DACs are centralized bottlenecks. A committee of known entities signs off on data availability, replacing the cryptoeconomic security of L1s with a multi-sig trust model. This reintroduces the validator centralization problem that blockchains were built to solve.
comparison-of-consensus-mechanisms
Blog
Why Data Availability Committees Are a Centralized Compromise
An analysis of how Data Availability Committees (DACs) used by L2s like StarkEx and Polygon zkEVM create a trusted layer, sacrificing decentralization for short-term scalability gains.
introduction
THE DAC DILEMMA
The Scalability Trap: Trading Trust for Throughput
Data Availability Committees sacrifice decentralization to achieve high transaction throughput, creating a systemic risk vector.
The trade-off is explicit. Projects like Celestia's sovereign rollups and EigenDA offer pure data availability layers. DACs, used by early Polygon PoS and Boba Network, are a pragmatic shortcut that optimizes for speed over security.
Failure modes are catastrophic. If a DAC colludes or gets compromised, it can censor transactions or withhold data, permanently freezing associated L2s like Arbitrum Nova. This creates a single point of failure absent in pure rollup designs.
Evidence: The security budget shrinks. A 7-of-10 multi-sig DAC has 120 private key combinations. A decentralized network like Ethereum has the combined hashpower of millions of miners/validators, making collusion economically impossible.
key-trends
THE CENTRALIZATION TRADEOFF
The State of the Compromise: Where DACs Are Used
Data Availability Committees (DACs) are a pragmatic, centralized solution for scaling blockchains, offering a middle ground between expensive on-chain data and trust-minimized alternatives like danksharding.
01
Celestia's Rollup-Centric Vision
Celestia pioneered the modular stack, using a DAC as a stepping stone to full danksharding. It provides a high-throughput data availability layer for sovereign rollups, but its current security model relies on a permissioned set of validators.
- Key Benefit: Enables sovereign rollups with minimal bootstrapping friction.
- Key Benefit: ~$0.001 per KB data posting cost vs. Ethereum's ~$1+.
- The Compromise: ~100 validators form the initial committee, a centralized point of failure.
~100
Validators
$0.001/KB
Data Cost
02
EigenDA: The Restaking Security Play
Built on EigenLayer, EigenDA leverages restaked ETH to secure its data availability layer. It's a DAC where operators are economically slashed via Ethereum, creating a cryptoeconomic trust layer rather than a purely social one.
- Key Benefit: Taps into Ethereum's $15B+ restaking economy for security.
- Key Benefit: Targets 10-100 MB/s throughput for hyper-scaled L2s like Arbitrum Orbit.
- The Compromise: Security is correlated with EigenLayer's centralization risks and operator set.
$15B+
Restaked TVL
10-100 MB/s
Target Throughput
03
Avail & Polygon Avail: The Proof-of-Stake DACs
These projects position their DACs as Proof-of-Stake (PoS) networks in their own right, with hundreds of validators. They aim for a higher decentralization threshold than typical committees but fall short of the data availability sampling (DAS) guarantees of full danksharding.
- Key Benefit: ~100-200+ validators, improving censorship resistance.
- Key Benefit: Native interoperability for rollups within their ecosystem (e.g., Polygon CDK).
- The Compromise: Security is siloed to the chain's own token, not Ethereum or a robust restaking base.
100-200+
Validators
Polygon CDK
Native Stack
04
The Problem: Validium's Inevitable Choice
Validiums (e.g., StarkEx, zkPorter) must store data off-chain to achieve ultra-low fees and high TPS. Using a pure DAC is often the only viable business decision, as on-chain DA can constitute >90% of a user's transaction cost.
- Key Benefit: ~$0.01 per trade vs. L1's ~$10+.
- Key Benefit: 9,000+ TPS achievable with off-chain data.
- The Compromise: Users trade Ethereum-grade security for a known, auditable entity's promise (e.g., StarkWare, Matter Labs).
>90%
Cost Saved
9,000+
TPS
05
Near's Nightshade Sharding
Near implements sharding (Nightshade) where each shard's data is validated by the entire network, but in practice, chunk-only producers act as a DAC for specific data segments. This creates a scalable but complex trust model.
- Key Benefit: Horizontally scalable throughput across multiple shards.
- Key Benefit: Single-shard security failure does not compromise the entire chain.
- The Compromise: Relies on a rotating subset of validators per chunk, introducing localized trust assumptions.
Horizontally
Scalable
Sharded
Architecture
06
The Solution: Danksharding's Endgame
Ethereum's danksharding (EIP-4844, Proto-Danksharding) is the trust-minimized counterpoint to DACs. It uses data availability sampling (DAS) allowing light nodes to verify availability, removing the need for a trusted committee.
- Key Benefit: Preserves Ethereum's full security for L2 data.
