Layer 2 adoption 2026 predictions: What will shape the next wave of Ethereum scaling

Ethereum ecosystem scaling has moved from theory to actual production-grade infrastructure. Optimistic Rollups, ZK Rollups, major Layer 1 upgrades like The Merge and The Dencun upgrade (EIP-4844), and other Layer 2 scaling solutions have helped transform Ethereum from a congested base layer into the core of a rapidly growing multi-layer ecosystem. 

The question is no longer if scaling for Ethereum will be possible, but what type of scaling will be implemented next? Which implementation will affect where developers build? How will rollup economics affect liquidity and value flows on-chain?

Let’s take a look at some key predictions that will accelerate Layer 2 adoption and define the next wave of Ethereum scaling in 2026.

Why 2026 is a critical year for Ethereum scaling

Surge in L2 adoption from 2023 to 2025

Layer 2 Total Value Locked has grown sharply since 2023. At the time, L2s collectively held just under $4 billion in value. By October 2025, that figure had climbed to roughly $47 billion. Daily transactions on L2s have followed suit, recording as high as 1.9 million transactions per day in 2025, eclipsing the Ethereum mainnet. 

The trend observed between 2023 and 2025 tells one story: a major shift is underway. Developers and users are increasingly turning to L2s for high-throughput use cases.

Ethereum hits mainstream payment & consumer applications

Between 2024 and 2025, Ethereum L2s experienced a surge in real-world use cases, with Base emerging as Coinbase’s consumer-focused rollup and a dominant network for Farcaster and lightweight SocialFi experiences. 

Likewise, gaming studios such as Atari, Lotte Group, Nexon, and Sky Mavis have adopted app-specific chains to reduce fees and enhance execution performance. This is not limited to gaming studios. Prediction markets, stablecoin payments, and tokenized real-world assets have also found sustainable homes on rollups, where gas costs no longer prohibit experimentation.

Bottlenecks

Despite the growth of Layer 2 solutions, which has resulted in improved transaction speeds and lower costs, certain scaling bottlenecks, such as data availability (DA), bridging efficiency, and price fragmentation, remain. The process of storing data required for transaction verification, in comparison to execution, is expensive, creating a scenario where many are shopping for cheaper data availability layers. 

Likewise, liquidity movement across Layer 2 bridges can often be subject to delays or require extra trust in third-party providers, which can sometimes complicate the process or expose users to risk. 

Ever heard about the same assets, like ETH, BTC, and USDT, existing on different chains, and with variable pricing? That’s liquidity fragmentation – an occurrence that reduces capital efficiency, spreads liquidity thin, and increases slippage, often forcing users to hop from one network to another, hunting for the best price.

These bottlenecks form the backdrop for the major predictions ahead.

The current state of layer 2s

Total L2 TVL

According to data from L2BEAT, the total L2 TVL at the start of 2025 was about $44B. It peaked at $49B in October, and in December, the value is at about $38B. Arbitrum One leads the pack with approximately 44% of L2 total value locked, followed by Base Chain at 33%, OP Mainnet at 6%, and Lighter at 3.5%.

The remaining TVL is distributed across ZKsync Era and emerging chains like Linea, Ink, Katana, and Starknet. Collectively, these networks support millions of daily active users.

Rise of consumer appchains

Consumer-focused appchains are emerging. They are designed and optimized around user experience, rather than strict decentralization trade-offs. Ethereum L2s like Base, Mantle, Blast, and Linea focus on simplified onboarding, low fees, and fast finality. Base, in particular, has become a general-purpose consumer chain and a soft entry point for non-crypto audiences. It leverages Coinbase’s infrastructure and integrates seamlessly with its 100 million+ users.

Modular L2 ecosystem expansion

As rollups scale, the modular blockchain thesis has gained traction, emphasizing the separation of execution, settlement, and data availability. Dedicated DA layers emerged as rollups sought alternatives to Ethereum’s limited and expensive DA capacity. Celestia, the first purpose-built data availability blockchain, offers significantly lower costs than Ethereum by focusing exclusively on DA and leveraging data availability sampling.

EigenDA approaches the problem differently by using Ethereum restaking to create a scalable DA layer secured by economic guarantees rather than native execution. This model enables a more flexible DA marketplace while maintaining close alignment with Ethereum’s security assumptions.

