Overview

Ethereum's rollup-centric roadmap promised L2s would function as trustless extensions of the base layer, secured by canonical bridges and cryptographic proofs. The market has delivered something entirely different: a parallel ecosystem where external bridges capture the majority of L2 volume, two-thirds of locked value exists outside Ethereum's security guarantees, and security councils retain emergency powers that can override on-chain settlement.

This report quantifies the gap between L2 architecture and market reality to trace how this divergence emerged from rational economic choices: users paying premium fees for instant bridging over seven-day security, projects issuing native tokens to capture value locally, and bridge providers building moats through network effects rather than Ethereum-derived security guarantees. The analysis demonstrates that this isn't a permanent fracture but a technological gap with a closing window—emerging solutions like based and native rollups and production ZK can soon reunify high-value rollup applications around Ethereum settlement, while sovereign architectures offer an alternative for use cases that genuinely don't require Ethereum's security properties.

Measuring Ethereum Alignment

To understand the current state of rollups, we need a framework for understanding where the ecosystem is headed. The recently proposed Ethereum Settlement Score (ESS) aims to measure factors like canonical bridge usage, asset composition, and censorship resistance to quantify how much an L2 actually leverages Ethereum versus operating as an independent system with nominal Ethereum branding. Current scores reveal that most Ethereum L2s today utilize Ethereum's security guarantees more in theory than in practice.

This measurement matters because the technology to reunite L2s with Ethereum is approaching production readiness. Native Rollups will eliminate security councils entirely by integrating directly with Ethereum's execution engine. Based sequencing will restore Ethereum's censorship resistance to L2s by enabling real-time force inclusion. Zero-knowledge proofs will make canonical bridges instant and competitive with third-party solutions. Together, these innovations chart a path from today's fragmented ecosystem to a future where L2s truly function as extensions of Ethereum.

The staging framework for this evolution extends beyond L2Beat's current model. While Stages 0-2 track the progression from fully centralized to partially decentralized systems, Stage 3 represents Native Rollups that completely eliminate intermediate trust assumptions like security councils and multisigs. Stage 4, the UltraSound L2 ideal, combines Native Rollup architecture with Based sequencing and perfect Ethereum utilization to achieve both architectural alignment and practical usage that maximizes Ethereum's security guarantees. Understanding this trajectory reveals two distinct paths for the rollup ecosystem: reunification with Ethereum through new technology, or embracing sovereignty through alternative architectures. Both are valid, but they serve fundamentally different purposes.

Rollups: Definitions, Dependencies, and Divergence

Semantics Matter

Despite the recent rhetoric, rollups are growing in popularity with L2Beat listing over 50 as of May 2025. This does not include the 80+ validiums, optimiums, and more. However, as the rollups landscape expands, discussions around scalability often devolve into teams talking passionately past each other, debating architecture, security assumptions, or performance claims without agreeing on a common vocabulary. Much of this friction stems from a lack of shared definitions. For this report, we define a rollup, in its simplest and most useful form, as a blockchain that posts its transaction data to another chain. In other words, a chain that relies on another chain for data availability. This definition draws a clear line between rollups and other scaling paradigms like validiums, sidechains, etc., that rely on different trust and availability models. 

Issuer Matters

L1-issued assets (e.g., ERC-20 tokens on Ethereum mainnet) inherit full Ethereum security. Their state, balances, and transfers are enforced directly by Ethereum’s consensus. However, L2-issued assets (e.g., tokens minted natively on Arbitrum or Base) are secured conditionally by the L1. Their security depends on the L2’s bridge and its ability to post accurate state roots or proofs to Ethereum. If the L2 operates honestly and the bridge is secure, the L2 asset can be redeemed or verified via Ethereum. However, this depends on the correctness of the fraud/validity proof mechanism, sequencer availability, and the bridge used. 

