The execution layer is the product. Rollups like Arbitrum and Optimism are not just scaling solutions; they are competing execution environments with distinct VMs, fee markets, and proving systems. The base layer's role is reduced to consensus and data availability, a shift formalized by Ethereum's Danksharding roadmap.
the-modular-blockchain-thesis-explained
Blog
Why the 'Rollup-Centric' Future is Really an 'Execution-Centric' One
Scaling was just the entry fee. The modular thesis unlocks a Cambrian explosion of execution environments—from parallel EVMs to app-specific VMs—by decoupling execution from base layer consensus.
introduction
THE SHIFT
Introduction
The industry's 'rollup-centric' roadmap is a misnomer; the real architectural battle is for execution supremacy.
Modularity commoditizes everything but execution. With shared DA from Celestia or EigenLayer and universal settlement on Ethereum, the only durable competitive edge is superior execution performance. This is why StarkWare's Cairo VM and Fuel's parallel UTXO model are existential bets, not incremental features.
Evidence: Arbitrum processes over 1 million transactions daily, but its Nitro stack and fraud proof mechanism define the user experience, not the underlying L1. The value accrual follows the execution, not the chain label.
key-trends
BEYOND THE ROLLUP NARRATIVE
The Execution Layer Explosion: Three Trends
The modular thesis is not about rollups; it's about commoditizing execution to unlock new application paradigms.
01
The Problem: The L2 Bottleneck is Execution, Not Data
Rollups like Arbitrum and Optimism are data-availability constrained, but their real scaling limit is single-threaded EVM execution. The bottleneck has shifted from the base layer to the rollup's own virtual machine.
- Sequencer congestion creates volatile fees and latency.
- Synchronous composability forces all apps into a single, slow queue.
- Monolithic VM design cannot scale with specialized hardware.
~500ms
Block Time
1x
Thread
02
The Solution: Parallel EVMs & Hyperchains
Projects like Monad, Sei, and Aptos are building parallel execution engines, while zkSync's Hyperchains and Arbitrum Orbit enable purpose-built execution layers. This fragments the monolithic block space.
- Horizontal scaling via parallel transaction processing.
- Vertical scaling via custom VMs for gaming or DeFi.
- Deterministic performance by isolating app-specific state.
10k+
TPS Target
10x
Efficiency Gain
03
The New Primitive: Intents & Solver Networks
Architectures like UniswapX, CowSwap, and Across abstract execution away from users. They shift the burden to off-chain solver networks competing on price and speed, making the chain a settlement layer.
- User expresses intent ("swap X for Y"), not a transaction.
- Solvers compete in a MEV-aware auction for best execution.
- Final settlement is batched and proven on-chain.
$10B+
Volume Routed
-90%
Slippage
deep-dive
THE EXECUTION LAYER
From Monolithic Prison to Modular Playground
The modular thesis shifts the bottleneck and value accrual from consensus to specialized execution environments.
The modular thesis is execution-centric. It decouples execution from consensus and data availability, enabling specialized environments like ZK-rollups, optimistic rollups, and validiums. This creates a competitive market for execution, where performance and cost are the primary metrics.
Monolithic chains are consensus prisons. They force all applications to share a single, slow execution environment. The Ethereum L1 is a settlement layer, not a scalable execution engine. Its role is to provide security and finality for rollups.
Execution layers compete on proving. ZK-rollups (Starknet, zkSync) and optimistic rollups (Arbitrum, Optimism) are the dominant models. Their competition drives innovation in prover efficiency, state management, and interoperability.
Evidence: Arbitrum processes over 1 million transactions daily, while Ethereum L1 handles ~1.2 million. The execution layer is where 90% of user activity occurs. The value accrual follows the activity.