By 2025, additional DA solutions such as Near DA and Avail expanded the ecosystem, giving rollups a growing menu of DA options optimized for lower costs, higher throughput, and varying trust tradeoffs.

Starknet & zkSync’s progress

Zero-knowledge rollups have matured significantly. Starknet improved throughput via Volition, Cairo optimizations, and compressed proofs. It also proved to be production-ready for complex applications, such as decentralized exchanges and gaming. Likewise, zkSync expanded with Hyperchains, scaled to handle over 100 transactions per second with sub-cent fees, and enabled L3 networks to deploy with customizable execution environments.

OP Superchain vision rollout

The Superchain is a long-term vision for Optimism to create a network of OP Stack blockchains that are connected, interoperable, and share standards, infrastructure, and governance. The idea is to get multiple isolated chains to behave like a single ecosystem. Networks like Base, Zora, Frax, and Mode are part of the evolving ecosystem. Also, developers can now build rollups that are compatible with Superchain standards.

It is important to note that not all OP Stack chains are fully integrated with every Superchain feature. Many are at different stages and not yet configured for messaging or shared sequencing. Also, certain interoperability features are not yet fully realized. By 2026, this vision could converge further into practical interoperability.

Prediction #1 — ZK Rollups become the default for high-value transactions

Lower proving costs

By design, Zero-knowledge rollups offer better privacy, security, and faster finality when compared to Optimistic rollups. However, proving on ZK rollups is complex and computationally expensive. This means ZK rollups, at the moment, are not an economically viable option for low-value transfers and high-frequency consumer transactions. Having said that, several implementations are already underway to reduce ZK’s costs. 

These include decentralized proving marketplaces, algorithmic improvements to reduce compute per transaction, a shift to hardware acceleration (GPUs, FPGAs, and ASICs), and parallelized proving that splits proof generation across multiple machines. Additionally, a major structural unlock, such as recursive ZK proofs, is also in the works. 

A full-scale implementation of these approaches by 2026 is expected to narrow the cost gap, making ZK rollups economically competitive for a wider range of applications, including low-value and high-throughput transactions.

Faster finality

Optimistic rollups (ORs), by design, assume transaction validity and rely on a fraud-proof system with a 7-day dispute window. While this execution seems simple and cheap, finality takes a long time, and withdrawals are delayed. ZK rollups, on the other hand, offer instant finality and quick withdrawals, albeit at a higher computational cost at the moment. However, the speed of finality offered by ZK rollups will be a competitive advantage, especially for institutional settlements, cross-chain transfers, and high-value DeFi operations.

zkEVM maturity

With multiple zkEVM implementations expected to reach production maturity in 2026, the execution gap between zkEVMs and native EVM chains is expected to narrow even further.

Scroll matures into a stable, production-ready zkEVM with strong developer mindshare. Polygon zkEVM integrates deeply with the Polygon AggLayer architecture. zkSync Era runs a robust L2 with expanding L3 deployments.

Institutional adoption of ZK-based L2s

Zero-knowledge Layer 2 scaling solution offers security, privacy, settlement certainty, easier proof-based auditing, and lower bridging risks. Institutions that prioritize these features will increasingly look to zk-based systems. Big banks, mainstream payment processors, and even asset managers fall into this category and will likely take decisive steps towards building core applications on Zk-based Layer 2s.

Prediction #2 — Massive growth in L3s and app-specific rollups

Why apps prefer owning the full stack

By 2026, a lot of successful decentralized applications will launch their own app-specific Layer 3s, in a bid to “own the full stack.” In simple terms, “owning the full stack” means that applications have direct control over fees, performance, and are not dependent on shared Layer 2s.

They can abstract or subsidize gas, use stablecoins or custom tokens for payments, isolate themselves from network congestion, and capture all sequencer fees. For high-volume applications like games, prediction markets, and decentralized social protocols, a migration to OP Stack L3s or Arbitrum Orbit is likely.

On-chain gaming L3s, DeFi-specific rollups

Layer 3s are generally understood as app-specific rollups that live on top of L2s. They grant developers the ability to make specific changes to execution environments, adjust fee models, and tune performance. All of these are ideal for use cases where extreme throughput, predictable execution, and custom logic are required. 

