Execution is the new commodity. The primary value is shifting from raw throughput to developer experience, measured by deployment speed and operational simplicity. This is the core thesis behind EVM rollups like Arbitrum and Optimism.
the-ethereum-roadmap-merge-surge-verge
Blog
Execution Layer Changes That Improve Developer UX
Forget the Surge. The real Ethereum revolution is in the Execution Layer. We analyze how EIP-7702, RIP-7560, and Verkle Trees are systematically dismantling the EVM's worst developer friction, from gas sponsorship to state management hell.
introduction
THE DEVELOPER PRIMITIVE
Introduction
Ethereum's execution layer is evolving from a raw compute engine into a refined platform that abstracts infrastructure complexity.
The EVM is a UX bottleneck. Its synchronous, single-threaded model forces developers to manage gas, nonces, and failed transactions. New paradigms like intent-based architectures (UniswapX, CowSwap) and parallel execution (Solana, Monad, Sui) abstract these concerns.
Account Abstraction (ERC-4337) is the catalyst. It decouples transaction validity from EOAs, enabling sponsored transactions, session keys, and batched operations. This reduces user friction and shifts gas management to dApp developers.
Evidence: Over 5.8 million ERC-4337 smart accounts have been created, with bundlers processing millions of UserOperations, demonstrating demand for this abstraction layer.
key-trends
EXECUTION LAYER CHANGES
The Three-Pronged Attack on Dev Friction
Modern execution environments are moving beyond raw speed to directly address the core pain points of building decentralized applications.
01
The Problem: Gas Abstraction
Users need native tokens to pay for transactions, creating massive onboarding friction and limiting application design.
- Solution: Account Abstraction (ERC-4337) enables sponsored transactions, gasless onboarding, and session keys.
- Key Benefit: Users interact with dApps using any token, or no token at all.
- Key Benefit: Developers can design novel fee models and subscription services.
~0
User Gas
ERC-4337
Standard
02
The Problem: Parallel Execution
Sequential block processing creates unnecessary bottlenecks, limiting throughput and increasing fees during congestion.
- Solution: Solana Sealevel & Sui Move process independent transactions concurrently, unlocking massive scalability.
- Key Benefit: 10,000+ TPS for non-conflicting operations like NFT mints or DEX swaps.
- Key Benefit: Predictable performance and lower fees under load, critical for consumer apps.
10,000+
TPS
Solana/Sui
Leaders
03
The Problem: State Bloat & Cost
Storing application state on-chain is prohibitively expensive, forcing devs into complex, insecure off-chain workarounds.
- Solution: Stateless Clients & Verkle Trees separate execution from state storage, dramatically reducing node requirements.
- Key Benefit: ~99% reduction in state growth pressure, enabling lighter nodes.
- Key Benefit: Paves the way for single-slot finality and cheaper contract storage, a core Ethereum roadmap goal.
-99%
State Growth
Verkle
EIP-6800
deep-dive
THE EXECUTION LAYER
Deconstructing the Pain: From EOA Hell to State Sprawl
Account abstraction and state management are the core execution-layer changes that will define the next generation of developer experience.
Account abstraction kills EOA hell. The Externally Owned Account (EOA) model is a UX dead-end, forcing users to manage private keys and pay gas natively. ERC-4337 and native implementations on chains like Starknet and zkSync enable smart contract wallets, allowing for social recovery, gas sponsorship, and batched transactions.
State sprawl cripples performance. Every new contract and user bloats the global state, increasing node sync times and hardware costs. Solutions like Verkle trees (Ethereum) and stateless clients aim to compress proof sizes, while Solana's aggressive state rent and Aptos' parallel execution demonstrate alternative architectural trade-offs.
Parallel execution is non-negotiable. Sequential processing, as in Ethereum's EVM, wastes hardware. Solana's Sealevel, Aptos' Block-STM, and Monad's parallel EVM show that concurrent transaction execution is the only path to scaling throughput without fracturing liquidity.
Evidence: The numbers don't lie. A Solana validator requires 128GB of RAM just to hold state, while an Ethereum archive node needs over 12TB. This is the direct cost of unchecked state growth that new execution layers must solve.
