EVM execution is no longer serial. Monolithic chains like Ethereum and Arbitrum process transactions one-by-one, creating a deterministic but slow global state. Parallel EVMs like Monad and Sei v2 break this by executing non-conflicting transactions simultaneously, decoupling throughput from consensus speed.
the-ethereum-roadmap-merge-surge-verge
Blog
Execution Layer Changes That Break Assumptions
The Ethereum execution layer is undergoing its most radical transformation since the Merge. EIP-7702, RIP-7560, and SSZ are dismantling long-held assumptions about accounts, transactions, and state. This is a technical deep dive for builders on what's changing and why it matters.
introduction
THE EXECUTION SHIFT
Introduction: The EVM's Unspoken Rules Are Being Rewritten
The core assumptions of EVM execution are being dismantled by parallelization, specialized chains, and new transaction formats.
Execution is becoming specialized. The 'one chain for all apps' model is fragmenting. Applications like dYdX and Aevo migrate to sovereign app-chains for custom fee markets and governance. This creates a landscape of vertical execution layers optimized for specific use cases.
Transaction intents bypass execution. Protocols like UniswapX and CowSwap abstract execution away from users. Instead of submitting a transaction, users sign an intent; a network of solvers competes to fulfill it off-chain, often via bridges like Across. The user's wallet no longer dictates the execution path.
Evidence: Monad's parallel execution engine demonstrates a 10,000 TPS benchmark for EVM-compatible smart contracts, a 100x leap over current serial processing limits on L2s.
key-trends
EXECUTION LAYER CHANGES
The Three Assumption-Breaking Trends
The monolithic execution model is being dismantled by specialized architectures that break core assumptions about speed, cost, and composability.
01
The Problem: Monolithic Bottlenecks
Single-threaded EVM execution serializes all transactions, capping throughput at ~50 TPS and creating volatile, auction-based fee markets. This makes predictable, low-cost execution impossible for mass adoption.
- Assumption Broken: Execution must be sequential and global.
- Result: High latency, unpredictable fees, and a hard scalability ceiling.
~50 TPS
EVM Ceiling
1000x+
Fee Spikes
02
The Solution: Parallel Execution Engines
Architectures like Aptos MoveVM and Sui's Object Model process independent transactions simultaneously. Solana's Sealevel runtime and Ethereum's emerging EIP-6480 (Parallel EVM) follow this trend, enabling orders-of-magnitude higher throughput.
- Key Benefit: 10,000-100,000+ TPS for non-conflicting transactions.
- Key Benefit: Dramatically lower latency (~400ms finality) and more stable fees.
100x
Throughput Gain
~400ms
Fast Finality
03
The Solution: Modular Execution & Specialized Rollups
Decoupling execution into sovereign layers (rollups) and further into dedicated app-chains (via Celestia, EigenLayer, AltLayer) allows optimization for specific use cases. dYdX (trading) and Immutable (gaming) exemplify this.
- Key Benefit: Custom VM & fee market per application (e.g., $0.001 stable fees).
- Key Benefit: Independent innovation cycles without monolithic chain upgrades.
$0.001
Stable Fee Target
1000+
Potential Rollups
deep-dive
THE NEW EXECUTION PRIMITIVE
Deep Dive: From Axioms to APIs
The EVM's core execution model is being superseded by new primitives that invalidate long-held architectural assumptions.
The EVM is no longer the center. The rise of intent-based architectures like UniswapX and CowSwap abstracts execution away from a single virtual machine. Users express a desired outcome, and a network of solvers competes to fulfill it, making the specific execution environment a secondary concern.
Block space is now a commodity. With parallel execution engines from Sui, Aptos, and Monad, the primary constraint shifts from sequential processing to state access patterns. This breaks the fundamental assumption that all transactions must be ordered and executed serially.
Execution is a delegated service. Protocols like EigenLayer and AltLayer enable restaking and rollup-as-a-service, turning execution layer security into a wholesale market. A rollup's security is no longer intrinsically tied to its own token or validator set.
Evidence: Arbitrum Stylus demonstrates this shift, allowing WASM-based programs to run alongside Solidity contracts, increasing throughput by 10-100x for non-EVM-native computations and fragmenting the execution monoculture.