For instance, an on-chain massive multiplayer role-playing video game or MMORPG requires frequent micro-interactions and executes thousands of state updates every second. A Layer 3 can batch and optimize these actions efficiently. But on a general-purpose L2, such workloads will struggle with network congestion, variable fees, and competition for blockspace.

Similarly, high-performance DeFi protocols like RFQ systems, options, and perpetual exchanges have needs that general Layer 2 networks, at the moment, do not optimize for, but L3s can facilitate. These needs include predictable execution, flexible liquidation logic, more control over transaction ordering, and the ability to capture MEV internally.

Arbitrum Orbit, OP Stack, zkSync Hyperchains

Arbitrum Orbit allows developers to create tailored chains that settle to Arbitrum, inheriting its security and tooling. The OP Stack already powers dozens of chains beyond Optimism itself, enabling networks like Base to emerge quickly and scale fast. zkSync Hyperchains offer a ZK-powered path for deploying app-specific rollups with strong security guarantees.

Together, these frameworks lower the complexity of launching new chains. By 2026, launching a rollup may approach the simplicity of deploying a smart contract, potentially sparking an explosion of specialized chains built for DeFi, gaming, RWAs, AI agents, high-speed trading, and more.

EVM standardization for L3 development

As new Layer 3 networks emerge, those already adhering to EVM standards will likely remain compatible with existing developer and user tooling. This means that developers can deploy Vyper or Solidity smart contracts without having to rewrite code. They can use existing tool stacks like MetaMask, Foundry, and The Graph, and rely on familiar workflows. 

In the case of users, it simply means the same wallet works across chains, transactions feel familiar, and there is no steep learning curve every time they try a new network.

By reducing the barrier to development and adoption, Ethereum Virtual Machine standardization for Layer 3s increases the likelihood that they will scale and integrate smoothly into the broader Ethereum environment.

Prediction #3 — Data availability wars intensify

Celestia vs EigenDA vs Ethereum’s Danksharding

By 2026, data availability is likely to become a competitive market. Celestia is projected to gain traction due to its low cost and modular data availability, which is favored by gaming L3s. On the other hand, EigenDA’s competitive advantage will be its Ethereum-aligned security guarantees via restaking.

Meanwhile, Ethereum’s danksharding roadmap continues to add more DA capacity, which in turn lowers costs while still retaining Ethereum’s security. The result is a multi-tier DA landscape in which rollups select DA solutions based on their specific cost, security, and performance requirements rather than relying on a single provider.

Cost differences drive L2 migration

Data availability represents 90% of L2 operating costs post-EIP-4844. We expect rollups to increasingly choose DA layers not based on ideology but economics. In other words:

Lower DA cost = lower rollup fees = more users.

This leads to the first major migrations between DA providers.

DA becomes the new scaling bottleneck

For rollups, batching will improve, execution will become more cost-effective, and fees are expected to decrease in 2026. Which means there will be more rollups posted back to Ethereum. The outcome of this? Blockspaces are quickly used up because Ethereum data availability (DA) is finite and expensive. Hence, DA becomes the new bottleneck limiting the efficiency of rollup operations. This constraint will drive innovation towards better data compression and validity proof schemes that minimize DA requirements.

Modular blockchains evolve into DA marketplaces

By late 2026, Data Availability no longer behaves like a static service, but it operates like a cloud computing marketplace. We’ll see variable demand-based pricing, competing fee schedules, different latency, trust, and security profiles. And the implication? Rollups will be able to select DA providers the same way developers choose between AWS, Google Cloud, or Cloudflare, based on cost, performance, and reliability.

Prediction #4 — The Superchain / unified liquidity vision gets real                                             

OP Superchain goes live across multiple chains

The Superchain is Optimism’s vision of getting many independent L2 chains to operate as a coordinated system. Layer 2 networks like Mode, Base, Metal, and the OP Mainnet are built on the OP Stack, which means that core components like messaging, execution, and upgrades are standardized across the board.

The fact that shared standards exist means that these chains can coordinate governance decisions, adopt common interoperability layers, and eventually move towards shared security and sequencing models. As this coordination improves, cross-chain liquidity is expected to deepen, and bridges become less visible to users.

