Do We Need Another L1? - Inside Monad’s Parallel EVM

Key Concepts

Monad: A new general-purpose Layer 1 (L1) blockchain focused on delivering high performance, fast finality, and a high degree of decentralization for the Ethereum Virtual Machine (EVM).
EVM Optimization: Monad's core engineering effort to significantly enhance the performance and efficiency of the EVM.
Decentralization: Defined technologically (control split among many accountable nodes) and socially (large number of watchdogs/contributors), with a practical emphasis on consumer-grade hardware for node operation.
Pipelining: A computer science technique applied across Monad's architecture to process multiple stages of work concurrently rather than sequentially, improving throughput.
Monad BFT: Monad's novel consensus mechanism incorporating pipelining.
Asynchronous Execution: Decoupling consensus and execution layers to run them in parallel.
Parallel Execution: Optimistically executing multiple transactions concurrently and then committing them serially, with re-execution for incorrect inputs.
JIT Compilation (Just-In-Time Compilation): Converting EVM bytecode into native machine code for more efficient CPU execution.
Monad DB: A custom-designed database optimized for native storage and efficient lookup of the Ethereum Merkel tree state on disk (SSDs).
Raptorcast: An efficient block propagation communication method for large blocks.
Consumer Hardware Chain: A blockchain designed to allow nodes to run on affordable, widely available hardware (e.g., a Costco MacBook).
MEV (Maximal Extractable Value): Value extracted by block producers through transaction ordering, inclusion, or censorship. Monad aims to mitigate toxic forms like sandwiching.
Shared Global State with Atomic Composability: Monad's design philosophy, emphasizing a single, integrated state layer where all applications can interact seamlessly.

Monad's Core Value Proposition and Purpose

Monad is presented as a significant engineering effort to bring new technologies to the Ethereum Virtual Machine (EVM), pioneering compatible and stackable improvements to enhance decentralization through software architecture. It aims to deliver performance, fast finality, and a high degree of decentralization right now, addressing struggles with slow finality in other ecosystems.

The fundamental purpose of a blockchain, in Monad's view, is to provide a means of coordination, transaction, value transfer, and asset issuance, enabled by a shared global state. This fosters self-verifiability and self-sovereignty, unlocking greater efficiency and opportunity globally, particularly in finance. The speaker emphasizes that crypto is about enabling a more efficient and opportunistic global economy, which requires a performant, decentralized, and permissionless Layer 1.

Why Monad as a Layer 1?

Monad is built as a Layer 1 (L1) rather than a Layer 2 (L2) because it focuses on critical optimizations needed at both the execution layer and the consensus layer. Layer 2s typically concentrate only on execution. Consensus, however, is crucial for a blockchain's decentralization and borderless nature, distributing network control across many global entities. The innovation at the consensus layer and its integration with execution is an "extremely important and underexplored aspect of crypto."

Monad's Approach to Decentralization

Decentralization is a core tenet for Monad, defined in two ways:

Technological Perspective: Control over the system is split among many independent entities (nodes), acting as "horcruxes" that hold each other accountable, enforce system rules, and ensure only valid state transitions occur. This trustless environment allows for more powerful applications.
Social Aspect: A large number of "watchdogs" contribute to the network by building applications, integrations, or simply observing, acting as the "white blood cells of the system."

A critical component of Monad's decentralization is the ability for "normal people" to run a node or validator permissionlessly from home, outside of data centers. This was a day-one constraint for Monad, aiming to be a consumer hardware chain, not a data center chain.

Specific Hardware Requirements for Monad Nodes:

32 GB of RAM
2 TB SSD
100 Mbps bandwidth (for a full node)
300 Mbps bandwidth (for a validator)
A reasonable CPU

These requirements are designed to be met by common devices like a "Costco MacBook," making node operation accessible and affordable (estimated machine cost: ~$1,500). This contrasts with many other high-throughput chains that effectively require data center infrastructure. Monad stands with Bitcoin and Ethereum in prioritizing this consumer-grade hardware accessibility.

EVM Optimization: The Six Major Improvements

Monad's three-and-a-half-year engineering effort focuses on optimizing the EVM, which is recognized as the dominant bytecode standard in crypto (over 80% of TVL, extensive tooling, and research). The project addresses fundamental inefficiencies in existing EVM implementations by introducing six major, stacked improvements to achieve over 10,000 Transactions Per Second (TPS) or 500 million gas per second at mainnet launch.

The overarching methodology is pipelining, a common computer science technique (illustrated with a laundry analogy) where different stages of work are processed concurrently rather than sequentially.

The six improvements, from highest to lowest layer, are:

Monad BFT (Byzantine Fault Tolerance): A new consensus mechanism that incorporates pipelining, specifically designed to address issues found in previous pipelined consensus protocols.
Asynchronous Execution: This improvement decouples the two main parts of a blockchain – consensus and execution – allowing them to run in parallel. In traditional blockchains, consensus waits for execution to complete, and vice-versa, leading to inefficiencies. Monad runs both in a pipelined fashion.
Parallel Execution: While transactions are officially defined to execute serially, Monad optimistically executes many transactions in parallel. It assumes correct inputs, produces pending results, and then commits these results in the original serial order, re-executing only if an input was incorrect. This leverages multi-core CPUs to identify and resolve database dependencies concurrently.
JIT Compilation (Just-In-Time Compilation): Smart contracts, typically written in Solidity, are compiled into EVM bytecode. This bytecode is not native machine code and requires a virtual machine for execution, similar to Java programs running on a JVM. Monad introduces a compiler that translates EVM bytecode into native machine code, significantly improving execution efficiency.
Monad DB: Ethereum stores its state in a Merkel tree, which enables concise verifiability (comparing Merkel roots to confirm identical state). However, updating and storing Merkel trees in generic databases (like LevelDB or RocksDB) is inefficient due to abstraction layers. Monad DB is a custom database specifically designed to store the Ethereum Merkel tree state natively on disk. This optimizes data lookups by packing relevant data closely together in pages, making retrieval much more efficient.
Raptorcast: A highly efficient block propagation communication method. It intelligently divides large blocks into chunks and sends them to different nodes, ensuring that all nodes receive enough chunks to reconstruct the original block quickly and reliably across the globe.