So, while Ethereum provides settlement guarantees for properly functioning rollups, L2 assets are not natively enforced by Ethereum. If the L2 halts or is compromised (e.g., through censorship, failed proofs, or bridge bugs), Ethereum doesn’t automatically step in to protect the asset’s integrity. In contrast, an L1 asset lives entirely within Ethereum’s trust boundary. Furthermore, forced bridging (a.k.a. forced withdrawal or escape hatch) is only guaranteed for L1-native assets bridged to the L2, not for assets that are natively issued on the L2. If you deposit ETH into an L2 like Base or Arbitrum, it's held in an L1 bridge contract. Even if the L2 sequencer is down or censoring, you can force an exit by submitting a transaction directly on Ethereum using the rollup's bridge contract and finality window (e.g., 7-day challenge period on optimistic rollups). This is possible because the canonical L1 bridge knows the L1 side of the transaction and can eventually verify your L2 state commitment via fraud or validity proofs.

However, L2-native assets don’t exist on L1 unless the issuer explicitly builds a canonical bridge or minting logic on L1. If the L2 halts, gets rugged, or censors, there is no guaranteed way to withdraw the token back to L1 unless the issuer or protocol has created a redeemable version on L1 (not common) or the token contract has logic supporting redemption or migration.

Below we show the percentage of canonical vs non-canonical assets on the two largest rollups by assets secured. Canonical assets comprise native ETH and ERC-20 tokens deposited through the official Ethereum L2 bridge, with balances recorded in the rollup’s bridge contracts on Ethereum.

We see that roughly two-thirds of assets on these rollups are non-canonical. This concentration highlights a growing dependency on rollup-native infrastructure and trust assumptions. As more applications launch tokens directly on rollups, whether for gas abstraction, ecosystem incentives, or user engagement, they bypass Ethereum's core base-layer protections. In practice, this creates fragmentation in security guarantees across the rollup stack, where a large portion of value is effectively siloed within environments that Ethereum cannot natively enforce. It also signals a shift in user behavior: convenience and low fees are increasingly prioritized over guaranteed finality and exit rights. 

As more assets are launched directly on rollups, it’s important to distinguish not just where they live, but where their liabilities are recorded and enforceable. This is especially critical for real-world assets (RWAs), where the value of the token is a direct claim on an offchain liability. If that liability is issued on an L2 without corresponding recognition on Ethereum L1, users have no guaranteed path to enforce redemption if the rollup becomes inaccessible or fails. In these cases, settlement is not secured by Ethereum, but instead hinges entirely on the issuer’s willingness to honor obligations offchain or across networks. The structure of issuance, whether liabilities are recorded on Ethereum or within an L2 silo, has direct implications for risk. For example, when $BUIDL (which represents a portion of Blackrock's liabilities) is issued onchain, the responsibility for honoring those liabilities in the event of network issues rests with Blackrock/Securitize, the issuer, and not with the validators. Furthermore, these external assets (and liabilities) also consist of assets brought in via third-party cross-chain protocols such as Circle-minted USDC, LayerZero USDT, and other interoperability solutions reflecting liquidity supplied by non-canonical bridge providers. Below we break down how much of the asset supply on major rollups falls into these categories.

Below we outline the differing security assumptions behind bridged and native assets, emphasizing that an asset’s trust model depends on how and where it is issued. In contrast to canonical assets, external assets such as USDC via Circle’s CCTP or OFT tokens via LayerZero rely on offchain issuers or third-party messaging endpoints, shifting trust away from Ethereum. Native assets, like KAITO on Base, are secured by the L2 itself, often governed by a multisig, making the L2's own security and governance the ultimate arbiter of user funds.

Native L2 assets like KAITO on Base represent a business success by offering agility, direct governance, and alignment with the L2's growth strategy. Projects can issue tokens without relying on Ethereum L1 or complex bridges, enabling faster iterations, tailored user experiences, and closer ties to platform incentives. The intention behind these assets is clear: to localize value capture, optimize for performance, and build self-sustaining economies within each L2. However, this model creates a growing disconnection across the Ethereum ecosystem. Each L2 operates with its own trust model, often secured by a multisig or custom governance, which fragments liquidity, complicates composability, and erodes the unified security assumptions that Ethereum once provided. As users and developers are forced to navigate a patchwork of isolated environments, the broader network effect of Ethereum weakens. This fragmentation opens the door for unification efforts. Native rollups with shared security (discussed later), standardized bridging protocols, or Ethereum-enshrined rollups could help preserve local autonomy while restoring global coherence. These initiatives aim to align asset issuance and transfer with Ethereum’s base-layer guarantees, offering the best of both worlds: localized innovation and ecosystem-wide trust.