Shared sequencing

By 2026, shared sequencing could enable atomic actions across multiple chains within a single transaction flow. A user could swap on Base, add liquidity on Optimism, and open a position on Mode, all at once, with every step succeeding or reverting together. This brings back Ethereum’s synchronous feel across the Superchain and removes the fragmentation that has slowed multi-chain DeFi.

Liquidity fragmentation reduces

The biggest user experience failure of 2024-2025 was fragmentation. Users have ETH on Base but can’t buy an NFT on Optimism without bridging. With the OP Superchain and similar shared architectures, liquidity becomes pooled across member chains. Market makers can deploy capital once and serve multiple networks, while traders tap into unified order books no matter which chain they’re on.

Coinbase/Base plays a major role

Base is expected to become the most widely used Layer 2 solution by 2026. Its massive Coinbase-powered user base and strong compliance profile make it the biggest onboarding funnel into the Ethereum ecosystem. Base also gains significant influence in Superchain governance and becomes the main hub for regulated, consumer-facing applications. As millions of new users enter through Coinbase, Base’s role in shaping the broader L2 landscape grows even more dominant.

Prediction #5 — AI agents drive huge on-chain activity

Autonomous agents using L2s for microtransactions

As Layer 2 fees continue to fall, they become ideal rails for autonomous agents. Machine-operated systems will be able to make constant microtransactions, pay for APIs, purchase model outputs, verify computations, and coordinate with other agents, all without human involvement. With cheap, permissionless payments and verifiable execution, L2s will emerge as the default economic layer for large-scale multi-agent networks.

zk rollups become preferred rails for AI

AI systems need verifiable results, privacy-preserving interactions, deterministic settlement, and predictable costs. ZK rollups offer all of these benefits, making them the ideal environment for autonomous agents. As AI-driven applications continue to grow, ZK-based networks will become the preferred execution layer for machine-to-machine transactions and verifiable computing.

AI-to-AI payments and coordination systems

Interaction between autonomous software agents, especially in the aspects of on-chain transactions, market making, and trading, is already underway today in decentralized finance systems. These AI-to-AI interactions are increasingly extending beyond trading to include participation in governance and payment for data or model access, all without requiring continuous human input. What is currently driving these sorts of interactions? Factors such as fast settlement, programmable execution, and low fees; requirements that are clearly within the capabilities of Layer 2 networks. 

By 2026, AI-to-AI payments and coordination systems are expected to expand even further. We can expect a growing share of on-chain activity likely to be handled by autonomous systems acting on behalf of users, organizations, and protocols.

Early examples: Autonolas, Fetch.ai, Ritual

Projects like Autonolas, Fetch.ai, Grass, and Ritual represent early building blocks for machine-to-machine coordination on the blockchain. For instance, Autonolas or Olas is a decentralized platform that facilitates the building, deployment and management of self-acting software entities, otherwise known as autonomous agents, for dApps, DAOs and off-chain services.

Fetch.ai also builds networks of autonomous economic agents capable of negotiating, transacting and optimizing outcomes across markets like finance, supply chain and energy. Ritual, meanwhile, focuses on bringing verifiable AI inference on-chain, such that smart contracts are able to trust the output of artificial intelligence models without the need for off-chain processes. 

While still evolving, these systems collectively highlight a growing trend, which is that blockchain is increasingly becoming a coordination and settlement layer for AI systems – more of which will be seen in 2026.

Prediction #6 — Consumer L2s explode

Farcaster frames + Base ecosystem

Farcaster frames are interactive components embedded in Farcaster posts. Farcaster itself is a decentralized social network. And Base? That’s a leading consumer-focused Layer 2 network. So, what’s the link?

Farcaster frames are able to host a variety of mini-applications like quizzes, polls, and live updates for games and products. Other than those, Frames also features the ability to integrate Web3 functionalities directly in social feeds, such that users can mint NFTs, tip creators, check airdrops, and make crypto payments, all within the same UI. 

For this model to scale, the underlying network needs to be cheap, fast, and reliable. That’s where Base comes in. It makes the in-feed actions on Farcaster feel instant and seamless. Users can smoothly transition between social actions and on-chain transactions without having to interface with the complexities of the blockchain.

