Derive Community Proposal Launches zkVM Version V3 and Shuts Down Old Chain
The Derive community has released a new DIP (Derive Improvement Proposal) suggesting the formal deployment and launch of Derive V3, while migrating all positions and balances from V2 to the genesis state of V3.
The proposal calls for transferring funds locked in the existing cross-chain bridge to Ethereum L1 and gradually shutting down the currently independently running Derive Chain (an application chain built on the Optimism tech stack), indicating a shift in the protocol's overall architecture from "self-built L2 application chain" to directly anchoring settlement on the Ethereum mainnet.
V3 is a zero-knowledge exchange based on zkVM (zero-knowledge virtual machine) that ultimately settles on the Ethereum mainnet, employing a risk management and settlement mechanism of "off-chain matching + ZK proof," which is completely different from Derive's previous model relying on its own application chain validation nodes to process transactions.
Margin calculations, pricing, and settlement logic, which are computationally intensive, will all be executed within the zkVM, generating corresponding zero-knowledge proofs off-chain; Ethereum L1 itself will not re-execute these calculations but will only verify the validity of the proofs, significantly reducing the computational load on L1.
Each state change will produce a state difference (state diff) that will be published to Celestia, with this independent data availability layer ensuring that transaction state data can be publicly accessed and verified, rather than storing all data on the Ethereum mainnet, which is a direct integration with Celestia's modular data availability solution.
The core trade-off of this architectural migration is between "trust minimization" and "operational costs": previously, Derive Chain as an independent application chain required the protocol party to bear the operational costs of the chain's own sequencer, security assumptions, and cross-chain bridge risks; after migrating to the zkVM architecture directly anchored to Ethereum settlement, the protocol's security directly inherits from the Ethereum mainnet, and transferring cross-chain bridge funds to L1 also means that the custodian of user assets shifts from a combination of "application chain + bridge" to the Ethereum mainnet contract itself, theoretically reducing the risk exposure of smart contracts and cross-chain bridges to attacks. Beneficiaries include existing users holding V2 positions who were previously concerned about cross-chain bridge security, as well as market makers and clearing participants needing to directly interoperate with other protocols in the Ethereum ecosystem; short-term pressures will be faced by traders needing to coordinate asset migration, who may encounter liquidity interruptions during the migration window, as well as operators of existing Derive Chain-related infrastructure (sequencers, node services).
Source: Public Information
ABAB AI Insight
Derive was previously a decentralized derivatives trading protocol that rebranded from the options protocol Lyra Protocol, and early on chose to build an independent application chain (Derive Chain) based on the Optimism OP Stack tech stack to achieve lower trading costs and autonomous control over clearing logic. This V3 proposal represents a structural reversal of the "self-built application chain" route—shifting from pursuing on-chain autonomy to directly relying on the security of the Ethereum mainnet.
This migration simplifies the custody path of user funds from "user assets → cross-chain bridge → Derive self-built application chain" to "user assets → Ethereum L1 contract". The shortening of the funding chain means that the investments made in infrastructure such as validation nodes and sequencers for the independent application chain can gradually exit, redirecting resources to the computational infrastructure needed for zkVM proof generation—the protocol's capital expenditure focus shifts from "operating a chain" to "generating and verifying zero-knowledge proofs."
Similar adjustments in the direction of "shifting from self-built application chains to relying on Ethereum mainnet settlement" vary: dYdX is going in the opposite direction (migrating from Ethereum L2 to a self-built Cosmos application chain), while several on-chain derivatives experiments using zkVM architecture (based on Risc Zero, SP1, etc.) are following the same path as Derive. Currently, the DeFi derivatives track has not converged between the two technical routes of "self-built sovereign application chains" and "relying on Ethereum mainnet settlement + ZK proof cost reduction"; Derive has clearly chosen the latter this time.
Structurally, this represents a technological substitution: using zero-knowledge proofs to replace the consensus and execution layers of the application chain to obtain trusted computing guarantees. The mechanism allows zkVM to transfer margin calculations, pricing, and settlement—originally requiring on-chain repeated execution and consuming large amounts of gas—off-chain, only verifying a proof on-chain, enabling the protocol to retain the low cost and high throughput of off-chain computation without relying on the independent trust assumptions of a self-built chain, but rather directly inheriting the security of the Ethereum mainnet—this is a technological path alternative for current DeFi derivatives protocols to address the dilemma of "performance versus trust minimization."
ABAB News · Law of Cognition
The road you built yourself will eventually merge back into the main road.
The heavier the calculation, the lighter the proof must be.
The endpoint of trust minimization is to return trust to the greatest common divisor.