Bridges Matter

As discussed, canonical bridges are foundational to Ethereum’s rollup-centric roadmap, offering strong security guarantees by relying on Ethereum for final settlement. However, the theoretical importance does not always translate into practical dominance. In practice, most users prefer external bridges, typically due to superior UX, faster finality, broader asset support, and integrated liquidity incentives. Crucially, most users may not even be aware of canonical bridges at all - they rarely surface as an option in front-end interfaces, in large part because the UX is simply not competitive. The data illustrates this divergence clearly. On Base, ~70% of bridged asset volume is routed through external options such as LayerZero, Wormhole, and others (though this percentage can change rapidly depending on actions from CEXs). Meanwhile, Arbitrum shows an even starker contrast, with a staggering 90% of flows through external bridges. This mismatch between architectural intent and user behavior highlights a revealed preference in the rollup ecosystem - convenience and composability over full Ethereum security. 

For a deeper look at Base’s bridge usage we see below that the plurality (41%) of bridged assets come from LayerZero’s Stargate bridge. 

For Arbitrum, the canonical bridge has the least volume (6%) among major bridges, although Hyperliquid’s bridge dominance understates the relative usage. 

To drive the point home, the overwhelming majority of users are bypassing the canonical bridges altogether. Less than 1% of unique addresses on both Base and Arbitrum interact with the official L1-to-L2 bridge contracts, indicating that almost all activity flows through third-party bridging solutions. When we consider that these external bridges often route high volumes through relatively few wallets, it becomes clear that most canonical bridge usage is likely driven by centralized exchanges and bridge aggregators, not retail users. This sharp divergence between volume and user count highlights a critical gap: while the canonical bridge is the only path that guarantees Ethereum-based settlement and exits, it's functionally ignored by the user base at large. The chart below visualizes just how stark this imbalance has become.

The L1 contract was intended to define the rollup, governing the canonical bridge, state transitions, node roles, and more. Given that we’ve seen that the majority of users sidestep these canonical pathways, and most assets and activity now reside outside the L1’s direct control, the original vision of the rollup as an L1-anchored extension starts to break down. This divergence raises a deeper question: what happens when applications or ecosystems no longer rely on Ethereum’s definitions at all and begin to define their own?

Two Cases for Rollup Evolution

The data reveals a disconnect between Ethereum's rollup vision and current reality. This divergence creates two distinct evolutionary paths.

Case 1: Reuniting with Ethereum Through Advanced Technology

Current Ethereum L2s represent a transitional phase rather than an end state. The proliferation of external bridges, native asset issuance, and persistent security councils reflects necessary intermediate steps made while core technologies mature. Three key innovations are converging to enable L2s to truly become extensions of Ethereum rather than adjacent systems with different trust assumptions.

Zero-knowledge technology promises to eliminate the seven-day withdrawal period that renders canonical bridges uncompetitive. By enabling instant, cryptographically verified state transitions, ZK proofs will allow canonical bridges to match the speed of third-party solutions while maintaining Ethereum's full security guarantees. This removes the primary friction point that drives the current market structure: today, users choose between Ethereum's security (via slow canonical bridges) or convenience (via fast external bridges), forcing them to abandon Ethereum as their asset's sole source of truth. 

When canonical bridges become instant through ZK verification, Ethereum can fulfill its intended role as both the issuance and settlement layer for L2 assets. Projects will be able to mint tokens on Ethereum, bridge them canonically to L2s, and maintain full L1 security throughout the asset lifecycle—eliminating the need for external issuers like Circle or bridge providers like LayerZero to create parallel security models. This technological shift fundamentally changes the economic equation: why accept external validator risk when Ethereum-native settlement offers the same speed with superior security?