The market can anticipate the emergence of more of this sort of integration that blurs the lines between financial activity and social interaction in 2026.

On-chain games using appchains/L3s

Games built fully on-chain require more throughput than what general-purpose Layer-2s currently offer. Elements that demand real-time processing, such as game-state updates, player actions, and asset transfer, often put pressure on shared networks. Application-specific L3s fix this by giving each game its own dedicated blockspace, while allowing developers to customize execution and even subsidize player fees. 

In 2026, on-chain games are expected to run on their own appchains/L3s, built specifically for high-speed, deeply interactive gameplay.

L2s with near-zero fees attract mobile users

Fees on Layer 2 are expected to approach zero in 2026. What does this mean for mobile users? Complex on-chain actions become affordable, further driving L2s toward becoming the default platform for global mobile adoption. With advances in compression, proof systems, and sequencing, social interactions, tipping, loyalty rewards, and micro-payments will become everyday on-chain behaviors.

UX improvements for non-crypto audiences

The goal remains to get users to interact with L2 apps without even knowing what a seed phrase or gas fees mean. Improvements in account abstraction, embedded wallets, and fiat on-ramps in 2026 will make the user experience with Layer 2 apps simpler and more intuitive, akin to a typical Web2 app, with seamless logins and instant actions.

Prediction #7 — Ethereum becomes a settlement layer, not a user layer

Rollup-centric roadmap fully realized

If Vitalik Buterin’s rollup-centric vision approaches full-scale actualization in 2026, the Ethereum mainnet will continue to remain the base layer for security, DA, and settlement, while day-to-day user activity happens on Layer 2 networks. This clear separation lets each layer operate more efficiently. The Ethereum mainnet serves as a settlement layer, while L2s, on the other hand, deliver a fast and low-cost experience.

99% of user activity moves to L2s

Almost all everyday activities are expected to shift to Layer 2 networks as they continue to outpace the Ethereum mainnet in terms of transactions, users, and developer deployments. This trend is already visible today and is accelerating quickly. Using Ethereum L1 directly becomes something reserved for large settlements, important governance decisions, or major protocol upgrades. 

That doesn’t mean Ethereum is failing. In fact, it means the opposite. L1 finally evolves into the secure foundation of the ecosystem, much like the internet’s base protocols that no one interacts with directly. 

ETH used as gas via EIP-4844 integration

Even as user activities are predicted to shift completely to Layer 2s, this does not replace the role of the Ethereum mainnet as the underlying infrastructure. Layer 2s still rely on Ethereum for security and DA. They also post transaction data back to Ethereum for settlement, and pay for execution in ETH. 

Likewise, validators and sequencers still hold ETH to pay DA fees, submit batches, and settle on L1. It does not matter whether a rollup abstracts gas fees, uses stablecoins, or its own tokens; at its core, it still relies on ETH to operate. This continues to fuel the demand for ETH at the infrastructure level, making it the “reserve asset” of Ethereum’s scaling stack.

Institutions settle on Ethereum, transact on L2s

An approach to expect will be financial institutions adopting a split model when using Ethereum. What does this mean? In simple terms, every day operations such as internal transfers, small payments, and similar activities that would be too slow or expensive for Layer 1s move to Layer 2s. On the other side, final settlement, balance reconciliation, custody, and similar high-value activities remain on Ethereum, where finality, decentralization, and security are strongest.

This gives institutions what they want, which is the security of Layer 1s, the execution of Layer 2s, and predictable compliance frameworks. The split between settling on Ethereum and operating on L2s will become the standard architecture for institutional blockchain use.

Prediction #8 — Layer 2 revenue models evolve

Sequencer profits grow

The role of sequencers is to order, bundle, and ensure that transactions are submitted to Layer 2 networks. In other words, L2 transactions pass through sequencers, which means, by nature, they sit at the center of MEV capture and fee collection. Hence, as Layer 2 adoption accelerates and monthly transaction volumes grow exponentially, sequencers are positioned to grow profit.

MEV markets mature

Maximal Extractable Value or MEV is the additional value that can be captured by controlling the inclusion, ordering, and exclusion of transactions in a block. This is an area that will see more structure to it in 2026.