- Key Benefit: Enables scalability without new trust assumptions.
- The Reality: Rollups will still use DACs for years until DAS is fully implemented and cost-competitive.
Ethereum
Security
DAS
Trust Model
deep-dive
THE CENTRALIZED COMPROMISE
Deconstructing the DAC: A First-Principles Breakdown
Data Availability Committees trade decentralization for scalability, creating a trust-based bottleneck that contradicts blockchain's core value proposition.
A DAC is a cartel. It replaces the permissionless data availability of a full L1 with a fixed, permissioned set of signers. This creates a single point of failure and trust, fundamentally centralizing the security of the rollup.
The trade-off is explicit. Projects like Arbitrum Nova and Polygon Avail (in its early design) use DACs to lower costs and increase throughput by orders of magnitude. The cost is accepting that liveness depends on a handful of known entities.
This violates crypto-first principles. The security budget of a DAC is social, not cryptographic. It relies on legal agreements and reputation, not economic staking and slashing like Ethereum or Celestia. A malicious majority can censor or withhold data.
Evidence: Real-world adoption. Despite the compromise, DACs are deployed because they work. Arbitrum Nova processes millions of transactions for gaming apps, proving the market prioritizes low fees over pure decentralization for specific use cases.
WHY DATA AVAILABILITY COMMITTEES ARE A CENTRALIZED COMPROMISE
DA Layer Comparison: Security vs. Cost Trade-Offs
A first-principles breakdown of Data Availability (DA) solutions, comparing the security guarantees and economic models of committees, validiums, and full rollups.
| Feature / Metric | Data Availability Committee (e.g., StarkEx, Polygon CDK) | Validium (e.g., StarkNet, zkSync) | Ethereum L1 (Rollup Mode) |
|---|---|---|---|
Data Custody Model | Multi-sig committee of 5-8 entities | Permissionless set of Proof-of-Stake validators | Entire Ethereum validator set (~1M ETH staked) |
Withdrawal Security | Committee can freeze funds | Validators can freeze funds | Cryptoeconomically secure; uncensorable |
Data Availability Cost | $0.01 - $0.10 per tx | $0.10 - $0.50 per tx | $1.00 - $5.00+ per tx |
Time to Finality | < 1 second | ~10 minutes (challenge period) | ~12 minutes (Ethereum block time) |
Data Redundancy | 3-8 copies held by committee | ~100 copies via DAS & committee | Thousands of copies via global node network |
Censorship Resistance | |||
Requires Trusted Third Parties | |||
Suitable for High-Value Assets (>$1B TVL) |
counter-argument
THE PRACTICAL TRADEOFF
The Steelman: Why Builders Choose DACs Anyway
DACs represent a deliberate, performance-first compromise for builders prioritizing speed and cost over pure decentralization.
Immediate Time-to-Market is the primary driver. Launching a fully decentralized rollup with a battle-tested Data Availability (DA) layer like Celestia or EigenDA requires complex integration and consensus engineering. A Data Availability Committee (DAC) is a pre-packaged, plug-and-play solution that shaves months off development.
Cost Predictability beats theoretical security for early-stage projects. DACs like those used by Arbitrum Nova or Mantle offer fixed, low fees. The variable cost of posting data to Ethereum or even an alternative L1 introduces financial risk a fledgling dApp cannot absorb.
The Security Spectrum is misunderstood. A 7-of-10 multisig DAC is not the same as a single operator. It provides a practical sybil-resistance threshold that deters casual attacks, creating a security model superior to many centralized exchanges but inferior to Ethereum.
Evidence: Arbitrum Nova, secured by a DAC including Google Cloud and ConsenSys, processes over 50% of Arbitrum's total transaction volume. This proves builders and users vote with their gas fees for this specific tradeoff.
risk-analysis
THE CENTRALIZATION TRAP
The Slippery Slope: Systemic Risks of DAC Dependence
Data Availability Committees (DACs) trade decentralization for scalability, creating hidden systemic risks for the L2s and rollups that rely on them.
01
The Cartel Problem
A DAC is a permissioned set of entities, often the L2's core team and VCs. This creates a single point of failure and regulatory capture.\n- Governance Capture: A majority can censor or halt the chain.\n- Regulatory Target: A defined group is legally liable, unlike a decentralized network.\n- Collusion Risk: Members can coordinate to extract MEV or manipulate state.
5-10
Typical Members
51%
Attack Threshold
02
The Data Blackout
If a DAC fails to sign or publish data, the rollup grinds to a halt. Users cannot prove fraud or withdraw funds, creating a fundamental liveness failure.\n- Withdrawal Freeze: No data means no Merkle proofs for the L1 bridge.\n- Chain Death Spiral: Sequencers stop producing blocks, killing all dApps.\n- Contagion Risk: A major L2 outage triggers panics across DeFi (e.g., MakerDAO, Aave).
