Commoditization of execution is the endgame. Rollups like Arbitrum and Optimism compete on UX, but their core function—processing transactions—converges into a low-margin utility. The real economic moat shifts to the underlying data layer that secures all transactions.
venture-capital-trends-in-web3
Blog
Why Data Availability Layers Are the Quiet Winners
Rollups get the hype, but the underlying Data Availability (DA) layer is the non-negotiable, fee-generating bottleneck. This analysis breaks down the inelastic demand, economic moats, and why VCs are quietly backing the plumbing.
introduction
THE DATA
Introduction: The Hidden Toll Bridge
Data Availability (DA) layers are capturing the fundamental value of scaling by commoditizing execution.
DA is the ultimate toll bridge. Every L2 transaction must post its data somewhere, paying a recurring fee. This creates a predictable revenue stream akin to AWS for web2, but for state verification. Celestia and EigenDA monetize this directly; Ethereum does so via blob fees.
The evidence is in the blobs. Ethereum's Dencun upgrade, which introduced EIP-4844 (proto-danksharding), cut L2 fees by over 90% overnight. This didn't kill DA value; it redirected it from calldata to a dedicated market, proving demand is inelastic. The toll is smaller per user, but the bridge handles exponentially more traffic.
thesis-statement
THE DATA PIPELINE
Core Thesis: Inelastic Demand Meets Modular Supply
Data Availability (DA) layers capture outsized value in a modular stack because their demand is inelastic while their supply is commoditized.
Inelastic demand is structural. Every transaction, rollup, and L2 must post data somewhere. This creates a non-negotiable cost floor for all blockchain activity, independent of market cycles.
Supply is becoming modular. With Celestia, EigenDA, and Avail, DA is now a pluggable resource. This commoditization drives down prices but consolidates demand onto the most efficient providers.
The winner captures the base fee. The dominant DA layer becomes the settlement layer for state. It extracts rent from every modular chain, similar to how AWS profits from all web2 apps.
Evidence: Ethereum as a DA layer for rollups generated over $400M in fees in 2023. Dedicated layers like Celestia reduce this cost by 99%, but the aggregate fee pool for the winning provider will scale with total blockchain usage.
market-context
THE BOTTLENECK
The DA Wars: Ethereum vs. The Alternatives
Data availability is the primary scaling bottleneck, and the battle to solve it defines the next generation of blockchain architecture.
Ethereum's DA is expensive. Every full node must download and verify all transaction data, creating a hard throughput cap. This cost is the primary driver of high L2 transaction fees, even for optimistic rollups like Arbitrum and Optimism.
Celestia pioneered modular DA. It separates execution from consensus and data availability, creating a sovereign rollup model. This allows chains like Manta and Dymension to post data at a fraction of Ethereum's cost, trading security for scalability.
EigenDA is Ethereum's response. It uses Ethereum's validator set for cryptoeconomic security but batches data off-chain, offering a middle ground. This is the core infrastructure for EigenLayer's restaking ecosystem and rollups like Mantle.
The trade-off is security vs. cost. Ethereum DA provides the highest security via full consensus. Celestia and Avail offer cheaper, probabilistic security. EigenDA sits in between, leveraging Ethereum's trust network without its full-node overhead.
key-trends
WHY DATA AVAILABILITY LAYERS ARE THE QUIET WINNERS
Three Trends Driving DA Dominance
The modular stack is here. Execution layers compete on speed, but the real infrastructure battle is for the data layer that secures them all.
01
The Modular Stack's Non-Negotiable Foundation
Rollups need to post data somewhere cheap and secure. On-chain Ethereum is too expensive, forcing a new market. DA layers like Celestia, EigenDA, and Avail provide a dedicated, optimized substrate for data publishing, decoupling security from execution costs.
- Cost is the Killer App: DA costs are ~99% cheaper than Ethereum calldata, the primary expense for L2s.
- Security as a Service: Provides crypto-economic security for data, enabling sovereign rollups to have their own execution fork choice.
- Interoperability Primitive: Shared DA layers enable native bridging and light client verification across rollups.
~99%
Cheaper vs ETH
10k+
TPS Capacity
02
The Blob-Centric Future Post-Dencun
Ethereum's Dencun upgrade introduced EIP-4844 (Proto-Danksharding), creating a dedicated data market with blobs. This validated the DA layer thesis but created a two-tier market: premium (Ethereum) and discount (external DA).
- Blobs as a Benchmark: Ethereum blobs set the gold standard for security, but supply is limited (~3-6 per block).
- The Discount Aisle: External DA layers compete on price for rollups willing to trade marginal security for ~90% cost savings.
