The trilemma is solved. The modular thesis of separating execution, consensus, and data availability (DA) has resolved the classic blockchain trilemma. Rollups like Arbitrum and Optimism now handle execution, while Ethereum L1 or Celestia provide consensus and DA. The bottleneck has simply shifted.
zk-rollups-the-endgame-for-scaling
Blog
Why the 'Scalability Trilemma' Is Now a 'Rollup Dilemma'
The old blockchain trilemma is solved. The new constraint is at the rollup layer: you must choose between high throughput, decentralized provers, and low user costs. This is the defining architectural battle for L2s.
introduction
THE ROLLUP DILEMMA
The Scaling Endgame Has a New Bottleneck
The scalability trilemma has collapsed into a single, more acute problem: rollups are now bottlenecked by their data availability layer.
The new bottleneck is data. Every rollup transaction requires publishing data to a DA layer. The cost and throughput of this data availability layer now defines the entire system's scalability and cost. This is the rollup dilemma.
Ethereum's DA is expensive. Using Ethereum's calldata for DA, as Arbitrum and Base do, anchors security but caps throughput and inflates transaction fees. EIP-4844 (blobs) provides temporary relief but does not solve the fundamental capacity constraint.
Alternative DA creates fragmentation. Using a cheaper, external DA layer like Celestia or Avail reduces costs but introduces a new security assumption and bridging complexity. This fragments liquidity and composability across the modular stack.
Evidence: The cost to post 1 MB of data to Ethereum is ~0.1 ETH, while on Celestia it is ~$0.01. This 1000x cost differential forces rollups to choose between security and scalability, creating the core trade-off of the rollup era.
thesis-statement
THE ROLLUP DILEMMA
The Core Argument: A Three-Body Problem for L2s
The scalability trilemma has collapsed into a zero-sum game between decentralization, security, and performance for Layer 2 rollups.
The trilemma is now local. Ethereum's base layer solved this globally, but each L2 now faces its own internal version. Optimizing for one vertex forces trade-offs in the other two, creating a zero-sum game for rollup architects.
Decentralization is the first sacrifice. High-throughput chains like Solana prove performance requires centralization. To compete, L2s like Arbitrum and Optimism centralize sequencers and provers, creating a single point of failure and censorship risk.
Security is the second casualty. Validiums and Optimiums (e.g., some StarkEx chains) trade Ethereum's data availability for lower fees. This reintroduces the very custodial risk that rollups were designed to eliminate.
Evidence: The L2Beat dashboard shows only 5 of the top 10 L2s by TVL are fully Ethereum-secured rollups. The rest are validiums or have centralized sequencers, proving the dilemma is already being resolved by compromising core values.
key-trends
FROM TRILEMMA TO TRADEOFF
The Three Axes of the Rollup Dilemma
The original blockchain trilemma (Security, Scalability, Decentralization) has been outsourced. Rollups now face their own three-way tradeoff: Sovereignty, Security, and Shared Liquidity.
01
The Sovereignty Trap: Appchains vs. Shared Sequencing
Rollups must choose between full control (own chain, custom VM) and shared efficiency (shared sequencer like Espresso or Astria). Sovereignty enables maximal extractable value (MEV) capture and tailored execution but fragments liquidity and security.
- Key Tradeoff: Customizability vs. Network Effects
- Example: dYdX V4 moved to its own Cosmos chain for sovereignty, sacrificing Ethereum's native composability.
100%
Control
-90%
Composability
02
The Security Subsidy: Opt-in vs. Inherited
Security isn't free. Rollups can pay for it directly via their own validator set (expensive, slow) or rent it from a parent chain (Ethereum, Celestia) via fraud/validity proofs. The choice defines your threat model and capital cost.
- Key Tradeoff: Capital Cost vs. Trust Assumptions
- Example: Arbitrum and Optimism pay ~$1M+/month in L1 data fees for Ethereum's security, while many alt-L1s bootstrap their own less-proven validator sets.
$1M+
Monthly Cost
7 Days
Challenge Window
03
The Liquidity Fault Line: Native vs. Bridged
Where does value live? Native assets on the rollup's own chain are secure but illiquid. Bridged assets from Ethereum (via LayerZero, Across) are liquid but introduce bridge risk. This dictates DeFi viability and user onboarding friction.
- Key Tradeoff: Capital Efficiency vs. Counterparty Risk
- Example: A rollup with only bridged USDC is hostage to the security of Stargate or Circle's CCTP, creating a weak point outside its own crypto-economic security.
