Zylos Logo
Zylos
2026-02-10

Blockchain Scalability in 2026: Layer 2 Solutions, Rollups, and the Path Forward

researchblockchainlayer-2rollupsscalabilityethereumzk-rollupsoptimistic-rollups

Executive Summary

Blockchain scalability has evolved from a theoretical challenge to a production-ready ecosystem in 2026. Layer 2 solutions, particularly rollups, have emerged as the dominant scaling paradigm, processing millions of transactions daily at a fraction of mainnet costs. The Ethereum Dencun upgrade (March 2024) with EIP-4844 proto-danksharding reduced L2 transaction costs by 90-95% through blob transactions. Major rollups like Arbitrum, Optimism, Base, zkSync, and Starknet now achieve 15,000-40,000 TPS with fees as low as $0.0001 per transaction. However, significant challenges remain: most rollups still rely on centralized sequencers, cross-chain bridges pose security risks (accounting for 40% of Web3 exploits), and full danksharding remains years away. The ecosystem is moving toward modular architectures, decentralized sequencers, and hybrid data availability solutions.

The Scalability Trilemma

Blockchain systems traditionally face a fundamental trade-off between three properties:

  1. Decentralization: No single entity controls the network
  2. Security: Resistance to attacks and censorship
  3. Scalability: High transaction throughput and low costs

Ethereum's base layer prioritizes decentralization and security at ~15-30 TPS, while Layer 2 solutions attempt to achieve scalability without compromising the other two properties.

Layer 2 Scaling Solutions: The Rollup Era

What Are Rollups?

Rollups execute transactions off-chain, bundle them into batches, and post compressed data to Ethereum Layer 1 for final settlement. They inherit Ethereum's security while achieving orders of magnitude higher throughput.

Two main types:

  1. Optimistic Rollups: Assume transactions are valid unless challenged (7-day fraud proof window)

    • Examples: Arbitrum, Optimism, Base
    • Throughput: 2,000-4,000 TPS
    • Average fees: ~0.1 Gwei ($0.001-0.01 per transaction)

  2. ZK Rollups: Use cryptographic validity proofs for immediate finality

    • Examples: zkSync, Starknet, Polygon zkEVM, Scroll
    • Throughput: 15,000+ TPS
    • Average fees: ~$0.0001 per transaction (though zkEVM: ~1.9 Gwei due to proof generation costs)

Performance Comparison 2026

RollupTypeTPSAvg FeeTVL ShareKey Strength
ArbitrumOptimistic~40,000~0.1 Gwei~50% of L2 DeFiMature ecosystem, Stage 1 rollup
OptimismOptimistic2,000-4,000~0.1 Gwei~27% (via Superchain)Modular OP Stack, ecosystem partnerships
BaseOptimistic2,000-4,000LowGrowingCoinbase backing, developer-friendly
zkSyncZK15,000+$0.0001GrowingFast finality, developer tools
StarknetZK15,000+$0.0001SpecializedStrong ZK tech, scalability potential
Polygon zkEVMZK~2,000~1.9 GweiSignificantFull EVM compatibility

Key market statistics:

  • Combined L2 user base: 6+ million active addresses (projected late 2026)
  • Institutional TVL on enterprise L2s: Forecast to exceed $50 billion
  • Base + Arbitrum: Represent 77% of L2 DeFi TVL
  • Activity consistently higher for optimistic rollups than ZK rollups (2022-2024): $186.4B vs $20.8B TVL bridged to Ethereum

Optimistic vs ZK Rollups: The Trade-offs

Verification Approach:

  • Optimistic: Assume validity, rely on fraud proofs and 7-day challenge period
  • ZK: Cryptographic validity proofs provide immediate finality

Finality:

  • Optimistic: 7 days for withdrawals to L1 (though some bridges offer fast exits)
  • ZK: Instant finality once proof is verified (ideal for gaming, trading)

Developer Compatibility:

  • Optimistic: Full EVM compatibility, easy migration from Ethereum
  • ZK: Improving but requires custom tooling; zkEVM initiatives (Scroll, zkSync, Polygon) working toward full compatibility

Computational Costs:

  • Optimistic: Lower computational overhead, higher gas for L1 batch submission
  • ZK: Higher proof generation costs, but more efficient L1 data posting

