In 2026, development teams abandon the monolith logic. They no longer pick a single blockchain like Ethereum, Solana or Avalanche to deploy their application and accept all the design trade-offs that choice imposes. Now they assemble their own stack with independent modular components. Infrastructure investors observe a shift in value capture: base layers unbundle their functions and each piece competes as a distinct financial product. Execution and data availability emerge as the two business lines that define token demand and profitability in this new cycle.
Unbundling the chain: why capital cares
A blockchain delivers four services: it orders transactions and reaches consensus, it executes code, it publishes the data that lets any participant verify the state, and, in some designs, it settles disputes on a more secure external layer. Until 2023, the dominant protocols bundled these services. Teams paid gas fees for the whole package and token holders captured the aggregated value of all functions.
Modularity breaks that bundle. A rollup can pay a specialized provider for consensus and data availability while choosing a different execution environment for its smart contracts. This separation expresses a financial reality: fee revenue no longer flows by default to the monolithic base layer. The market fragments value capture. Investors who understand the new structure allocate capital to the layers that charge recurring fees for specific services, not to those that attempt to do everything.
Execution: the engine that charges per verified instruction
The execution layer processes transactions and computes smart contracts. It operates as a verifiable CPU. For years the Ethereum Virtual Machine (EVM) concentrated almost all activity and revenue. Tokens of EVM-compatible chains accumulated value precisely because users paid gas to execute code on that standard.
In 2026, the execution market diversifies, not for technological whims but for profitability needs. The EVM imposes gas limits, high costs on complex operations and a fee architecture that penalizes long or intensive computations. This pushes entire application categories out of the market: decentralized artificial intelligence, financial simulations, on-chain game engines and privacy protocols using homomorphic encryption.
New execution environments capture those workloads with more efficient fee models. The Cartesi Machine, for example, packages a full Linux system and charges per verifiable instruction outside the EVM’s constrained environment. Developers write contracts in Python, C++ or Rust and use standard libraries. The market pays for that execution because it lowers operational cost and enables applications that previously required centralized servers.
Other environments, such as programmable zero-knowledge machines, charge for processing encrypted data without revealing information. Privacy becomes an execution product with its own fee model.
Investors look at this layer as a verifiable compute infrastructure market. A token’s valuation multiple no longer depends solely on EVM compatibility; it depends on the sustained demand from applications running on that engine, the resources they consume and the fees they generate.
Data availability: the verifiable cost per byte published
The data availability (DA) layer publishes the information that any observer needs to reconstruct the chain’s state and verify transaction validity. Without public DA there is no auditability and no real decentralization. But publishing data costs money: every byte an application writes to the DA layer consumes fees and competes for block space.
Ethereum operates as the most secure and most expensive DA layer. Blob fees — introduced with EIP-4844 — lowered costs for rollups, but keep limited capacity and a pricing dynamic linked to mainnet congestion. For a social application publishing millions of posts daily or an on-chain game with thousands of moves per second, the Ethereum DA bill squeezes margins and makes the business model unviable.
Specialized DA layers enter the market with a drastically lower price per byte and large blocks. Avail DA, for instance, implements data availability sampling that allows verifying block publication without downloading it entirely. It combines that with confirmations of roughly 20 seconds assisted by validity proofs and plans blocks up to 10 GB. The cost per published byte falls two orders of magnitude compared to Ethereum.
Celestia, EigenDA and other networks compete in this same segment with different fee structures, decentralization levels and speeds.
Institutional capital reads this situation as a commoditization of the DA service. Ethereum’s inherited security keeps a premium for high-value settlements. Alternative layers capture the volume from applications with tight margins and high data frequency. Investors who analyze DA tokens do not look only at market capitalization; they examine fee revenue, the volume of data published by rollups and client applications, and the long-term retention of those clients.
The investment question formulates as follows: who monetizes the verifiable data flow that Web3 needs?
Aggregation and fragmentation risk: the financial bottleneck
The specialization of execution and DA introduces a risk the market already prices into certain valuations: liquidity and user fragmentation.
Each team picks its own stack and that produces disconnected islands. A user who holds assets on a rollup with EVM execution and Celestia DA cannot interact directly with an application running on a ZK machine that publishes its data to Avail without going through bridges, intent protocols or chain abstraction solutions.
This problem has a direct financial reading: fragmentation raises cross-ecosystem transaction costs, slows capital velocity and drains value toward interoperability intermediaries.
Investors detect an opportunity there. Aggregation protocols, shared validity proof providers and unified settlement layers capture fees for solving fragmentation. The modular stack does not eliminate intermediaries; it redistributes their commissions among specialized layers. The interoperability market becomes an infrastructure sector with its own revenue, correlated but not identical to that of execution and DA layers.
Value flows to the layers that bill
Modularity is not a software architecture fad. It represents a change in the revenue structure of crypto infrastructure. In 2026, development teams buy execution and data availability as separate services. The protocols that offer the best price per unit of verifiable compute or the lowest cost per published byte — with sufficient security — capture market share and generate recurring fee flows.
My opinion is that the market increasingly punishes the tokens of base layers that try to force bundling. Capital rotates toward the modular pieces that demonstrate sustained fee traction and toward the aggregators that mitigate fragmentation.
Investment teams that understand the modular stack do not ask whether Ethereum or Solana will win; they ask which execution engine collects the fees from the next million users of decentralized AI and which DA layer bills the bytes of those applications.
The answer builds infrastructure portfolios in 2026.
