Introduction

Ethereum’s scaling challenge isn’t a bug—it’s a design choice with consequences that ripple through every layer of the network. Throughput caps and state bloat force hard tradeoffs. Validators shoulder storage burdens that climb toward terabytes, yet execution stays anchored to predictable, verifiable limits that keep home staking within reach. This tension between accessibility and performance defines the current roadmap. It also shapes how value flows, how applications compose, and who captures the ordering rights that sit just beneath the surface of every block.

Base Layer Limits and the Rollup-Centric Roadmap

Ethereum’s L1 doesn’t chase raw speed. The block gas target caps throughput intentionally, leaving each 12-second slot with constrained execution room. State size grows relentlessly, forcing validators to store hundreds of gigabytes just to stay synchronized. That combination isn’t accidental—it nudges computation off-chain while preserving security at the foundation.

It’s a deliberate architecture. Still, it raises costs for home stakers even as research pursues pruning designs and stateless validators that could ease the burden without sacrificing verification guarantees. The picture isn’t entirely clear. Bandwidth demands persist regardless of storage optimizations, and latency asymmetries favor operators with robust connectivity over those running nodes from residential networks.

Worth noting: EIP-4844’s proto-danksharding reshaped the fee landscape overnight. March 2024’s Dencun upgrade introduced blob space, letting rollups publish data far more cheaply than before. Fees that once hovered between $0.50 and $1.50 per rollup transaction collapsed to roughly $0.01–$0.05 after blobs went live. With data availability priced low, rollups can batch thousands of transactions without overwhelming L1 calldata. Most user activity now migrates to Layer 2, anchoring only proofs and checkpoints on mainnet.

This cements Ethereum’s evolution. L1 becomes settlement plus data layer, not a high-throughput execution engine. Rollups handle the volume; Ethereum guarantees finality and stores the commitments that let users reconstruct state or challenge fraud.

PeerDAS and data availability sampling moved to testing on Fusaka Devnet-3 in July 2025. Developers pushed gas limits to 45 million while trialing Peer-to-Peer Data Availability Sampling alongside Verkle commitments. The goal? Let nodes verify data availability without downloading every byte, enabling higher throughput without centralizing hardware requirements.

If successful, mainnet could raise gas targets while keeping verification accessible to diverse validator sets. If not, the network faces a choice between staying constrained or accepting drift toward professional-grade infrastructure. That’s tension worth acknowledging.

Bridges, Standards, and Cross-Chain Flow

ERC token standards underpin composability across DeFi, NFTs, and real-world assets. ERC-20, ERC-721, and emerging ERC-6551 give developers predictable interfaces—fungible tokens, collectibles, token-bound accounts. Because these standards hold consistent on both L1 and L2, liquidity and tooling carry across environments with minimal friction. Wallets, explorers, and custody platforms don’t need custom logic for every deployment.

That consistency explains why DeFi protocols and RWA issuers anchor on Ethereum even as execution migrates elsewhere. Standards create network effects. They reduce switching costs and enable protocols to compose without reinventing primitives.

Bridges and cross-chain messaging enable multi-chain workflows, but they introduce exploit risk that sits outside Ethereum’s base security guarantees. Research catalogues losses in the hundreds of millions: Wormhole lost 120,000 ETH, Ronin lost $620 million, Nomad bled $156 million. These incidents highlight how validator multisigs and message verification become attractive attack surfaces when bridge design fails to match the rigor of L1 consensus.

Bridges remain necessary for moving assets between rollups and alternative Layer 1 chains. Yet every hop adds trust assumptions and failure modes that bypass the economic security validators provide on Ethereum itself. It’s easy to overlook this when bridges work smoothly, but the risks compound with each additional chain in a workflow.

Tokenbound accounts—formalized in ERC-6551—expand NFT utility into identity and gameplay. These wallets let NFTs control other assets or permissions directly, turning collectibles into programmable identities that hold funds, approve transactions, or gate access. That deepens interoperability across applications. Games and platforms can treat an NFT as both character and wallet simultaneously, settling everything on Ethereum standards without custom bridges or wrappers.