THE END OF THE MONOLITH
Execution Layer Specialization Matrix
Comparing execution environments by their core architectural trade-offs and specialization. The 'rollup-centric' future is a misnomer; the real competition is between different execution paradigms.
| Core Execution Paradigm | Monolithic L1 (e.g., Solana, Sui) | General-Purpose ZK Rollup (e.g., zkSync Era, Starknet) | Application-Specific Rollup (e.g., dYdX, Aevo) | Parallelized VM / SVM Rollup (e.g., Eclipse, Monad) |
|---|---|---|---|---|
Architectural Goal | Maximize hardware utilization for all tasks | EVM equivalence & developer familiarity | Optimize for a single application's logic | Maximize throughput via parallel execution |
State Access Pattern | Global, contentious | Global, but can be sharded | Localized & predictable | Partitioned via software (e.g., Solana-style) |
Throughput (TPS) Theoretical | 50,000+ | 200-2,000 | 10,000+ (for its specific app) | 10,000+ |
Time to Finality | < 1 sec | ~10 min (L1) + ~10 sec (ZK proof) | < 1 min (with fast DA) | < 1 sec |
Sovereignty / Forkability | ||||
Custom Fee Token / MEV Capture | ||||
Primary Cost Driver | Global state bandwidth | L1 Data Availability & proof generation | Data Availability | Compute & State Access |
Key Innovation | Horizontal scaling via validator hardware | Zero-knowledge proofs for trustless bridging | Vertical integration & tailored economics | Deterministic parallel execution & async I/O |
counter-argument
THE EXECUTION-CENTRIC FUTURE
The Monolithic Counter-Punch: Solana's Integrated Model
The 'rollup-centric' vision is a misnomer; the real competition is between specialized execution layers, and Solana's monolithic architecture is its optimized entry.
The core competition is execution. The 'rollup-centric' narrative focuses on data availability and settlement, but user experience is defined by execution speed and cost. Solana's monolithic design vertically integrates these functions on a single state machine.
Integration eliminates coordination overhead. Rollups like Arbitrum and Optimism must pay for L1 security and bridge latency. Solana's single global state provides atomic composability without the fragmentation and bridging costs inherent to a multi-chain Ethereum L2 ecosystem.
Hardware scaling is the real bottleneck. The debate isn't data vs. settlement; it's whether specialized hardware can scale a single state machine. Solana's parallel execution via Sealevel and localized fee markets demonstrate this path, contrasting with Ethereum's planned enshrined rollups.
Evidence: User and developer traction. Solana consistently processes more daily transactions than all major Ethereum L2s combined, with Jupiter and Raydium enabling complex, atomic swaps impossible across fragmented rollups without LayerZero or Wormhole bridges.
takeaways
EXECUTION IS THE BOTTLENECK
TL;DR for Builders and Investors
The 'rollup-centric' roadmap is a misnomer; the real competition is for execution supremacy, with rollups as just one implementation.
01
The Problem: Rollups are a Feature, Not a Destination
Rollups are a specific scaling solution, not the end-state. The market demands generalized execution environments that can be any L1, L2, or L3. The winning architecture will be the one that provides the most performant, flexible, and cost-effective execution layer, regardless of its label.
- Modularity Wins: Decoupling execution from consensus/settlement/data availability is the trend.
- Execution is the Product: Users and developers care about finality speed and cost, not the underlying data-availability layer.
1000+
EVM Chains
-90%
Cost vs. L1
02
The Solution: Sovereign Execution Layers (Monad, Fuel, Eclipse)
New architectures are bypassing the EVM's legacy constraints to build hyper-optimized, parallel execution engines. These can exist as standalone L1s or be integrated as L2s/L3s via shared security.
- Parallel Execution: Enables 10,000+ TPS by processing non-conflicting transactions simultaneously.
- State Minimization: Reduces node hardware requirements, lowering barriers to decentralization.
10,000+
Theoretical TPS
~1s
Finality Target
03
The Investment Thesis: Bet on Execution Clients, Not Consensus
The value accrual is shifting from the base layer's native token (e.g., ETH for security) to the software that runs the execution environment. This mirrors the Prysm vs. Lighthouse dynamic in Ethereum consensus, but for execution.
- Client Diversity is Critical: Avoids single points of failure and fosters innovation.
- New Business Models: Execution-layer tokens could capture fees from MEV, sequencing, and premium throughput.
$1B+
MEV Annually
5-10x
Client Multiplier
04
The Builder's Playbook: Abstract the Chain
Developers should build for portable execution environments, not specific chains. Use intent-based architectures (like UniswapX) and universal smart accounts to let users' intents be fulfilled across any optimal execution layer.
- Intent-Centric Flow: Users specify what they want, not how to achieve it, enabling cross-chain atomicity.
- Execution Marketplace: Solvers (like in CowSwap, Across) compete to provide the best route and price.
50%+
Gas Saved
~500ms
Solver Latency
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