EXECUTION LAYER
Dev UX Upgrade Matrix: Impact & Timeline
Comparison of key Ethereum execution layer upgrades and their direct impact on developer experience, measured by concrete metrics and capability unlocks.
| Feature / Metric | Pre-EIP-1559 | Post-EIP-1559 / Merge | Post-EIP-4844 / Dencun | Post-4844 + Verkle (Prague/Electra) |
|---|---|---|---|---|
Gas Price Predictability | Volatile, First-Price Auction | Base Fee + Priority Tip Model | Base Fee + Priority Tip Model | Base Fee + Priority Tip Model |
Max Blob Data per Block | ~90 KB (calldata) | ~90 KB (calldata) | ~1.3 MB (blobs, ~0.375 MB full nodes) | ~1.3 MB (blobs, ~0.375 MB full nodes) |
L1 Calldata Cost for L2s | ~$1-10 per tx (variable) | ~$1-10 per tx (variable) | < $0.01 per tx (target) | < $0.01 per tx (target) |
State Growth Pruning | Archive Nodes Only | Archive Nodes Only | EIP-4444 (Execution Clients) | Stateless via Verkle Proofs |
Dev Client Sync Time | Days to Weeks | Days to Weeks | Hours to Days (EIP-4444) | < 1 Hour (Verkle Stateless) |
Witness Data for Validation | Full State Trie | Full State Trie | Full State Trie | ~1 KB Verkle Proof |
MEV Protection for Devs | Manual |
|
| Native PBS Integration (in-protocol) |
protocol-spotlight
EXECUTION LAYER UX
Who Wins? The New Primitive Landscape
The next wave of adoption will be won by chains that make building feel like Web2, without sacrificing Web3's core value propositions.
01
The Parallel EVM Thesis
Sequential execution is the bottleneck. Parallel EVMs like Monad, Sei, and Solana process independent transactions simultaneously, unlocking massive throughput.\n- Developer Benefit: No more gas wars or frontrunning for non-conflicting ops.\n- User Benefit: Predictable, low latency (~200-500ms) and stable fees, even during mempool congestion.
10k+
TPS Potential
~200ms
Finality
02
Account Abstraction (ERC-4337) Wallets
EOAs are a UX dead-end. Smart contract wallets like Safe, ZeroDev, and Biconomy abstract away seed phrases and enable sponsored transactions.\n- Developer Benefit: Offload gas fees, enable social logins, and batch operations in a single tx.\n- User Benefit: No more ETH for gas, seamless onboarding, and programmable security (e.g., 2FA, spending limits).
~$0
User Gas Cost
1-Click
Onboarding
03
Intent-Based Architectures
Requiring users to specify how (complex transactions) is a failure. Intent-centric systems like UniswapX, CowSwap, and Across let users declare what they want.\n- Developer Benefit: Build order-flow auctions and delegate routing complexity to specialized solvers.\n- User Benefit: Guaranteed best execution, MEV protection, and failed transaction refunds, all from a simple swap interface.
~20%
Better Prices
MEV-Proof
Execution
04
Modular Rollup Stacks
Monolithic L1 development is over. Stacks like Arbitrum Orbit, OP Stack, and zkStack let teams launch app-specific rollups in weeks.\n- Developer Benefit: Customizable data availability (Celestia, EigenDA), sequencers, and governance without forking a codebase.\n- User Benefit: Native token for gas, ultra-low fees, and tailored security models for the application's needs.
Weeks
To Launch
-99%
vs L1 Cost
05
State Expiration & Stateless Clients
Perpetual state growth is a scaling killer. Proposals like Verkle Trees (Ethereum) and stateless validation separate current state from historical data.\n- Developer Benefit: Enables lightweight clients, removing hardware requirements as a barrier to running a node.\n- User Benefit: Preserves decentralization and censorship resistance while enabling sustainable long-term chain growth.
TB → GB
State Size
Raspberry Pi
Node Hardware
06
Universal Gas Tokens
Paying gas in the chain's native token is a liquidity fragmentation tax. Systems like Particle Network's Universal Account and LayerZero's Omnichain Fungible Token (OFT) standard enable paying for any chain's gas with a single asset.\n- Developer Benefit: Simplify user onboarding and abstract away multi-chain gas management complexity.\n- User Benefit: Fund one wallet, interact with any supported chain. Eliminates the need to bridge native gas tokens for every new network.
