| Authors | Sei Labs |
|---|---|
| Year | 2022 |
| Project | Sei Network |
| License | Apache-2.0 |
| Official Source | https://github.com/sei-protocol/sei-chain/blob/main/whitepaper/sei_whitepaper.md |
This page is an educational summary and analysis of an official whitepaper or technical paper, written for reference purposes. It is not a verbatim reproduction. CryptoGloss does not claim authorship of the original work. All intellectual property rights remain with the original author(s). The official document is linked above.
Sei is a Cosmos SDK blockchain described in a 2022 whitepaper by Sei Labs, purpose-built for trading applications. Sei’s thesis: general-purpose blockchains are overbuilt for most use cases but specifically underbuilt for trading — trading applications need sub-second finality, high throughput, and trading-specific protocol primitives that general chains do not provide.
Two core technical contributions:
- Twin-Turbo Consensus: Combines optimistic block processing (nodes speculatively execute before consensus finishes) with intelligent block propagation (transactions are pre-shared among validators before block proposal, reducing redundant data transfer)
- Native On-Chain CLOB: A built-in order book matching engine at the protocol layer, available to any contract without reimplementation
A later upgrade (Sei v2) added parallelized EVM execution with optimistic parallelism — marking Sei as one of the first chains to offer parallel EVM alongside full EVM compatibility.
> Whitepaper: github.com/sei-protocol/sei-chain/blob/main/whitepaper/sei_whitepaper.md
Publication and Context
Sei Labs was founded in 2022 by Jeff Feng and Jayendra Jog. The project raised $35M in its Series A (August 2022) led by Multicoin Capital and Jump Crypto. Sei mainnet (v1) launched September 2023.
By 2022, the Cosmos ecosystem had established IBC as a viable cross-chain standard, and Osmosis had proven that specialized Cosmos SDK appchains could build productive DeFi ecosystems. Sei targeted the next specialization: an appchain optimized not just for a specific DeFi application but for the entire trading primitive set (order books, matching, clearing).
The Injective whitepaper (2020) had similar goals; Sei differentiated through the consensus performance focus (twin-turbo) and later EVM compatibility (Sei v2 targeting Ethereum developers).
Twin-Turbo Consensus
Component 1: Optimistic Block Processing (OBP)
Standard Tendermint BFT:
- Proposer broadcasts block proposal
- Validators receive block → execute transactions → vote
- Wait for 2/3+1 votes → finalize
With OBP: Validators begin executing transactions as soon as they receive the block proposal, before consensus is reached. Since ≥95% of proposed blocks are valid (proposer has been selected by honest consensus), this speculation is rarely wrong. The result: consensus time and execution time overlap rather than being sequential.
This is similar to how CPUs use branch prediction — speculatively execute the likely path and only roll back on misprediction.
Component 2: Intelligent Block Propagation (IBP)
Standard Tendermint: the block proposer sends the full block (including all transaction data) to all validators. For high-TPS chains, this data transfer becomes a bottleneck.
With IBP: Before the block round begins, validators pre-share transactions via gossip. When the proposer proposes a block, they only send transaction hashes (not full data) in the proposal message. Validators already have the full transactions from prior gossip, so they can reconstruct the block without re-downloading data.
Combined: twin-turbo consensus targets <400 millisecond end-to-end block time (receive tx → finalize) on Sei v1.
Native CLOB Protocol Primitive
Sei includes a built-in on-chain CLOB at the module level:
- Any smart contract can place orders into the protocol’s order book via a CosmWasm interface
- Orders are matched by the protocol at the end of each block (batch clearing, similar to Injective FBA)
- Contracts build DeFi products (perpetuals, spot, options) using the protocol CLOB rather than implementing matching logic themselves
Difference from Injective: Injective’s CLOB is core to the chain’s primary use case (exchange). Sei’s CLOB is a library primitive that any contract can optionally use — the chain hosts general DeFi, not just exchange operations.
Sei v2: Parallelized EVM
Sei v2 (2024) introduced two major upgrades:
Full EVM Compatibility
Optimistic Parallelism
- Execute all transactions in a block simultaneously in parallel threads
- After execution, check for state conflicts (two txs reading/writing the same account)
- Re-execute conflicting transactions serially
For most DeFi workloads, most transactions are independent (touch different accounts/contracts) and never conflict — they execute fully in parallel. Only a small fraction require serial re-execution.
This is similar to Solana’s Sealevel and Aptos’s Block-STM but applied to EVM bytecode, which requires additional complexity because EVM does not natively declare access lists.
Key Technical Properties
| Property | Value |
|---|---|
| Consensus | Tendermint BFT + Twin-Turbo |
| Block time | <400ms (target) |
| Native CLOB | Yes (batch-matched per block) |
| EVM compatible | Yes (Sei v2) |
| Parallel execution | Yes (optimistic parallelism, Sei v2) |
| Smart contracts | CosmWasm (Rust) + EVM (Solidity) |
| Token | SEI |
| Mainnet | September 2023 (v1); v2 2024 |
Reality Check
Sei’s twin-turbo consensus achieved genuine performance improvements over standard Tendermint. The parallel EVM is a meaningful technical addition when deployed.
Challenges:
- DeFi ecosystem size: Sei’s DeFi ecosystem launched smaller than anticipated. Major trading apps did not prioritize Sei deployment, limiting the “fastest L1 for trading” narrative.
- EVM competition: By the time Sei v2 launched with EVM compatibility, it competed with Base, Arbitrum, Optimism, and dozens of other EVM environments with far larger existing ecosystems. Being a new EVM L1 is increasingly difficult.
- SEI token performance: SEI price performance after main listing was weak relative to other Cosmos ecosystem tokens.
- Twin-turbo in practice: Benchmarked performance in production conditions was lower than theoretical targets; real-world block times varied with network load.
Legacy
Sei is an important data point for the specialized appchain thesis. It demonstrated that Cosmos SDK chains can achieve sub-second finality with consensus-layer optimizations. The twin-turbo design influenced thinking about how to improve Tendermint performance without replacing the consensus algorithm entirely. Sei v2’s parallel EVM is a production reference implementation for EVM parallelization.
Related Terms
Research
- Sei Labs. (2022). Sei: The Fastest Layer 1 for Trading. github.com/sei-protocol.
— Primary whitepaper; describes twin-turbo consensus, native CLOB, and the trading-optimized appchain design philosophy.
- Buchman, E. (2016). Tendermint: Byzantine Fault Tolerance in the Age of Blockchains. MSc Thesis, University of Guelph.
— Tendermint BFT specification that Sei builds on; understanding Tendermint is prerequisite to understanding twin-turbo’s optimizations.
- Gelashvili, R., Spiegelman, A., Xiang, Z., Danezis, G., Li, Z., Malkhi, D., Xia, Y., & Zhou, R. (2022). Block-STM: Scaling Blockchain Execution by Turning Ordering Curse to a Performance Blessing. PPoPP 2023.
— Block-STM parallel execution (from Aptos); Sei v2’s optimistic parallelism is related in design and serves as benchmarking context.