Naturally, new transaction ordering mechanisms and faster block times will result in more MEV opportunities, albeit requiring coordination and democratization, which platforms like Flashbots already do. 

Stakeholders can expect an open, transparent and competitive transaction ordering market driven by cross chain MEV auctions and shared sequencers. A mature MEV market will ensure that value extraction aligns well with protocols and users.

Additionally, the openness, transparency, and democratization of the market will lead to benefits such as fairer fees and, where applicable, user rebates.

Shared revenue with builders

Optimism’s retroactive public goods funding model is expected to spread across the L2 ecosystem. More networks are beginning to direct a portion of sequencer revenue back to builders, rewarding teams based on the impact of their work. This ensures steady funding for open-source infrastructure, developer tools, and high-value community projects. 

By 2026, several L2s are predicted to adopt formal revenue-sharing systems that support L3 builders, service providers, and major protocol teams, strengthening the entire ecosystem.

L2 tokens: shift from “governance” to real cashflow assets

By design, most Layer 2 tokens were intended as governance tokens used to decentralize decision-making. L2 tokens are, however, expected to do more than just be governance tools and begin capturing real economic value. Today, rollups already generate revenue from DA costs, sequencing fees, and MEV, tying native L2 tokens to real cash flow.

By 2026, more networks will introduce revenue-sharing models, sequencer profit distribution, and yield tied to actual network usage. Governance will still matter, no doubt, but cash flow will be a significant part of what gives L2 tokens long-term value.

Prediction #9 — L2 interoperability becomes safe and native

Trust-minimized bridges 

One of the drawbacks of early cross-chain bridges that creates a major security concern is their reliance on trusted validators or multisigs. This means that users trust a group of operators to confirm transactions. And if these parties are compromised or act maliciously, the transaction could be at risk. 

Trust-minimized ZK bridges are shifting the narrative, with solutions like Succinct’s Telepathy already promoting L2 interoperability by utilising validity proofs to cryptographically verify state changes across chains without the need for trusted intermediaries. This approach is expected to become mainstream as it will make bridging faster, permissionless, reliable, and more secure.

Cross-chain intents

Intent-based systems like Anoma and Radius are expected to remove the complexity of choosing chains or handling bridges. Instead of telling the network how to execute a transaction, users simply state what they want, such as swapping tokens at the best price, moving liquidity to the highest-yield chain, or getting a specific amount of USDC on Arbitrum. Solver networks then figure out the optimal cross-chain route and execute it behind the scenes.

Shared blockspace

To further drive the goal of L2 interoperability, individual L2s like Arbitrum, OP, etc., will no longer order transactions, but shared sequencers will coordinate and finalize the process across multiple L2s simultaneously and instantaneously. 

On the user side, wallets display a single combined balance and unified transaction history, while apps automatically route transactions through the most suitable chain. In practice, users do not get to interact with the complexities; everything happens seamlessly in the background.

Prediction #10 — Regulation pushes U.S. builders toward L2s

Base emerges as the compliant L2 for U.S. teams

Base will emerge as the preferred L2 for US teams due to the strong regulatory standing and compliance-driven nature of Coinbase, its parent infrastructure. The link between Coinbase and Base means that the latter enjoys transaction monitoring, easy onboarding, built-in Know Your Customer (KYC) and AML support, and most importantly, a clearly spelled out legal and operational framework. These are the sort of features suited to fintech integrations, enterprise deployments, and consumer applications.

Stablecoin payments flow into L2s

As regulatory clarity continues to unfold for the already thriving stablecoins ecosystem, in terms of what can be issued and how they can be used, payment companies gain confidence to transition on-chain, especially to Ethereum L2s. But why Layer 2 networks? L2s become an attractive option because of their Ethereum-grade security, near-instant finality, and low fees.

These are a combination of features that appeal to teams handling high-volume payments, cross-border transfers, and payroll. Everyday payments are not left out. Compliance-driven stablecoins like USDC and PYUSD are predicted to move across L2s in the background, completing transactions worth billions of dollars monthly, while wallets abstract the complexity of the process (bridges, chain selection, and gas fees). 

What could slow Layer 2 adoption?