- Hybrid Models Win: Protocols like Arbitrum and zkSync use Ethereum for finality but can leverage external DA for scale, creating a fluid, multi-DA ecosystem.
3-6
Blobs/Block
~90%
External DA Savings
03
Restaking Creates Hyper-Secure, Vertically Integrated DA
EigenLayer's restaking model allows ETH stakers to secure new services, creating a new security primitive. EigenDA is the flagship Actively Validated Service (AVS), leveraging restaked ETH to bootstrap trust.
- Capital Efficiency: Tap into Ethereum's $50B+ staked ETH for security, avoiding new token issuance.
- Vertical Integration: Rollups like Mantle and Celo use EigenDA, bundling execution, settlement, and DA into a cohesive, restaking-secured stack.
- Security Flywheel: More AVSs and rollups attract more restaked capital, increasing the cost of attack for all.
$50B+
Securing Pool
1.6 MB/s
Throughput Target
THE COST OF VERIFIABLE DATA
DA Layer Economics: A Comparative Snapshot
A first-principles comparison of cost structures and economic guarantees across leading data availability solutions.
| Feature / Metric | Ethereum (Blobs) | Celestia | EigenDA | Avail |
|---|---|---|---|---|
Current Cost per MB (USD) | $0.40 - $1.20 | $0.01 - $0.03 | $0.001 - $0.005 | $0.02 - $0.05 |
Data Availability Sampling (DAS) | ||||
Proof of Custody | ||||
Throughput (MB/sec) | 0.375 |
|
|
|
Economic Security (Stake) | $110B (ETH Staked) | $2B+ (TIA Staked) | $20B+ (EigenLayer Restaked) | $0.2B (AVAIL Staked) |
Settlement Finality Time | ~12 min (Ethereum) | ~15 sec (Celestia) | Instant (Ethereum Finalized) | ~20 sec (Avail) |
Native Interoperability | All L2s | Rollups via TIA | AVSs via EigenLayer | Rollups & Validiums |
Primary Revenue Model | Basefee Auction | Pay-per-byte + Inflation | Pay-per-byte | Pay-per-byte + MEV |
deep-dive
THE DATA LAYER
The Moat: Why This Isn't a Commodity
Data Availability layers are defensible infrastructure because they create a physical and economic moat that pure execution layers cannot replicate.
DA is physical infrastructure. Unlike virtualized execution environments, DA layers require globally distributed node networks, specialized hardware for data sampling, and massive bandwidth. This creates a capital-intensive moat that new entrants cannot spin up overnight, unlike a new L2 rollup on Ethereum.
The market consolidates to the cheapest, most secure provider. Rollups like Arbitrum and zkSync will route data to the DA layer that minimizes their fixed costs while maintaining security guarantees. This creates a winner-takes-most dynamic similar to AWS in cloud computing, not a fragmented commodity market.
Evidence: The Celestia flywheel. Celestia's modular design and data availability sampling enable exponential scalability for rollups. Its adoption by major chains like Arbitrum Orbit and Polygon CDK demonstrates that DA is a foundational primitive, not an interchangeable component. The network effect of rollup security attracts more rollups, which deepens the moat.
risk-analysis
THE DA THREAT MATRIX
The Bear Case: What Could Break the Thesis
Data Availability is the foundational bottleneck for scaling blockchains; these are the scenarios where the thesis fails.
01
The Modular Stack Re-Consolidates
The core assumption is that modularity (separating execution, settlement, consensus, DA) is optimal. A monolithic chain like Solana or a super-optimized L1 could achieve sufficient scale and low enough costs to make external DA irrelevant. If a single chain can process 100k+ TPS at <$0.001 per transaction, the complexity tax of a modular stack becomes unjustified.
100k+
Monolithic TPS
<$0.001
Target Cost/Tx
02
DA Becomes a Commodity Race to Zero
If Data Availability is just about storing and attesting to blobs of data, it risks becoming a low-margin utility. Competition between Celestia, EigenDA, Avail, and even Ethereum Danksharding could drive prices to near-zero, destroying the economic moat for specialized DA layers. The winner is the user, but the DA projects become low-value infrastructure.
~$0.10
Current Cost/MB
→ $0.001
Future Cost/MB
03
Security & Censorship Catastrophe
A critical failure in a major DA layer (e.g., a successful data withholding attack on Celestia or a liveness failure in EigenDA) would cascade through every rollup that depends on it, freezing billions in TVL. This systemic risk could trigger a mass migration back to Ethereum L1 for security, proving that decentralized security is non-negotiable and cannot be outsourced cheaply.