Still, compromised NFTs could move funds if permissioning isn’t carefully designed. Wallets and marketplaces must adapt UX and security models to account for the new attack vectors this composability unlocks. The standard is powerful; the safety considerations are non-trivial.

MEV, PBS, and Fairness Experiments

MEV-Boost separates builders from proposers, coordinating bids through relays that ensure validators pick the highest-value block without seeing its contents before commitment. Specialized builders optimize transaction ordering to maximize MEV—capturing arbitrage, liquidations, sandwich opportunities—then submit bids. Validators select the top bid, earning revenue that would otherwise vanish into miner or searcher profits.

This captures value at the protocol level rather than letting it leak to uncoordinated extractors. It also inserts relay trust into the critical path between block construction and proposal. Relays coordinate fairness, but they’re third parties. They operate outside core consensus rules.

Collusion and low-bid cartels threaten decentralization more than most users realize. If builders coordinate to submit artificially low bids, they starve proposers of revenue while concentrating market power among a handful of sophisticated operators. Research warns this dynamic could centralize block production even as validator counts rise, because the real leverage sits in construction, not proposal.

Enshrined PBS would move block auctions on-chain, reducing reliance on external relays and shrinking censorship surfaces through protocol-level enforcement. Until that lands, validator freedom depends on a small set of relays behaving honestly—a setup that doesn’t quite fit with Ethereum’s broader decentralization goals. The tension persists.

MEV strategies—sandwich attacks, arbitrage, liquidations—drive fee dynamics and fuel censorship debates. Builders reorder mempool transactions to capture profit from price slippage, front-running user swaps or triggering liquidations milliseconds before they’d happen organically. These tactics shape gas auctions and can motivate selective censorship if certain transaction flows prove more lucrative than others.

Community debates center on whether MEV markets can remain open without entrenching gatekeepers. When OFAC-aligned validators refuse certain transactions, the boundary between protocol neutrality and regulatory compliance becomes visible. Some blocks skip transactions tied to sanctioned addresses; others don’t. MEV incentives interact with compliance pressures in ways that aren’t yet fully resolved.

Composability Stress and L2 Fragmentation

Rollup proliferation splits liquidity across execution environments, yet every L2 inherits Ethereum’s finality and security through settlement on L1. Multiple rollups compete for users, fragmenting liquidity pools and forcing frequent bridging. Settlement on Ethereum keeps ultimate security consistent, but fractured markets increase slippage and operational complexity for protocols trying to maintain coherent user experiences.

Deploying across multiple rollups or relying on aggregators becomes necessary to reach users wherever they transact. This adds coordination overhead that wasn’t present when everything ran on a single chain.

Data cost cuts push activity to L2 while L1 anchors settlement and checkpoints. With blob pricing in place, everyday swaps and NFT mints live on rollups where fees measure in cents rather than dollars. Ethereum stores the proofs and data commitments that let users exit or challenge invalid state. That reduces average transaction costs dramatically, but it makes UX dependent on sequencer reliability and bridge uptime—dependencies that didn’t exist in the old model where L1 handled everything directly.

It also shifts revenue composition. More fees come from L2 data blobs posted to Ethereum; less from direct calldata or execution gas on mainnet. This changes the economics for validators and raises questions about long-term value capture if most activity happens off-chain yet pays minimal rent to L1.

Cross-rollup MEV and fragmented UX remain open design tensions. As rollups multiply, MEV surfaces at bridge and router layers where ordering games can span multiple domains. Users face differing sequencer rules, varying fee markets, and inconsistent finality windows depending on which rollup they’re using. That creates cognitive load and trust fragmentation.

Researchers and builders explore shared sequencing or intent-based routing to unify the experience without sacrificing neutrality or performance. No canonical solution balances all three yet—neutrality, cost efficiency, and simplicity. The ecosystem is still searching.