Security assumptions around L3s

Layer 3s depend on the security architecture of the Layer 2s they are built on. Likewise, Layer 2s rely on the security of the Ethereum (L1) mainnet. The cascading trust model could create a major security risk. Should a major L2 fail occur, the dependent L3s are similarly affected.

Sequencer centralization

Most Layer 2 scaling solutions still use centralized sequencers run by their core teams. With centralization comes censorship risks, a single point of failure, and exposure to regulatory pressure. If meaningful progress toward sequencer decentralization doesn’t happen by 2026, it could weaken the core value proposition of L2s and limit their long-term trust and resilience.

DA layer failures or outages

Data availability layers ensure that rollups make transaction data available at a reasonable cost. To achieve this, some DA layers rely on a different or, in some cases, weaker security architecture than Ethereum’s. This introduces a reliability risk, such that if a cheaper DA experiences a network outage or consensus failure, the dependent rollups face data fragmentation and eventual state inconsistency.

Liquidity fragmentation returning

Five L2s launch today, and two L3s launch a couple of days after. While this in itself provides users with variety, it’s a precursor to liquidity fragmentation. Instead of funds being concentrated in one place, they are spread across several networks, making trading less efficient and increasing borrowing costs. Even with shared sequencing and Superchain-style coordination, incentives can still pull capital in different directions, making liquidity fragmentation an ongoing risk.

U.S. regulatory ambiguity

Crypto regulation in the US is often regarded as ambiguous. Why? Multiple state regulators and agencies always have something to say, from the SEC to the CFTC, Treasury, and FinCEN. Also, the rules around crypto in general, Layer 2 tokens, and stablecoins remain unclear. 

For developers and enterprises, enforcement risks remain difficult to assess, creating operational and legal uncertainty. The outcome? Adoption slows down, builders are forced to adapt via a compliance-first design, restrict features in the US, or shift efforts toward other jurisdictions with clearer regulatory frameworks.

What does 2026 look like for Ethereum scaling?

A multi-L2 world dominated by Base, Arbitrum, Optimism, and zkSync

In the coming months, the Ethereum L2 ecosystem is expected to experience less fragmentation and consolidate around a few dominant players. First would be Base, an L2 that has positioned itself strategically as the hub for consumer-focused applications. 

Base’s integration with Coinbase and its hundreds of millions of users simplifies crypto onboarding for a non-crypto audience, guarantees fiat access, and regulatory oversight. This is attractive to mainstream applications.

Arbitrum will likely continue to hold its base as the home of DeFi and gaming due to its deep liquidity, solid developer community, and mature tooling. It is also renowned for its complex financial applications and high-throughput use cases. 

Optimism’s Superchain and enterprise rollups are also expected to thrive, becoming the backbone for interoperable application ecosystems. The fourth, zkSync, which anchors ZK-based Layer 2s and Layer 3s, will take the lead in high-value transactions,  institutional use cases, and privacy-sensitive applications.

Ethereum as a global settlement, DA + rollups as execution

Ethereum’s role evolves into a global settlement and data availability layer, securing billions of transactions happening on L2s. Direct user activity on the mainnet becomes minimal, while rollups handle day-to-day activities, such as consumer apps, trading, payments, gaming, and AI interactions. This architecture delivers massive scale while keeping Ethereum’s core security and decentralization intact.

Millions of daily active users across L2s

Taking a cue from the growth observed in user and daily transactions as of 2025, the numbers are expected to surpass current levels by 2026. Ecosystems like Base and Arbitrum will process millions of transactions daily, with users onboarded from around the world: DeFi traders in Asia, gamers in Europe, payment users in Latin America, and global social media participants. 

Rollups become the new app stores

Rollups will evolve from just being a tool to cut down fees and scale Ethereum into a holistic platform that hosts ecosystems of applications, as is the case with today’s app stores.

Beyond just bundling transactions off-chain and posting them back to Layer 1, rollups will host decentralized applications, help users discover them, provide built-in tools and liquidity, and handle execution in the background.

Additionally, as L3 frameworks grow on top of L2 rollups, these networks begin to feel less like a single-purpose scaling layer and more like a full application ecosystem where distribution, monetization, and execution are coordinated at the platform level.