$10B+
Bridged TVL Risk
~20 mins
Withdrawal Delay
SCALABILITY TRILEMMA EVOLVED
The Rollup Dilemma in Practice: Protocol Trade-Offs
Comparing core architectural choices for rollups, highlighting the inherent trade-offs between decentralization, security, and performance.
| Architectural Dimension | Optimistic Rollups (e.g., Arbitrum, Optimism) | ZK-Rollups (e.g., zkSync, StarkNet) | Validiums (e.g., Immutable X, Sorare) |
|---|---|---|---|
Data Availability | On-chain (Ethereum) | On-chain (Ethereum) | Off-chain (DAC/Committee) |
Withdrawal Finality (Time to L1) | 7 days (Challenge Period) | < 1 hour (ZK Proof Verification) | < 1 hour (Proof + Data Signatures) |
Inherent Trust Assumption | 1 honest validator in 7 days | Cryptographic (ZK Proof Validity) | Data Availability Committee Honesty |
EVM Compatibility | Full bytecode equivalence | Custom ZK-EVM (bytecode-level emerging) | Application-specific (often custom VM) |
Transaction Throughput (Max TPS) | ~4,000-5,000 (theoretical) | ~2,000-3,000 (current, proof-gen bound) | ~9,000+ (off-chain data unconstrained) |
L1 Security Cost (per tx data) | ~$0.10 - $0.30 (calldata) | ~$0.15 - $0.40 (calldata + proof) | < $0.01 (signatures only) |
Prover Centralization Risk | Low (Sequencer optional, no special hardware) | High (Specialized proving hardware required) | Medium (Committee operation & proving) |
deep-dive
THE NEW REALITY
Deconstructing the Dilemma: Why You Can't Have All Three
The classic blockchain trilemma has collapsed into a rollup-centric dilemma, forcing a choice between decentralization, sovereignty, and shared security.
The trilemma is now a dilemma. Ethereum's L1 prioritizes decentralization and security, pushing scalability to rollups. This creates a new trade-off: rollups must choose between sovereign execution and shared security.
Sovereignty demands sacrifice. An Arbitrum Nova or Starknet appchain gains full control over its stack and forkability. This sovereignty comes at the cost of fragmented liquidity and a weaker security budget than Ethereum L1.
Shared security limits control. Using a shared sequencer like Espresso or a validium with Ethereum DA (e.g., via EigenDA) inherits L1 security. This forces protocol developers to cede operational control and upgrade keys to a centralized entity.
Evidence: The market vote is clear. Over 90% of rollup TVL resides on Arbitrum and Optimism, which chose shared security and sacrificed full sovereignty for network effects. Sovereign chains like dYdX remain isolated islands.
protocol-spotlight
THE ROLLUP DILEMMA
Architectural Responses to the Dilemma
The scalability trilemma has collapsed into a rollup-specific dilemma: how to scale execution without fragmenting liquidity or compromising security. These are the core architectural responses.
01
The Sovereign Rollup Gambit
Celestia, EigenLayer, and Avail champion a radical split: decouple execution from consensus and data availability. The base layer provides cheap, secure data, while rollups handle execution and define their own fork-choice rules.
- Key Benefit: Unlocks unprecedented sovereignty and customization for app-chains.
- Key Benefit: Scales data availability independently, enabling ~$0.001 per transaction data costs.
100x
Cheaper DA
Sovereign
Fork Choice
02
The Shared Sequencer Collective
Espresso Systems, Astria, and shared sequencer networks like those proposed for Arbitrum and Optimism solve the atomic composability and MEV fragmentation problem. They provide a neutral, decentralized sequencing layer for multiple rollups.
- Key Benefit: Enables atomic cross-rollup transactions without slow bridges.
- Key Benefit: Mitigates MEV extraction and provides credible neutrality for rollup operators.
~500ms
Cross-Rollup
Neutral
MEV Market
03
The Intent-Centric Settlement Layer
Proposed by Anoma and embodied in protocols like UniswapX and CowSwap, this inverts the transaction model. Users declare a desired outcome (intent), and a solver network competes to fulfill it optimally across fragmented liquidity pools.
- Key Benefit: Abstracts away liquidity fragmentation, acting as a native cross-rollup AMM.
- Key Benefit: Users get better execution via solver competition, often surpassing simple limit orders.
Optimal
Execution
Abstracted
Liquidity
04
The Hyper-Integrated L2 Stack
Arbitrum Stylus and Optimism's OP Stack represent the integrated counterpoint to sovereign rollups. They deeply couple execution environments with their parent chain's security and tooling, prioritizing developer experience and network effects.
- Key Benefit: Seamless EVM+ compatibility allows developers to use Rust, C++, and other languages.
- Key Benefit: Leverages existing $10B+ TVL and security assumptions of Ethereum, reducing adoption friction.
