Monad is a Layer-1 blockchain that runs unmodified Ethereum bytecode at roughly 10,000 transactions per second with one-second block times and single-slot finality. It does this by rebuilding the four pieces of a blockchain client that have always bottlenecked the EVM — execution, consensus, mempool, and state storage — while keeping bytecode and RPC compatibility so any Solidity contract or MetaMask user works on day one.
​

The project raised $225M in April 2024 in a round led by Paradigm with participation from Dragonfly, Electric Capital, Greenoaks, and others, the largest infrastructure round of that cycle. Public testnet went live in February 2025; mainnet launches in 2026 with the native token MON.
​

Monad is a high-throughput, EVM-compatible Layer-1 designed by ex-Jump Trading engineers. It targets 10,000 TPS by running transactions in parallel using optimistic concurrency control, decoupling execution from consensus (asynchronous execution), replacing LevelDB with a custom Merkle-trie database called MonadDB, and pipelining a HotStuff-derived consensus called MonadBFT. Solidity contracts deploy without changes; wallets and tooling work as-is.
​

Monad was founded in 2022 by Keone Hon, James Hunsaker, and Eunice Giarta — three engineers from Jump Trading's crypto desk. Hon led high-frequency trading systems at Jump for seven years; the Monad team's pitch is that the EVM is a great developer surface but a slow runtime, and most of the slowness is in the client, not the spec. Rebuilding the client unlocks the EVM without breaking the contract layer. Paradigm led the seed and the Series A; the April 2024 round at $225M from Paradigm, Dragonfly, Electric Capital, Greenoaks, Coinbase Ventures, and others made Monad one of the best-capitalized L1 launches of the cycle.
​

Monad's throughput comes from four engineering choices that work together. None of them change the EVM specification — they change how a node executes against it. Together they take a modern server (16-core, NVMe, 32 GB RAM) from the EVM's typical ~100 TPS to a target of 10,000 TPS at one-second blocks.

Ethereum executes transactions one at a time so that any contract can read the result of the previous transaction. Monad runs them concurrently using optimistic concurrency control: every transaction executes speculatively against a snapshot, and the scheduler commits results in canonical order. If two transactions touched the same state, the later one re-executes against the first's output. In practice most blocks have low conflict rates, so almost all transactions clear on the first pass.
​

In Ethereum, each block is fully executed before the next consensus round begins. Monad separates the two pipelines: validators agree on the ordering of transactions first, then execute the previous block while the next round is already running. This parallelism between consensus and execution roughly doubles effective throughput on the same hardware.
​

Most EVM clients store state in LevelDB or RocksDB, which were designed for log-structured workloads, not the random-access patterns of Merkle Patricia tries. MonadDB is a purpose-built database that stores the trie natively, uses asynchronous I/O directly against NVMe, and parallelizes state reads across the storage device. Reading a slot becomes one disk seek instead of multiple LevelDB lookups.
​

Monad's consensus is a pipelined variant of HotStuff (the BFT family also used by Diem and Aptos). It targets one-second block times with two-block finality (so roughly two seconds end-to-end), with a leader-based round structure that overlaps the propose, vote, and commit phases. Validator counts on testnet have run in the low hundreds.
​

Monad runs EVM bytecode unchanged, which is the strict definition of bytecode-equivalent. Solidity contracts compile and deploy with the same toolchain (Hardhat, Foundry, Remix), MetaMask connects with a chain ID switch, and standards like ERC-20, ERC-721, and ERC-4337 work out of the box. The only differences are gas-metering refinements and the new precompiles Monad adds for parallel-execution hints. For developers, the migration cost from Ethereum or any EVM L2 is effectively zero.
​

Monad sits in the same throughput tier as Solana and the Move-language chains, but with a different developer surface. The trade-off: keep the EVM ecosystem (where most developers, contracts, and tooling already live) and rebuild the runtime, instead of building a faster runtime from scratch and asking developers to learn Rust or Move.

TPS figures are project-stated targets or sustained mainnet throughput; real-world workloads vary. Solana's stated 65,000 TPS is a theoretical peak — sustained throughput is closer to 3,000–5,000 TPS on production traffic.
​

MON is Monad's native token and serves three roles: paying gas fees on every transaction, staking by validators to secure consensus, and governance. The full tokenomics — supply, distribution, vesting — will be published with the mainnet launch in 2026. Testnet uses MON faucets that have no real value.
​

Anything you can build on Ethereum — but with throughput high enough to do things the EVM has historically excluded. The early ecosystem on testnet has clustered around three categories. High-frequency DeFi: orderbook DEXs (Kuru, Madness Finance), perps, and prediction markets that need sub-second matching. Onchain games: games where every action is a transaction (Fantasy Top, Curvance) and gas needs to round to zero. Stablecoin payments: low-fee transfer rails that compete with Tron's TRC20 USDT throughput on EVM rails. Monad's low fees and EVM compatibility make it a candidate for stablecoin issuers who want EVM compatibility without Ethereum L1 costs — see our overview of the best stablecoin L1 chains in 2026.
​

Monad is one of several high-profile L1 launches landing in 2026, each with a different bet. Monad bets on parallel-EVM throughput. MegaETH bets on real-time L2 with 100k TPS via specialized sequencer hardware. Plasma is Tether's stablecoin-native L1 with zero-fee USDT transfers. Converge is Securitize and Ethena's institutional RWA chain. And the Hyperliquid stack splits its custom L1 from its EVM-compatible HyperEVM. Different optimization targets, same broad thesis: the EVM is the developer surface but its 2015 runtime is the bottleneck.
​

The public testnet launched February 2025 and has been running since. Mainnet is targeted for 2026; the Monad Foundation has not committed to a specific date, but the testnet has been processing real workloads (DEXs, NFT mints, payment apps) for over a year of public stress-testing. Developers can deploy today against the testnet using docs.monad.xyz; tokens, mainnet-only contracts, and final tokenomics arrive at TGE.
​

Monad is a Layer-1 — its own validator set, its own consensus, its own state. It is not an Ethereum rollup and does not post data to Ethereum. This is a deliberate choice: rolling up to Ethereum would inherit Ethereum's data-availability costs, which would cap throughput well below Monad's 10k TPS target. Monad gets EVM compatibility from the spec, not from Ethereum's settlement.
​

Architecture details: official Monad documentation at docs.monad.xyz and the Monad blog at monad.xyz/blog. Funding figures: Paradigm's April 9, 2024 announcement of the $225M round and reporting in The Block and Bloomberg. Comparative TPS and block-time figures: each project's own documentation and DeFiLlama as of May 2026. Founder background: Jump Trading staff page archives and Keone Hon's public talks. Testnet metrics: testnet.monadexplorer.com. Token mechanics not yet published; figures in this article are restricted to what the Monad Foundation has confirmed publicly. Snapshot date: 2026-05-04.
​