Native Rollups build upon ZK technology to remove the reliance on security councils and multisigs. By integrating directly with Ethereum's execution engine through a proposed EXECUTE precompile, these rollups enable Ethereum validators to verify L2 state transitions using the same ZK proving systems that will make canonical bridges instant. This creates a unified architecture where L2s don't just settle on Ethereum—they are verified by Ethereum itself, using Ethereum's native EVM as the ultimate arbiter of correctness. 

ZK proofs provide the cryptographic guarantees that make instant verification possible, while Native Rollups embed these proofs directly into Ethereum's consensus layer, eliminating any intermediate trust assumptions. This architectural shift transforms L2s from systems that post data to Ethereum into systems that are computationally validated by Ethereum—true extensions of the base layer rather than affiliated chains with separate security models. When combined with instant ZK-verified bridges, Native Rollups deliver L2s that inherit Ethereum's full security stack without sacrificing performance.

Based sequencing completes this transformation by returning Ethereum's censorship resistance properties to L2s. Rather than relying on centralized sequencers that can censor or reorder transactions, Based rollups delegate sequencing to Ethereum's own validators and enable real-time force inclusion. This ensures L2s inherit Ethereum's complete value proposition: not just its security model but also its fundamental permissionlessness and liveness guarantees.

Together, these technologies chart a path toward L2s that achieve perfect alignment with Ethereum both architecturally and in practical usage patterns. These are active areas of development, with teams like Fabric, Succinct, Gattaca, Gelato, and many more building the infrastructure to support this transition. The goal of this effort is to restore Ethereum's role as a true security and censorship-resistance exporter for high-value, permissionless applications that require its unmatched guarantees.

Case 2: Embracing Sovereignty Through Alternative Architectures

Not every application requires or benefits from Ethereum's security model. For many use cases, the patterns we observe today, native asset issuance, external bridge usage, and isolated ecosystems, aren't problems to be solved but features that reflect genuine user preferences and application requirements. Gaming platforms with self-contained economies, social networks with native tokens, or specialized protocols may gain more from complete architectural freedom than from inheriting Ethereum's constraints.

This realization points toward Sovereign Rollups that embrace independence from the start rather than maintaining the overhead of unused Ethereum alignment. These systems recognize that for applications prioritizing performance, customization, and sovereignty over shared security, a fundamentally different architecture makes more sense.

Sovereign Rollups: An Architectural Shift

Sovereign rollups represent a fundamental departure from the traditional Ethereum L2 model. Rather than deferring to Ethereum for settlement and state finality, sovereign rollups validate and finalize their state internally. They can still use DA and consensus layers like Celestia, Ethereum or Avail, but discard the need for L1 contracts as the arbiter of truth, offering a fully self-contained execution and verification stack.

It's worth noting that while Celestia pioneered the Sovereign Rollup concept, these systems can also use Ethereum for data availability and consensus. This creates an interesting dynamic where Ethereum serves purely as a DA layer rather than a settlement layer, demonstrating that Sovereign Rollups represent an architectural choice, not allegiance to any particular ecosystem.

Advantages of Sovereign Rollups

• Performance - By eliminating the need to post data or proofs to Ethereum, sovereign rollups can unlock significantly higher throughput, making them far more suitable for high-frequency applications. Despite Ethereum raising the blob limit from three to six, its data capacity still can’t suffice to match the demand from all L2s. This is why Celestia has been taking market share away from Ethereum over time.
• Sovereignty - Sovereign rollups can evolve on their own terms. They support independent upgrades, forks, and application-specific governance, without being constrained by Ethereum’s upgrade schedules or centralized bridge multisigs.
• Modularity - Builders are free to select their preferred DA layer, bridging protocols (e.g., LayerZero, Hyperlane), and execution environments. This modular architecture aligns security and performance with each application's unique requirements.
• Cost Savings - Sovereign rollups bypass Ethereum’s gas costs, avoiding the expense of posting data or verifying fraud/ZK proofs on L1. This results in significantly lower operational costs, particularly for high-volume applications. 