Current Adoption:

  • Optimistic: Higher TVL and activity due to earlier maturity and easier development
  • ZK: Growing rapidly as tooling improves and finality advantage becomes critical

Ethereum's Scaling Roadmap: Danksharding

Proto-Danksharding (EIP-4844) - Implemented March 2024

EIP-4844 introduced blob transactions, a new transaction type carrying large data chunks (blobs) that are:

  • Cheaper than calldata: Blobs cost significantly less because they're pruned after ~18 days (4096 epochs)
  • Not accessible to EVM: Blobs are for data availability only, not execution
  • KZG-committed: Enable data availability sampling without full downloads

Impact:

  • L2 transaction costs reduced by 90-95%
  • Daily L2 operating costs dropped to $135,100 (down from $1M+ pre-Dencun)
  • Transaction volumes surged from 500,000 (Feb 2024) to 2.1 million (Oct 2025)
  • Each block can include up to 6 blobs (target: 3, max: 6)

Blob Gas Market:

  • Separate fee market from regular gas
  • Dynamic pricing based on blob demand
  • Target: 3 blobs per block for predictable costs

Full Danksharding - Future (2027+)

Full danksharding will increase blobs per block from 6 to 64, enabling:

  • 100,000+ TPS across all rollups
  • Proportionally lower costs per transaction
  • True Ethereum-native sharding without compromising security

Prerequisites:

  1. PeerDAS (EIP-7594): Peer Data Availability Sampling, coming in Fusaka upgrade (Q4 2025)

    • Validators sample small data chunks from peers
    • Reduces full-node storage requirements by 80%
    • Enables safer blob scaling beyond 6 per block

  2. Distributed Validator Technology: Further decentralizes validation

  3. Upgraded Client Software: Handling increased data throughput

Timeline:

  • 2024: Proto-danksharding (EIP-4844) ✅
  • 2025: Pectra upgrade (validator improvements) ✅, Fusaka (PeerDAS) 🔄
  • 2026: Glamsterdam (gas efficiency, throughput)
  • 2027+: Full danksharding

Alternative Scaling Solutions

State Channels

Concept: Conduct most transactions off-chain between parties, use blockchain only as dispute arbiter.

Examples:

  • Lightning Network (Bitcoin): Near-instant, low-fee BTC transfers
  • Raiden Network (Ethereum): Off-chain ERC20 token transfers

Benefits:

  • Thousands of off-chain transactions with only open/close on-chain
  • Near real-time speed
  • Minimal fees

Use Cases:

  • Micropayments
  • Gaming (fast state updates)
  • IoT device-to-device payments

Limitations:

  • Requires locking funds in channels
  • Best for frequent interactions between same parties
  • Capital inefficiency for one-off transactions

Validium: Off-Chain Data Availability

Concept: Similar to ZK rollups but stores data off-chain, only posting validity proofs to L1.

Characteristics:

  • Higher throughput: Thousands of TPS, lower costs than rollups
  • Lower security guarantee: Data availability depends on trusted parties
  • Use cases: High-frequency trading, gaming where speed >> security

Example: StarkEx powers platforms like dYdX (derivatives exchange) and Sorare (NFT gaming)

Volition: Hybrid Data Availability

Concept: Users choose per-transaction whether data goes on-chain (rollup mode) or off-chain (validium mode).

Benefits:

  • Flexibility: High-value transactions use rollup mode; microtransactions use validium
  • Cost optimization: Users pay for security level they need
  • Seamless switching between modes

Example: StarkEx V4.5 introduced volition, allowing dynamic DA selection

Modular Blockchain Architecture

Concept: Separate consensus, data availability, execution, and settlement into specialized layers.

Key Components:

  • Execution Layer: Rollups (Arbitrum, Optimism, zkSync)
  • Data Availability Layer: Ethereum (via blobs), Celestia, EigenDA, Avail
  • Settlement Layer: Ethereum L1
  • Consensus Layer: Ethereum beacon chain

Benefits:

  • Horizontal Scalability: Run multiple parallel chains sharing security
  • Specialization: Each layer optimizes for specific function
  • Interoperability: Easier cross-chain communication within modular stack

Example: Validium networks can deploy like containers in modular stacks, allowing new ecosystems to join with reduced friction.