1 Token
All Chains
-100%
Bridge Gas Tax
counter-argument
THE TRADEOFF
The Devil's Advocate: Complexity, Fragmentation, and New Risks
Execution layer innovations that simplify user experience systematically shift complexity and risk to the protocol and infrastructure layer.
Abstraction creates systemic risk. Account abstraction (ERC-4337) and intent-based architectures (UniswapX, CowSwap) delegate transaction construction to third-party bundlers and solvers. This centralizes execution logic into opaque, high-stakes relayers that become single points of failure and censorship.
Fragmentation is the new lock-in. Multi-chain smart accounts (via Safe{Core}) and cross-chain intents (via Across, LayerZero) promise unified UX but create vendor dependencies. Developers now manage a fragmented security model across multiple bridging protocols and signature schemes instead of a single chain's consensus.
The gasless fallacy. Paymasters (like Biconomy, Pimlico) subsidize fees to onboard users, but this introduces subsidy economics and MEV extraction vectors. Relayers front-run profitable user operations, creating a hidden tax that undermines the promised cost savings.
Evidence: The ERC-4337 EntryPoint contract has over $100M in staked ETH, creating a massive honeypot. A single bug in this singleton contract would compromise every AA wallet on Ethereum.
takeaways
EXECUTION LAYER UX
TL;DR for Busy Builders
The EVM is evolving beyond gas wars and opaque state. Here are the key execution changes that let you build, not beg.
01
Parallel Execution (Solana, Sui, Aptos, Monad)
The Problem: Sequential execution creates congestion and unpredictable gas spikes. The Solution: Process independent transactions simultaneously using optimistic concurrency control.
- Throughput: Scales with cores, enabling 10k-100k+ TPS.
- Predictability: Eliminates gas auctions for non-conflicting operations.
100x
Throughput
~0.1s
Finality
02
Account Abstraction (ERC-4337, Safe, Biconomy)
The Problem: Seed phrases and gas payments in native tokens are a UX dead-end. The Solution: Decouple transaction validation and payment logic via smart contract wallets.
- Sponsorship: Let apps pay gas fees or use stablecoins.
- Batch Operations: Single signature for multiple actions (-90% user ops).
-90%
User Ops
0 ETH
Gas Needed
03
Intent-Based Architectures (UniswapX, CowSwap, Across)
The Problem: Users specify low-level transactions (swap A for B on DEX C) and get front-run. The Solution: Users declare desired outcome; a solver network finds the optimal path.
- Better Prices: Solvers compete, capturing MEV for the user.
- Gasless: Users sign a message, not a transaction.
+5-20bps
Price Improvement
0 Slippage
Guaranteed
04
Modular DA & Blobs (Ethereum Dencun, Celestia, EigenDA)
The Problem: Storing all data on-chain is expensive, making L2s and high-throughput apps unviable. The Solution: Offload data availability to specialized layers, keeping only commitments on-chain.
- Cost: L2 tx fees drop from $0.50 to <$0.01.
- Scale: Enables truly high-frequency on-chain apps.
-99%
L2 Fees
~128 KB
Per Blob
05
State Expiry & History Pruning (Ethereum Verkle, Stateless Clients)
The Problem: Full nodes require ~1TB+ of state, centralizing infrastructure. The Solution: Make old state 'inactive' and require witnesses for access, drastically reducing hardware requirements.
- Node Requirements: Drop from TBs to ~100GB.
- Access: Historical data remains available via portals.
-90%
State Size
$500
Node Cost
06
Global Preconfirmations (Espresso, SUAVE, Shutter)
The Problem: Users have no guarantee their transaction will be included or won't be front-run. The Solution: A network of sequencers provides a signed, enforceable promise of future inclusion.
- Latency: Guarantee received in ~500ms.
- Fairness: Encrypted mempools prevent predatory MEV.
~500ms
Guarantee
0 Front-run
Protected
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