EXECUTION LAYER SHIFT
Assumption Breakdown: Before vs. After
Comparing the core assumptions of a monolithic execution environment (Before) against the new reality of a modular, specialized execution landscape (After).
| Architectural Assumption | Before (Monolithic EVM) | After (Modular Execution) |
|---|---|---|
Execution is a Public Good | ||
State is Globally Synchronous | ||
Gas Price is the Sole Resource | ||
Settlement is On-Chain | ||
L1 is the Primary Liquidity Hub | ||
MEV is Extracted by Validators | ||
Client Diversity Ensures Censorship Resistance | ~50% Geth dominance | Proposer-Builder-Separation (PBS) required |
Throughput Scaling Path | L1 Hard Forks (e.g., London, Shanghai) | Rollups & Alt-DA (e.g., Celestia, EigenDA) |
protocol-spotlight
EXECUTION LAYER SHIFTS
Protocol Implications: Who Wins, Who Rewrites?
The move to parallel execution and shared sequencing invalidates core assumptions about MEV, composability, and state access.
01
Solana & Sui: The Parallel Execution Vanguard
Their native parallel execution models (Sealevel, Block-STM) become the de facto standard, not an exotic feature. Blockchains that serialize execution (EVM) face obsolescence without major re-architecture.\n- Winners: High-frequency DEXs, on-chain games, social apps.\n- Losers: Monolithic EVM chains with congested state access.
10,000+
TPS Potential
-90%
Contention Delay
02
The Death of Generalized Frontrunning
Shared sequencers (like Espresso, Astria) and fast finality break the dark forest of mempool MEV. Searchers must compete on inclusion, not latency.\n- Winners: End-users, intent-based protocols (UniswapX, CowSwap).\n- Losers: Simple arbitrage bots, traditional MEV relays.
~500ms
Finality Time
$1B+
MEV Redistributed
03
EVM-Equivalent Chains Must Pivot or Perish
Chains like Polygon, Avalanche C-Chain, and Arbitrum face a forking dilemma: maintain EVM compatibility and suffer performance ceilings, or break it for parallelism.\n- Solution: Aggressive L2 rollup integration or adopting new VMs (Move, Fuel).\n- Risk: $50B+ TVL trapped in architectures with inherent bottlenecks.
2-5s
Current Block Time
High
Migration Risk
04
App-Specific Rollups as the New Norm
The cost of sovereign execution collapses. Every major dApp (dYdX, ApeCoin) becomes its own rollup or hyperchain, optimized for its own state model.\n- Winners: Rollup-as-a-Service providers (AltLayer, Caldera), custom VM developers.\n- Losers: Generic smart contract platforms competing for "blockspace".
1000+
Rollups by 2025
-99%
OpEx vs. L1
05
The Shared Sequencer Power Grab
Control over transaction ordering becomes the most valuable real estate in crypto. Entities that operate dominant shared sequencers (EigenLayer, Near) capture economic value previously held by L1 validators.\n- Implication: A new trust layer emerges, potentially re-centralizing power.\n- Battleground: Decentralized sequencing proofs vs. institutional validators.
>$10B
Potential Fee Market
Oligopoly
Risk Profile
06
State Access Becomes a Paid API
With parallel execution, uncontended state reads are cheap, but hot spots (e.g., major NFT collection, popular token) require explicit pricing. The "gas" model evolves into a state rental market.\n- Winners: Dedicated state providers, data availability layers (Celestia, EigenDA).\n- Losers: DApps that assume uniformly cheap global state.
Micro-cents
Cold Read Cost
100x
Hot Spot Premium
future-outlook
THE ASSUMPTION BREAK
Future Outlook: The Modular Execution Endgame
Modular execution layers will render today's monolithic scaling assumptions obsolete.
Execution is a commodity. The current L2 race for dominance will shift to a competition for specialized, hyper-optimized execution environments. Monolithic chains like Solana and Ethereum L1s will compete on security and settlement, not raw throughput.