~0s
Recovery Time
100%
TVL Frozen
03
Celestia vs. EigenDA: The False Dichotomy
The market frames this as a choice, but both models centralize risk. Celestia's validator set is permissionless but small (~$2B staked). EigenDA relies on EigenLayer's ~$20B in restaked ETH, creating massive re-staking systemic risk. The real solution is Ethereum's danksharding (EIP-4844) or validity-proof-based DA like Avail.
~$20B
EigenDA TVL Risk
100-200
Active Validators
04
The Regulatory Kill Switch
A DAC provides a clear legal entity for regulators to target with sanctions or shutdown orders. This undermines crypto's core value proposition of censorship resistance.\n- OFAC Compliance: Members must censor transactions or face penalties.\n- Securities Law: A controlled group makes the L2 look like an unregistered security.\n- Precedent: This is how traditional finance is controlled.
1
Subpoena Needed
100%
Enforceable
future-outlook
THE CENTRALIZATION TRAP
The Path Forward: Beyond the Committee
Data Availability Committees are a temporary scaling solution that reintroduces trust assumptions, creating a centralization bottleneck for the next generation of rollups.
Committees are centralized bottlenecks. They replace a decentralized network of nodes with a small, permissioned set of signers. This creates a single point of failure and regulatory capture, directly contradicting the trust-minimization ethos of Ethereum.
The trust model regresses. A DAC-based rollup like a Celestia-powered chain or an early-stage Arbitrum Nova relies on committee honesty for data availability. This is a step backward from Ethereum's cryptoeconomic security, which requires no such social consensus.
The industry is moving to pure DA. Protocols like EigenDA and Avail provide cryptoeconomically secured data availability without committees. This shift mirrors the evolution from multi-sig bridges to trust-minimized solutions like Across and Chainlink CCIP.
Evidence: The total value secured by DAC-based systems is a fraction of Ethereum's. For long-term viability, rollups must graduate to enshrined rollups or Ethereum-aligned DA to inherit L1's security guarantees.
takeaways
THE DAC DILEMMA
TL;DR for CTOs and Architects
Data Availability Committees offer a pragmatic scaling path, but their security model is a centralized compromise that introduces systemic risk.
01
The Problem: Data Availability is the Scalability Bottleneck
Full on-chain data storage for rollups like Arbitrum and Optimism is expensive. Posting all transaction data to Ethereum L1 creates a ~90% cost burden for users. This forces a trade-off: security for scalability.
~90%
Cost Burden
1.5MB/s
Ethereum Limit
02
The Solution: A Trusted Quorum
DACs like those used by StarkEx and Polygon Avail's "validium" mode replace cryptographic guarantees with a legal/multisig model. A known set of entities (e.g., 5-10 reputable nodes) sign off that data is available, allowing processing without full L1 publication.
- Key Benefit 1: Enables ~10,000+ TPS and near-zero fees.
- Key Benefit 2: Pragmatic for high-throughput, low-value applications (gaming, perps).
10,000+
TPS
5-10
Trusted Nodes
03
The Centralized Compromise: Liveness over Safety
DACs prioritize liveness (keeping the chain running) over Ethereum's safety guarantees. If the committee colludes or goes offline, funds can be frozen or stolen. This reintroduces the custodial risk that decentralization aimed to solve.
- Key Risk 1: Single point of failure via committee multisig.
- Key Risk 2: No cryptographic proof of data withholding.
1
Multisig Failure
0
Cryptographic Proof
04
The Alternative: Modular DA Layers
Projects like Celestia, EigenDA, and Avail (in full mode) provide decentralized data availability with cryptographic guarantees. They use Data Availability Sampling (DAS) and erasure coding, allowing light nodes to verify data is published without downloading it all.
- Key Benefit 1: Trust-minimized security for sovereign rollups.
- Key Benefit 2: ~100x cheaper than Ethereum calldata, but more secure than a DAC.
100x
Cheaper vs ETH
1000+
Light Nodes
05
The Architect's Choice: Risk-Weighted Design
Your application's value-at-risk dictates the DA solution. Use a risk matrix: High-value DeFi (Uniswap) needs Ethereum or Celestia. High-volume gaming/NFTs can tolerate a reputable DAC. Hybrid models like "volition" (zkSync) let users choose per transaction.
High-Value
DeFi = L1/Celestia
High-Volume
Apps = DAC
06
The Future: Proofs Kill Committees
The endgame is cryptographic, not social. Ethereum's EIP-4844 (blobs) reduces L1 DA cost. zk-proofs of DA (like from Avail) and succinct fraud proofs (Optimism's Cannon) will make trusted committees obsolete for all but the most latency-sensitive use cases.
EIP-4844
Cost -90%
zk-proofs
Future Standard
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