$B+
TVL at Risk
7 Days
Escape Hatch Time
04
The Interoperability Nightmare
A fragmented DA landscape with multiple, non-interoperable standards (Celestia's Blobstream, EigenDA's attestations, Avail's proof system) creates immense complexity for cross-chain applications. Rollups become siloed to their chosen DA provider, breaking composability and fragmenting liquidity, negating a core promise of the modular ecosystem.
4+
DA Standards
Fragmented
Liquidity & State
05
Regulatory Capture of Data
DA layers are, by design, global data broadcast networks. A regulator could target a dominant DA provider with legal action to censor or surveil transactions for entire rollup ecosystems. Unlike base layers like Ethereum with stronger decentralization claims, a VC-backed DA project could be a more vulnerable single point of control.
1
Legal Entity
100+
Rollups Impacted
06
The Throughput Illusion
The advertised throughput of DA layers (e.g., Celestia's 100 MB/s) is a theoretical maximum for data posting, not useful transaction processing. Real-world rollup throughput is gated by execution and proving (ZK or Fraud Proof) bottlenecks. If these other layers cannot keep pace, the excess DA bandwidth provides zero marginal utility, making its capacity a moot selling point.
100 MB/s
Theoretical DA
~100 TPS
Practical Rollup Limit
investment-thesis
THE DATA LAYER
Capital Allocation: Betting on the Plumbing
Data availability layers are the fundamental, fee-generating infrastructure that underpins the modular blockchain thesis.
Data availability is the bottleneck. Execution layers scale by offloading data, creating a persistent demand for cheap, secure blob space from Celestia, EigenDA, or Avail. The winner captures fees from every rollup transaction.
The bet is on commoditization. Execution is a competitive, low-margin business, but the data availability market is a natural oligopoly. It consolidates around a few providers with the cheapest, most reliable proofs, similar to AWS in cloud computing.
Evidence: Ethereum's blob fee market now generates millions in daily revenue post-Dencun, proving the model. Arbitrum and Optimism have migrated billions in transaction data to these external layers, validating the economic separation.
takeaways
THE INFRASTRUCTURE BOTTLENECK
TL;DR for the Time-Poor CTO
Execution is commoditized. The real scaling war is over who secures the data.
01
The Problem: Ethereum is a $3.5M/GB Hard Drive
Storing call data on L1 is the primary cost for rollups like Arbitrum and Optimism. This creates a direct trade-off between security and scalability.\n- Cost: ~$100k+ daily in L1 fees for major rollups.\n- Bottleneck: Limits TPS and makes micro-transactions economically impossible.
$3.5M
Per GB Cost
>90%
Rollup Cost
02
The Solution: Celestia & EigenDA
Separate data publishing from consensus. These specialized layers provide cryptoeconomic security for data at ~1% of L1 cost.\n- Modular Stack: Enables sovereign rollups and high-throughput execution layers.\n- Market Effect: Drives down costs for everyone, from Arbitrum Orbit chains to zkSync.
100x
Cheaper Data
~$20/GB
Sample Cost
03
The Killer App: Hyper-Scalable L3s & Appchains
DA layers unlock viable economics for application-specific chains. This is the endgame for vertical integration.\n- Throughput: Enchains can target 10k+ TPS without L1 friction.\n- Sovereignty: Teams control their stack, from MEV policy to gas token.
10k+
Potential TPS
L3s
Primary Use-Case
04
The Risk: Security vs. Cost Trade-Off
Not all DA is equal. Ethereum (highest security) vs. Validium (lowest cost) represents a spectrum. The market will stratify.\n- Security Slice: Projects choose based on asset value (DeFi vs. Social).\n- Interop Challenge: Bridges and oracles must adapt to multi-DA environments.
Spectrum
Security Model
Critical
For DeFi
05
The Metric: $/MB/sec
Forget TPS. The new infrastructure KPI is cost per megabyte per second of data throughput. This measures raw scalability economics.\n- Comparison: EigenDA targets ~$0.01/MB/day vs. Ethereum at ~$3500.\n- Driver: Directly enables sustainable high-frequency applications (gaming, perps).
$/MB/sec
New KPI
>100,000x
Improvement
06
The Bottom Line: DA is a Commodity Moat
Winning DA layers become low-margin, high-volume utilities—like AWS S3 for blockchains. The moat is in integration depth and liquidity of attestations.\n- Winner-Take-Most: Network effects from rollup adoption and tooling.\n- VC Bet: Not on the tech, but on which standard the ecosystem coalesces around.
Utility
Business Model
Ecosystem
Real Moat
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