Monad's Performance Statistics

At launch, Monad aims to deliver:

Block Times: 400 milliseconds
Finality: 2-block finality, resulting in 800 milliseconds
Gas Limit per Block: 200 million gas
Throughput: 500 million gas per second
Transactions Per Second (TPS):
- Approximately 24,000 simple transfers per second (assuming 21,000 gas per transfer).
- Approximately 10,000 complex transactions per second (assuming 50,000 gas per transaction).

Monad vs. Other Blockchains

Monad vs. Solana: Monad is described as "the EVM's answer to Solana." While both share a spirit of optimizing for throughput, Monad fundamentally differs by prioritizing low hardware requirements for decentralization, enabling consumer-grade machines to run nodes. Solana, conversely, has high hardware requirements, effectively operating as a "data center chain" (e.g., 10 Gbps bandwidth for validators).
Monad vs. Ethereum's ZK Path: Ethereum's roadmap focuses on ZK scaling, where validators become verifiers of proofs, allowing for home-based verification but shifting the heavy computation elsewhere. Monad, however, believes in maximizing performance from a single node while maintaining decentralization. Monad DB's reliance on SSDs (cheap, performant, scalable to 30-100 TB) is key to this, as opposed to RAM-intensive solutions which are expensive and limit state growth. The speaker notes that some of Monad's innovations, like asynchronous execution, are also being explored by Ethereum researchers, and Monad's open-source work can inform the broader EVM ecosystem. For example, making asynchronous execution compatible with EIP7702 (account abstraction) was a "massive effort" whose solution is now available for study.

Decentralization, Latency, and MEV

The speaker agrees with Vitalik Buterin's critique that "low latency is the inherent cost of decentralization," acknowledging that pushing block times too low (e.g., 50ms) would lead to centralization due to physical limitations (e.g., 200ms transit time across the globe). Monad's 400ms block time is chosen as a "nice happy medium" that is "close to imperceptible" for human users while still allowing for globally distributed validators and resisting the centralizing forces seen in High-Frequency Trading (HFT) environments (e.g., co-location incentives, data center rent-seeking). Monad's "line" for latency is "wherever there is a compromise on decentralization."

Regarding MEV (Maximal Extractable Value), Monad aims to mitigate toxic forms like sandwiching and front-running.

Short-term: Monad's asynchronous execution means block leaders build on a lagged state, reducing immediate MEV opportunities. The default client uses a priority gas auction (similar to early Ethereum). Third-party MEV solutions on Monad are initially limited to "size two" bundles, which are generally less toxic (e.g., liquidations triggered by oracle updates).
Long-term: The goal is to implement pre-trade privacy, where block builders commit to a block without knowing the transaction contents until after commitment, thereby addressing the MEV problem at its root.

Monad's Design Philosophy and Launch

Monad champions a design with "shared global state that has atomic composability," which can be considered a "monolithic" approach in the sense of an integrated, seamless state layer, but without the negative connotation of requiring chunky, centralized hardware.

At mainnet launch, Monad will have no kill switches, admin keys, or multi-sigs, with all rules enforced by code and a decentralized validator set. The launch ecosystem will be fully backward compatible with existing EVM dev tooling (e.g., Tenderly, MetaMask, Chainlink, stablecoin issuers), though this required significant retrofitting by Monad Labs (now Category Labs) due to Monad's entirely new tech stack. While launching with a single client from Category Labs, the long-term vision is to foster multiple client implementations and an even larger, globally distributed validator set.

The MON Token

The MON token is integral to the Monad network as a Proof-of-Stake L1 blockchain.

Distribution: Monad is the first project to utilize Coinbase's new token sales platform, aiming for broader and fairer distribution to retail users, mitigating issues with airdrop hunters and bot farms. This reflects a shift in the regulatory environment towards more mature, mainnet-ready projects with stringent disclosures.
Utility: MON is used for gas fees and staking.
Staking & Issuance: Insider tokens are locked and ineligible for staking until unlocked, a unique aspect that directs staking rewards to public participants. The block reward is 25 MON per block, totaling approximately 2 billion MON per year from a total supply of 100 billion. This results in an initial 2% annual inflation rate, which slightly decreases over time. This low inflation rate is chosen to reward stakers adequately while minimizing dilution for non-stakers and reducing the effective cost of capital for the asset in DeFi.
Token Valuation: The speaker emphasizes that the most important factor for L1 token valuation is network effects and the value unlocked for end-users. Value will ultimately derive from transaction fee processing and the token's role as the native currency of a substantial and growing economy.

Vision for 2030 and Potential Failure Modes

If successful, Monad envisions a 2030 where:

Breakout decentralized applications are widely used.
Many more people globally have access to financial tools (e.g., dollars, competitive yield markets, capital markets) not limited by their local economies.
A more interconnected world coordinates on a decentralized, trustless layer.

The primary failure mode for Monad would be if the market ultimately "ends up being that no one cares about the properties that we value deeply" – specifically, the commitment to decentralization and high performance on consumer hardware. Execution and speed are identified as critical factors for success.

Mainnet Launch

Monad's mainnet is scheduled for November 24th. Users are encouraged to explore the validator map (searchable for "Monad validator map") to observe the network's decentralization (300+ geographically distributed validators) and performance.