Critical Challenges in 2026

1. Sequencer Centralization

Problem: All major Ethereum L2 rollups currently rely on centralized sequencers.

Risks:

  • Censorship: Centralized operators can exclude transactions
  • MEV Extraction: Sequencers see all transactions and can extract undisclosed MEV
  • Single Point of Failure: Downtime if sequencer fails

Solutions in Development:

Decentralized Sequencers:

  • Multiple validators share transaction ordering responsibilities
  • Metis launched alpha decentralized sequencer (2024)
  • Arbitrum, Optimism working toward decentralization roadmaps

Shared Sequencers:

  • Espresso: Decentralized network providing sequencing for multiple rollups
  • Astria: Shared sequencer with permissionless participation
  • Benefits: Cross-rollup atomic transactions, MEV resistance, censorship resistance

Current Status (2026):

  • Most L2s still use centralized sequencers (treated as "acceptable centralization" for now)
  • Decentralization viewed as long-term goal, not immediate priority
  • Stage 1 rollups (like Arbitrum) have fraud proof systems reducing sequencer trust requirements

2. Cross-Chain Bridge Security

Problem: Bridges account for 40% of Web3 security incidents.

Key Risks:

  • Smart Contract Vulnerabilities: Bugs in bridge contracts exploited for hundreds of millions
  • Relayer Failures: Centralized relayers can fail or be compromised
  • Economic Security: Insufficient bonding may not deter attacks

Security Pillars:

  1. Economic Security: Bonded validators with skin in the game
  2. Implementation Security: Audited smart contracts, formal verification
  3. Environment Security: Monitoring, incident response, upgrade processes

2026 State:

  • Security improvements post-2022 exploits (e.g., Wormhole infrastructure strengthened)
  • Intent-based bridging emerging (Across Protocol uses intent-based execution)
  • Chain abstraction reducing need for manual bridging
  • Interoperability standards like ERC-7683 improving cross-chain reliability

Best Practices:

  • Use official bridges (e.g., Arbitrum Bridge for Arbitrum ↔ Ethereum)
  • Verify audits and security track records
  • Consider bridge age and TVL as trust signals
  • For large transfers, accept slower but more secure canonical bridges

3. ZK Rollup Centralization

Problem: ZK rollups depend on centralized proving circuits and prover infrastructure.

Risks:

  • Single prover or small set of controlled circuits
  • Most networks operate "Stage 0" rollups (missing decentralization features)

Progress:

  • zkSync experimenting with external/distributed prover networks
  • Prover decentralization roadmaps in development
  • Hardware acceleration (GPUs, FPGAs) making proving more accessible

4. Transaction Finality Trade-offs

Optimistic Rollups: 7-day withdrawal period for L1 finality

  • Fast bridges offer liquidity but introduce custodial risk
  • Users must trust bridge operators or wait full challenge period

ZK Rollups: Instant finality but higher computational costs

  • Proof generation can be slow for complex transactions
  • Centralized provers currently bottleneck

5. Interoperability Fragmentation

Problem: Dozens of L2s create fragmented liquidity and user experience.

Solutions:

  • Superchains: Optimism's OP Stack enables shared infrastructure (Base, World Chain, Soneium, INK, UniChain)
  • Cross-L2 Messaging: Native protocols for L2 ↔ L2 communication
  • Chain Abstraction: Users interact with apps without knowing underlying chain
  • Shared Sequencers: Atomic cross-rollup transactions via Espresso/Astria

Economics and Adoption

Transaction Cost Evolution

Pre-Dencun (before March 2024):

  • Ethereum L1: $5-50 per transaction (during congestion)
  • L2s: $0.10-1.00 per transaction

Post-Dencun (March 2024 - present):

  • Ethereum L1: $1-20 per transaction (improved base layer efficiency)
  • Optimistic Rollups: $0.001-0.01 per transaction (0.1 Gwei average)
  • ZK Rollups: $0.0001 per transaction (except zkEVM: higher proof costs)

Future (with Fusaka PeerDAS, late 2025):

  • Further 30-50% cost reduction expected
  • More predictable blob gas pricing
  • Higher blob capacity without centralization

Future (with Full Danksharding, 2027+):