EVM+
Compatibility
$10B+
Shared TVL
05
The Universal Interop Hub
LayerZero, Chainlink CCIP, and Axelar are building canonical messaging layers that treat every chain and rollup as a sovereign state. They provide the secure communication rails, forcing the dilemma to be solved at the application layer via omnichain contracts.
- Key Benefit: Creates a standardized security primitive for cross-chain messaging, audited once.
- Key Benefit: Enables true omnichain applications where state is synchronized across hundreds of environments.
Canonical
Messaging
Omnichain
Apps
06
The Validium & Volition Compromise
StarkEx, zkSync, and Aztec offer a spectrum of data availability choices. Validiums trade some Ethereum security for lower costs by using off-chain DA, while Volitions let users choose per transaction. This directly addresses the cost/security trade-off.
- Key Benefit: ~10-100x lower fees than full zkRollups by moving data off-chain.
- Key Benefit: User-controlled security sliders (Volitions) provide flexibility for different asset classes.
~100x
Cheaper Tx
Adjustable
Security
future-outlook
THE ROLLUP DILEMMA
The Path Forward: Specialization and Hybrid Models
The scalability trilemma has collapsed into a rollup-specific tradeoff between sovereignty, security, and interoperability.
Sovereignty creates fragmentation. A sovereign rollup like Celestia or Avail provides maximal autonomy but inherits the hardest interoperability problem. This forces a choice between slow, trust-minimized bridges and fast, trust-required bridges like LayerZero or Wormhole.
Security demands standardization. A rollup secured by Ethereum's base layer, like Arbitrum or Optimism, trades sovereignty for native composability. The shared settlement layer becomes the interoperability hub, but innovation in data availability or execution is constrained.
Hybrid models are inevitable. The future is not a single winner. Projects like Eclipse and Polygon CDK demonstrate the hybrid approach: a rollup can use Celestia for data and Ethereum for settlement, optimizing for specific cost and security profiles.
Evidence: The market is voting for specialization. Over 60% of new rollups now use modular stacks, with Celestia capturing 30% of DA market share within six months of mainnet. This proves developers prioritize configurable trade-offs over monolithic dogma.
takeaways
THE ROLLUP DILEMMA
TL;DR for Builders and Investors
The scalability trilemma is dead. The new fight is over who controls the value and security of the modular stack.
01
The Problem: Shared Sequencers = Shared Sovereignty
Outsourcing block production to a neutral third party like Espresso or Astria trades sovereignty for interoperability. You get atomic cross-rollup composability but cede your chain's most valuable asset: its block space and MEV. This creates a new centralization vector and a political dilemma for rollup teams.
~500ms
Finality Target
1 Entity
Single Point
02
The Solution: App-Specific Sequencing
Rollups like dYdX and Aevo run their own sequencers. This preserves maximum value capture and customizability (e.g., front-running protection). The trade-off is severe: you become an isolated liquidity island, breaking native composability with the broader Ethereum L2 ecosystem and forcing users through slow, insecure bridges.
100%
MEV Capture
-90%
Interop Speed
03
The Hybrid: Based Sequencing & EigenLayer
Based rollups (e.g., using Espresso) let Ethereum L1 proposers sequence blocks, inheriting Ethereum's economic security and credibly neutralizing the sequencer centralization debate. EigenLayer AVSs offer an alternative, creating a marketplace for decentralized sequencer sets. Both models sacrifice some customizability for credible neutrality and shared security.
L1 Security
Backing
Market Rate
Cost
04
The Investor Lens: Value Flows to the Base Layer
The modular thesis shifts value accrual down the stack. Ethereum (via based sequencing and blob fees), Celestia/EigenDA (via data availability fees), and shared sequencer networks become the fundamental cash-flow assets. Rollups become low-margin, high-volume commodities. Investment shifts from 'which L2?' to 'which modular primitives win?'
$10B+
DA Market
Commodity
Rollup Margin
05
The Builder's Choice: Optimism's Superchain vs. Solo
OP Stack's Superchain (Optimism, Base, Zora) represents a coordinated vision: shared security (via fault proofs), a shared sequencer (eventually), and a unified liquidity pool via the Optimism Gateway. This is a direct counter to the fragmented Cosmos model. The choice is binary: join a coalition for network effects or go solo for maximum short-term extractable value.
1-Click
Chain Deploy
Shared TVL
Liquidity
06
The Endgame: Intents & Solver Networks
The ultimate bypass to the sequencer dilemma. Protocols like UniswapX, CowSwap, and Across use intents—users declare what they want, competitive solver networks figure out how. Execution moves off-chain, making the underlying rollup's sequencer choice less relevant. This abstracts the chain away, turning the rollup into a settlement commodity.
~1s
User Experience
Auction
Execution
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