Ethereum's Execution and Settlement Layers: What a CFO Needs to Know

On February 18, the Ethereum Foundation published its 2026 protocol priorities, consolidating execution and data-availability scaling under a single "Scale" track and targeting a gas limit above 100 million -- more than triple the 30 million ceiling that held from the Merge through early 2025. For any institution evaluating blockchain infrastructure, the restructuring raises a practical question: how does Ethereum actually process and finalize transactions, and what does that mean for reporting, controls, and reconciliation?

How Execution and Settlement Work Together

Ethereum operates as a two-layer machine. The execution layer handles computation -- running smart contract logic, updating account balances, and ordering transactions within each block. The consensus layer handles agreement -- validators stake ETH as collateral, attest to proposed blocks, and finalize the chain's state. These layers run on separate client software (execution clients like Geth or Nethermind and consensus clients like Prysm or Lighthouse), but they communicate through an internal Engine API on every node.

When a user submits a transaction, it enters the execution layer's mempool. A block proposer -- selected pseudo-randomly from the validator set -- bundles transactions into a block, executes them, and passes the result to the consensus layer. Attesters then vote on the block's validity. Under normal conditions, a block reaches "justified" status within two epochs (roughly 12.8 minutes) and "finalized" shortly after. Once finalized, reverting that block would require destroying at least one-third of all staked ETH -- currently over $30 billion at approximately 34 million ETH staked -- making settlement economically irreversible rather than merely probabilistically unlikely.

For a CFO's purposes, this distinction matters. Finality on Ethereum is not a timestamp; it is an economic guarantee backed by validator collateral. That changes how you think about settlement risk compared to traditional payment rails, where finality depends on institutional rules and business-day conventions.

What Changes for Reporting and Controls

The operational difference between Ethereum settlement and traditional finance shows up in three areas that any finance team should understand.

First, reconciliation. On legacy rails, settlement occurs on a T+1 or T+2 cycle, and reconciliation involves matching entries across custodian statements, clearinghouse records, and internal ledgers. On Ethereum, the canonical ledger is the chain itself. Every transaction carries a unique hash, block number, and timestamp. A finalized transaction is visible to every participant simultaneously. This compresses the reconciliation window but demands that internal systems can read on-chain data and map it to general-ledger entries in near real time.

Second, controls. Ethereum transactions are authorized by cryptographic signatures, not by a name on a wire instruction. Operational controls shift from "who approved this in the ERP system" to "who holds the private key, and what policy governs its use." Multi-signature wallets, time-lock contracts, and on-chain governance modules can enforce approval workflows, but the control framework is fundamentally different from traditional authorization matrices.

Third, audit trails. Every state change on Ethereum is permanently recorded and publicly verifiable. This is more transparent than most internal ledgers, but it also means that errors -- a misrouted payment, an incorrect contract call -- are equally permanent. There is no reversal mechanism equivalent to a chargeback or wire recall. Risk mitigation depends on pre-execution controls such as simulation, allowlists, and spending caps rather than post-execution remediation.

Where It Breaks

The architecture is not without constraints. Finality requires network liveness; if more than one-third of validators go offline simultaneously, the chain stops finalizing (though it continues producing blocks). This happened briefly on Ethereum's Holesky testnet in early 2025 during the Pectra fork testing and prompted the Foundation to add a dedicated "Harden the L1" track to the 2026 roadmap. Mainnet has not experienced a finality failure since the Merge in September 2022, but the risk is nonzero and should be modeled.

Gas costs introduce another friction. Each computation and storage operation consumes gas, and fees fluctuate with demand. The Fusaka upgrade in December 2025 raised the block gas limit to 60 million, roughly doubling throughput from its post-Merge baseline. The 2026 target of 100 million and beyond, combined with parallel execution in the planned Glamsterdam upgrade, may ease congestion further -- but fee volatility remains an operational variable that traditional settlement systems do not have.

Finally, the Layer 2 ecosystem adds complexity. Most high-volume activity now occurs on rollups that post compressed data back to Ethereum for settlement. L2 transactions are faster and cheaper, but they introduce a dependency chain: the rollup's sequencer, the data-availability layer, and L1 finality all factor into the end-to-end settlement timeline.

What Is Improving

The 2026 roadmap addresses several of these frictions directly. The unified Scale track is pushing gas limits higher while introducing enshrined proposer-builder separation (ePBS) to improve block production fairness. The Improve UX track targets native account abstraction, which may simplify institutional key management by embedding programmable authorization logic at the protocol level. And the Harden the L1 track, with its "Trillion Dollar Security Initiative," explicitly aims to make Ethereum robust enough to serve as settlement infrastructure for global-scale financial activity.

Ethereum execution and settlement -- key components from submission to finality:

• Execution Layer (Geth, Nethermind) -- processes transactions, runs smart contracts, updates state. Users pay gas fees; validators earn priority fees. Current gas limit: 60 million, targeting 100 million+ in 2026.
• Block Proposer -- bundles transactions into a block every 12-second slot, executes them against the current state, and passes the result to the consensus layer via the Engine API.
• Consensus Layer (Prysm, Lighthouse) -- validates blocks and manages finality through proof-of-stake. Approximately 34 million ETH (~$30 billion+) is staked as collateral. Validators are slashed for misbehavior.
• Attesters -- vote on block validity. Finality is reached in roughly 12.8 minutes (2 epochs). Reversal after finality would require destroying at least one-third of all staked ETH.
• Layer 2 Rollups -- execute transactions off-chain and post compressed data plus proofs back to L1 via blobs. Over 530 million monthly L2 transactions were recorded in late 2025.

Ethereum's settlement architecture splits computation from consensus. Validators stake collateral to guarantee finality, creating an economic security model distinct from traditional clearinghouse structures. The December 2025 Fusaka upgrade raised the gas limit to 60 million, with 100 million+ targeted in 2026.

Ethereum's execution and settlement split may look unfamiliar to teams accustomed to omnibus custodians and centralized clearinghouses, but the underlying logic -- separate the work of processing from the work of agreeing -- is not new. What is new is that the agreement layer is secured by economic collateral rather than institutional trust, and the entire ledger is open to inspection by any participant at any time. Whether that tradeoff suits a given institution depends on the use case, but understanding the mechanism is the prerequisite for making that judgment.

For informational purposes only. Not an offer to buy or sell any security. Available only to accredited investors who meet regulatory requirements.

Sources:

1. Ethereum Foundation, "Protocol Priorities Update for 2026," blog.ethereum.org, February 18, 2026.
2. Ethereum.org, "Fulu-Osaka (Fusaka)" upgrade documentation, ethereum.org/roadmap/fusaka.
3. Fidelity Digital Assets, "The Fusaka Upgrade: Scaling Meets Value Accrual," research report, October 2025.

‍