  • 10x blob capacity (6 → 64 blobs per block)
  • Proportional cost reduction
  • Sub-cent transactions even for complex operations

Use Case Enablement

Now Viable on L2s:

  • DeFi: Decentralized exchanges, lending, yield farming (Base + Arbitrum: 77% of L2 DeFi TVL)
  • NFTs: Low-cost minting and trading
  • Gaming: Fast in-game transactions, play-to-earn
  • Social Apps: Blockchain-based social media with microtransactions
  • Payments: Real-world merchant payments with <$0.01 fees
  • IoT: Device-to-device microtransactions (Lightning, Raiden)

Institutional Adoption (2026):

  • Forecast: $50B+ institutional TVL on enterprise L2s by late 2026
  • Tokenized assets on ZK rollups (15,000+ TPS enables institutional-grade compliance)
  • Privacy-preserving enterprise transactions

The Road Ahead: 2026 and Beyond

Near-Term Milestones (2025-2026)

  1. Fusaka Upgrade (Q4 2025): PeerDAS implementation

    • 80% reduction in full-node storage requirements
    • Safer blob scaling foundation
    • Paves way for increasing blob count beyond 6

  2. Glamsterdam Upgrade (Early 2026): Execution layer improvements

    • Gas efficiency optimizations
    • Simplified transaction processing
    • Network throughput boost

  3. Sequencer Decentralization: Multiple rollups transitioning from centralized to decentralized/shared sequencers

    • Metis, Arbitrum, Optimism roadmaps progressing
    • Espresso and Astria shared sequencer networks maturing

  4. Rollup Interoperability: Native cross-L2 messaging and liquidity sharing

    • Optimism Superchain expanding (Base, World Chain, etc.)
    • Standards like ERC-7683 gaining adoption

  5. ZK Prover Decentralization: Distributed proving networks launching

    • zkSync external prover experiments
    • Hardware acceleration making proving accessible

Long-Term Vision (2027+)

  1. Full Danksharding: 64 blobs per block

    • 100,000+ TPS aggregate throughput across all rollups
    • Sub-cent transactions as default
    • Data availability solved at scale

  2. Stage 2 Rollups: Fully decentralized rollups with:

    • Decentralized sequencers
    • Permissionless provers
    • Governance-controlled upgrades only

  3. Seamless Multi-Chain UX:

    • Chain abstraction hides blockchain complexity
    • Users interact with apps, not chains
    • Assets flow automatically via intent-based protocols

  4. Enterprise-Grade Infrastructure:

    • Compliance-ready ZK rollups for regulated institutions
    • Privacy-preserving transactions at scale
    • Interoperable with traditional finance systems

  5. Sustainable Economics:

    • L2 revenue models mature (sequencer fees, MEV sharing)
    • Ethereum L1 benefits from L2 blob demand
    • Aligned incentives across modular stack

Conclusion

Blockchain scalability in 2026 is no longer a future promise but a present reality. Layer 2 rollups have successfully scaled Ethereum to millions of daily transactions at sub-cent costs while inheriting mainnet security. Proto-danksharding (EIP-4844) marked a turning point, slashing L2 costs by 90-95% and enabling mass adoption of DeFi, gaming, NFTs, and payments.

However, the journey is far from complete. Critical challenges remain: most rollups still rely on centralized sequencers, cross-chain bridges pose persistent security risks, and full danksharding remains years away. The ecosystem is actively addressing these through decentralized sequencers, shared infrastructure (Optimism Superchain), improved bridge security, and modular architectures that separate concerns.

The next phase focuses on decentralization without sacrificing performance, interoperability without fragmentation, and institutional adoption without compromising blockchain's core values. With Fusaka (PeerDAS) coming in late 2025, Glamsterdam in early 2026, and full danksharding on the horizon for 2027+, Ethereum's rollup-centric roadmap aims to deliver 100,000+ TPS while maintaining credible neutrality.

For developers, builders, and users, the message is clear: Layer 2 is production-ready today, and the infrastructure supporting it will only improve. The scalability trilemma hasn't been solved but has been cleverly sidestepped through modular design, cryptographic innovation, and pragmatic trade-offs. The question is no longer whether blockchains can scale, but how fast the ecosystem can mature to realize its full potential.

Sources