Shared sequencers fragment state. The rise of shared sequencer networks like Espresso and Astria decouples ordering from execution. This creates a new market for block space and forces execution layers to compete purely on VM performance and cost.
Parallel execution is non-negotiable. The performance ceiling for single-threaded EVMs is ~5k TPS. SVM, MoveVM, and FuelVM demonstrate that parallel execution and state access lists are prerequisites for the next order-of-magnitude scaling.
Proof aggregation changes economics. Validiums and sovereign rollups using proof aggregation networks like Risc Zero or Succinct will make fraud proofs and ZK-proving the dominant security model, not optimistic fraud windows.
Evidence: Arbitrum Stylus already demonstrates that a single sequencer can support multiple VMs (EVM, WASM), proving execution environments are becoming pluggable components, not foundational chains.
takeaways
EXECUTION LAYER CHANGES
Key Takeaways for Builders
The EVM is no longer a monolith. New execution paradigms are breaking fundamental assumptions about transaction ordering, state access, and finality.
01
The Problem: MEV is a Protocol Tax
Traditional block-building is a black box, allowing searchers to extract $1B+ annually from users via front-running and sandwich attacks. This is a direct tax on protocol activity and user trust.
- Solution: Adopt MEV-aware architectures like SUAVE, Flashbots Protect, or CowSwap's batch auctions.
- Benefit: Return value to users via MEV redistribution or eliminate it entirely through encrypted mempools.
$1B+
Annual Extract
>90%
Sandwichable Txs
02
The Solution: Parallel Execution is Non-Negotiable
Sequential EVM execution caps throughput at ~50 TPS and is the primary bottleneck for scaling. This forces protocols into unsustainable L2 fragmentation.
- Adopt: Runtimes with parallel execution like Solana's Sealevel, Aptos' Block-STM, or Monad's parallel EVM.
- Impact: Achieves 10,000+ TPS by processing non-conflicting transactions simultaneously, unlocking new application design space.
200x
Throughput Gain
<100ms
Latency Target
03
The Problem: Gas is a Broken Abstraction
Gas fees are volatile and unpredictable, creating terrible UX. The gas opcode also exposes a side-channel for MEV, revealing user intent mid-execution.
- Solution: Move to gasless transactions via account abstraction (ERC-4337) or adopt intent-based systems like UniswapX and Across.
- Benefit: Users sign what they want, not how to do it. Solvers compete on execution, leading to better prices and guaranteed success.
~$0
User Gas Cost
5-10%
Price Improvement
04
The Solution: Native Account Abstraction is Inevitable
EOAs are insecure and inflexible, requiring seed phrases and lacking programmability. This hinders mass adoption and smart contract wallet integration.
- Adopt: ERC-4337 for EVM chains or build on natively abstracted chains like Starknet and zkSync.
- Impact: Enables social recovery, session keys, gas sponsorship, and batch transactions as primitive features, not hacks.
~0
Seed Phrases
1-Click
Batch Txs
05
The Problem: State is a Centralizing Force
Global state growth (currently ~1TB+ for full nodes) creates prohibitive hardware requirements, centralizing node operation and harming decentralization.
- Solution: Implement stateless clients (Verkle Trees) or state expiry. Architect protocols to be state-minimal.
- Benefit: Reduces node requirements to <1 TB, enabling home staking and preserving the credibly neutral base layer.
1TB+
State Size
<1 TB
Target Size
06
The Solution: Intent-Based Architectures Win
Pushing raw transactions is like giving a taxi driver turn-by-turn directions. It's inefficient and exposes you to exploitation.
- Adopt: Design for intent-centric flows. Let users declare outcomes (e.g., "swap X for Y at best price") and delegate execution to competitive solvers via UniswapX, CowSwap, or Across.
- Impact: Optimal execution emerges from solver competition, abstracting away complexity and capturing cross-chain liquidity seamlessly.
~$10B
Volume (UniswapX)
Multi-Chain
Native Flow
ENQUIRY
Get In Touch
today.
Our experts will offer a free quote and a 30min call to discuss your project.
NDA Protected
Confidentiality24h Response
Time to ValueDirectly to Engineering Team
No Sales Fluff10+
Protocols Shipped
$20M+
TVL Overall