via Calldata and now mostly via Blobs. This allows for useful properties of unauthorized state reconstruction and withdrawal from L2 to Ethereum Mainnet.
In contrast, Optimiums outsources DA to off-chain networks such as Eigenda, Celestia, etc. and relies on proofs to prove that data is available. They may settle for Ethereum using evidence of fraud, but Ethereum adds a layer of trust that cannot be directly verified.
Vitalik Buterin is free to use “L2” and in fact we are particularly in good company there, as we call CELO moving to “L2”, but the difference is more than semantics.
The Ethereum Scaling Tracker L2Beat now classifies most optims as “other.” This is all categories of chains that do not meet strict rollup standards. For example, CELO completed the migration in March 2025 to run on the OP stack on Eigenda.
According to L2Beat, two important checks for rollup status fail.
- There are fewer than 5 external actors who can submit fraud proof
- There are no DA bridges to post eigenda availability evidence on Ethereum
Results: If data is missing, there is no L1 intensive therapy.
The risk of moving to Ethereum in late 2024 is also categorized into this bucket for similar reasons – reliance on allowed challenger sets – posting data on Ethereum.
According to Sreeram Kannan, founder of Eigenlayer, Eigenda is “scaling protocols in Blob spaces,” while rollups can be thought of as “scaling protocols for execution.” In both cases, data and logic occur outside of Ethereum’s protocol rules, but only a rollup ensures that Ethereum can directly enforce validity.
“When a certificate from Eigenda appears, I believe that it means that the data is available,” he told BlockWorks.
That trust model does not have the chain enforcement today.
“If the DA claims that the data is available and that the data is unavailable, there is no request,” Kannan said.
So, is “L2” just marketing alongside Ethereum? Or should it reflect the ability of the protocol to enforce both accuracy and availability?
L2Beat’s current stance is that Ethereum needs to reject the roots of its poor state and be able to independently verify data availability. Other things require more trust, whether it’s a Dacoolam, the Governance Council, or a set of permitted challengers.
The best ones, either state verification or DA bridge, remain in the “Other” category. However, as the technique matures, some people may graduate to “barium and optimal” or even acquire rollup status. To do this, there is a need for a robust fallback mechanism and a way to turn DA failure into an “objectively attributed defect.” This is a difficult problem to crack.
