The L2BEAT report on Lighter has been released, and I would like to discuss L2BEAT's rating criteria and Lighter's rating situation here. L2BEAT evaluates a Layer 2 based on two main aspects: risk analysis and the stage of the Rollup. ————————————————————— 1⃣ Risk analysis, which is like a big pizza pie for L2BEAT, evaluates in 5 aspects with green, yellow, and red (good, medium, poor). 1) Sequencer failure What happens if the L2 sequencer fails? Lighter is rated green; users can initiate transactions through Ethereum, and if the sequencer is down for more than 18 days, they can withdraw funds directly using Ethereum. Here, Lighter has an interesting setting in its code: if the sequencer ignores any user deposits, withdrawals, order openings, or closings, it enters "desert mode" and shuts down to prevent hacker operations. Lighter is rated green ✅ 2) State validation For state validation, Lighter relies on Succinct's ZK Proof, so this point is error-free. For more on Succinct's zkVM, please read. Lighter is rated green ✅ 3) Data availability Lighter uses Ethereum's official blobs data structure for the DA layer, and all data can be verified on the Ethereum chain. Lighter is rated green ✅ 4) Proposer failure If a transaction submitted by a user on Lighter is rejected, Lighter will shut down, and then the user can exit trustlessly on Ethereum by submitting a zero-knowledge proof of funds. Lighter is rated green ✅ 5) Exit window If a smart contract is upgraded normally or maliciously, how long is the exit window for users to withdraw their funds? More than 7 days is yellow, more than 30 days is green. Here, Lighter has no upgrade exit window, so it is rated red. This is quite normal at the current stage, as mature L2s like Base and Arbitrum have not reached green either. Lighter is rated red ❌ ————————————————————— 2⃣ The stage of the Rollup is a Layer 2 rating standard proposed by Vitalik and practiced by L2BEAT, which includes: Stage 0: The project is still heavily reliant on centralized control (like a centralized sequencer). Even if some mechanisms are in place, they may not fully prevent the sequencer from abusing power or intervening. Stage 1: "Intermediate stage"—proof, challenge, and exit mechanisms are present, but some "human intervention" capability remains, and the Exit Window must be greater than 7 days. For example, Arbitrum uses a security committee with enough external personnel, and Arbitrum's Exit window is 10 days, so it is rated Stage 1. Optimism stack has an Exit window of 0 days, so it is Stage 0. Stage 2: "Ultimate goal"—in this stage, centralization is minimized, and the system mainly relies on automation/smart contracts/permissive verification to ensure security and correctness, achieving complete decentralization in Layer 2. Very few Layer 2s on the market reach Stage 2. Looking back at Lighter, it uses a fully centralized sequencer, and Lighter's current security committee only uses a 3/5 multi-signature, which is not enough and consists mostly of insiders, and the Exit window is 0, so Lighter is undoubtedly rated Stage 0. ❓ Does being rated Stage 0 mean it's not viable? Not necessarily. Although Base and Optimism are rated Stage 0 by L2BEAT, they carry a significant amount of Layer 2 activity. ————————————————————— In addition, during the rating process, Lighter was also found to have: a. Some of Lighter's contracts are unaudited. b. Key circuit code is not open-sourced. c. The sequencer fully trusts the prices from external oracles, but these oracles have not yet been certified. d. MEV issues (a problem common to centralized sequencers). The team has promised to open-source solutions for issue b within 1-2 weeks, and they also plan to address issue c (most likely by adopting mature oracles). ————————————————————— Summary and views Lighter's rating on L2BEAT is quite average because Lighter, as a PerpDEX Appchain, is being rated alongside specialized Layer 2s. The team optimizing the code may elevate it to all green or Stage 1, but achieving Stage 2 is very difficult, as handling high-frequency trading speed clearly contradicts communication with decentralized nodes. Additionally, there are many areas in the code that need to be open-sourced and audited, and the team needs to ensure proper disclosure. As an AppChain, Lighter has already performed well, with its TVL nearing that of many well-known Layer 2s even before the TGE.