The Future of Composability Requires Incentive-Agnostic Layers

The dominant composability layer isn't Ethereum or Solana—it's the opaque network of searchers, builders, and relays that reorder transactions for profit. This creates ~$1B+ annual extractable value and user-negative externalities like front-running.\n- Fragments State: Each chain's MEV ecosystem is a silo.\n- Distorts Incentives: Builders optimize for their cut, not protocol health.

Paradigms like UniswapX and CowSwap shift the burden from users (complex execution) to solvers (competitive fulfillment). This creates a declarative composability layer where the network's job is to satisfy a goal, not just process a tx.\n- User Sovereignty: Express what, not how.\n- Solver Competition: Drives efficiency and better prices.

A neutral, protocol-owned sequencer (e.g., Espresso, Astria) is the prerequisite for credible neutrality. It provides atomic composability across rollups without granting a single entity the power to censor or extract. This is the base layer for intent settlement.\n- Atomic Cross-Rollup Bundles: Unlocks new app paradigms.\n- Incentive-Agnostic Core: Execution is a public good, not a profit center.

Bridges like LayerZero and Axelar introduced the oracle/relayer model, but their token incentives often misalign with security. The future is light-client bridges with economic insurance (e.g., Succinct, Polymer) that minimize trusted components and verifiably slash for faults.\n- Verifiable Security: Cryptographic proofs > staked signatures.\n- Cost Structure: Fees cover risk, not marketing budgets.

Applications like The Graph and Goldsky index and serve blockchain data, but their pay-per-query models create walled gardens. The endgame is decentralized data lakes (e.g., Celestia Blobstream, EigenDA) where data availability is a cheap commodity, enabling anyone to compute on the raw state.\n- Unbundled Data Access: Separates storage from compute.\n- Permissionless Indexing: No gatekeepers on historical state.

The monolithic "do-everything" chain dies. The future is hyper-specialized execution layers (gaming, DeFi, social) connected by incentive-agnostic transport layers. Composability happens at the intent layer, not the VM layer. This is how we scale to 1B users without systemic risk.\n- Vertical Integration: Apps own their stack for performance.\n- Horizontal Composability: Secure, cheap communication between them.

Modern composability is subsidized by MEV extraction, creating a system where user experience is secondary to validator/searcher profit. This leads to predictable failures:

• Front-running and sandwich attacks degrade trust.
• Latency races centralize infrastructure around a few players.
• Protocol logic is distorted to maximize extractable value, not utility.

Bridges and L2s compete by issuing incentive tokens, not superior tech. This creates vendor-locked liquidity silos that undermine the cross-chain vision.

• TVL is rented, not owned, fleeing at the end of farm programs.
• Security is diluted across dozens of weakly-audited bridge contracts.
• User experience fragments as they chase yields across unstable corridors.

Composability's security ultimately rests on oracle price feeds. A system built on incentivized actors is inherently vulnerable to coordinated manipulation.

• DeFi legos become a house of cards; a critical oracle failure cascades.
• Cross-chain intent systems like UniswapX and Across introduce new trust assumptions in relayers.
• The solution isn't more oracles, but fewer points of failure through cryptographic verification.

Composability allows protocols to be drained of value by adjacent layers. Searchers and aggregators capture the margin, leaving core developers underfunded.

• The value capture stack is inverted: Infrastructure (e.g., LayerZero, EigenLayer) extracts more than the application.
• Sustainable business models are impossible when every function call is a leak.
• Innovation shifts from new primitives to new extraction techniques.

You can only optimize for two: Security, Decentralization, Universal Connectivity. Current solutions sacrifice one, creating systemic risk.

• Fast bridges (LayerZero) trust a small validator set.
• Secure bridges (IBC) are limited to similar consensus engines.
• Fully decentralized bridges are slow and expensive, killing UX.

The bear case is that composability cannot evolve beyond its Ponzi-like incentive phase. The only viable future is infrastructure layers that do not require token bribes to function.

• Verification, not validation: Systems must be trust-minimized by design (ZK proofs, light clients).
• Fee-for-service, not farm-and-dump: Sustainability must come from utility fees, not inflation.
• If this fails, we get a network of isolated, high-fee walled gardens.

Today's DeFi legos are glued together by searchers and builders who extract ~$1B+ annually in MEV. This creates toxic arbitrage loops, front-running, and unpredictable execution that breaks smart contract logic.

• Unreliable State: Your protocol's function is a profit vector for others.
• Fragmented Liquidity: Intents and orders are routed for extractive value, not optimal execution.

Shift from transaction-based to outcome-based systems. Users express what they want, not how to do it. Protocols like UniswapX, CowSwap, and Across use solvers competing in a free market to fulfill intents.

• Incentive-Agnostic Core: The protocol layer doesn't profit from routing, eliminating built-in rent-seeking.
• Better Execution: Solvers compete on price, leading to ~5-20 bps better rates for users.

The value accrual shifts from the bridge or sequencer to the verification and data availability layer. EigenLayer, Celestia, and Espresso are building credibly neutral layers that don't compete with applications.

• Shared Security: Apps inherit security from a pooled validator set, reducing ~90%+ of bootstrap costs.
• Universal Composability: A neutral base layer enables trust-minimized communication across rollups and app-chains via protocols like Hyperlane and LayerZero.

Monolithic blockchains (L1s) are replaced by a modular stack: execution, settlement, consensus, and data availability. This allows each layer to be incentive-agnostic and hyper-optimized.

• Execution Rollups: Arbitrum, Optimism, zkSync compete purely on performance and cost.
• Data Availability: Celestia, Avail, EigenDA provide bandwidth as a commodity, not a profit center.

Value capture will bifurcate. Application-layer tokens will face compression as their utility becomes a commodity. Long-term value accrues to the base layers that enable permissionless innovation without taking a cut.

• Avoid Application Tokens with embedded tolls; they will be arbitraged away.
• Invest in Infrastructure that enables incentive-agnostic composability and credible neutrality.

Build your protocol as if the underlying layers are public goods. Use intents, leverage shared sequencers like Astria, and rely on decentralized oracles like Pyth and Chainlink for verifiable data.

• Design for Solvers: Expose clear fulfillment criteria, don't bake in fees.
• Integrate Horizontally: Use cross-chain messaging that doesn't lock you into one ecosystem's extractive bridge.