• Bootstrapping - Cosmos appchains offer sovereignty, but at the cost of bootstrapping validator sets. Sovereign rollups offer similar flexibility but lower bootstrapping costs as the largest costs are limited to consensus and DA.

These benefits make sovereign rollups particularly well-suited for use cases that prioritize speed, scalability, and customization over Ethereum-native security guarantees. Applications with high throughput demands or latency sensitivity, such as gaming, social platforms, or high-frequency trading, can thrive without being bound to Ethereum’s finality constraints. Sovereign rollups are inherently well-positioned for native issuance, bypassing the need to bridge assets from Ethereum entirely. They are also powerful in ecosystems that value rapid innovation, as they enable seamless forking, upgrading, and experimentation without being restricted. Additionally, projects that require bespoke cross-chain experiences gain the flexibility to implement non-canonical bridging solutions, free from the limitations of standard L2 bridge semantics. 

As these ecosystems mature and grow their own liquidity and infrastructure, it becomes a useful thought experiment to consider whether some are trending toward functioning more as “Rollup L1s” than traditional L2s, especially if native issuance, external bridging, and asset ecosystems continue to expand independently of Ethereum. This is because, despite the “rollup” label, sovereign rollups function much more like L1s than traditional L2s. They do not inherit settlement security from Ethereum or any other base chain, instead, they define their own canonical state through their own network of full nodes. Just like an L1, a sovereign rollup maintains its own execution logic, validates its own state transitions, and achieves finality based on its chosen consensus and data availability layer. There is no external contract dictating which chain is “correct”; the source of truth resides within the rollup itself. This makes sovereign rollups fundamentally different from Ethereum L2s, which depend on smart contract verification on Ethereum for state resolution. In essence, a sovereign rollup is a new kind of L1: one that outsources consensus and data availability, but retains full autonomy over execution and settlement.

In recent years, high-profile projects have explored the rollup path to gain more control or users. For example, dYdX, the perps DEX, decided to leave Ethereum L1 and launch its L2 (then launch a Cosmos L1 for even more control). Celo transitioned to an L2 for greater distribution. Starknet, although still an Ethereum L2, has longer-term aspirations for greater sovereignty, possibly with optional Ethereum and Bitcoin settlement. The trend towards L2s of all kinds is well underway, hence RaaS providers are increasingly in demand.

Gelato

Gelato RaaS offers comprehensive support across the entire spectrum of rollup architectures, from Ethereum-aligned L2s to Sovereign Rollup L1s. On one end of the spectrum, Gelato supports leading Ethereum L2s like Ink to implement based rollup designs leveraging Gattaca, while supporting OP Succinct and preparing for full ZK rollup deployments as the technology matures. On the other end, the ABC Stack enables sovereign rollups that prioritize performance and autonomy. Between these extremes, Gelato RaaS powers various configurations of OP Stack and Arbitrum Orbit for projects utilizing Ethereum with customizable parameters. This full-spectrum approach allows developers to choose the optimal balance of security, sovereignty, and performance for their specific use cases. By extending the boundaries on both ends, from Ultrasound L2s with maximum Ethereum alignment to ABC Sovereign Rollup L1s with complete autonomy, Gelato ensures developers have access to the most suitable architecture for their applications, whether they prioritize Ethereum's security guarantees or require the flexibility of independent execution environments.

Conclusion

The rollup ecosystem has reached an inflection point where market structure must reconcile with technological reality. Current usage patterns reveal that L2s operate more as affiliated chains than true Ethereum extensions. This isn't a failure of vision but a reflection of rational market behavior given existing technological constraints.

It’s important to distinguish between L2s positioned to leverage emerging technologies—native rollups that eliminate trust assumptions, based sequencing that prevents censorship, production-ready ZK that enables instant settlement—and those whose usage patterns indicate a choice to not align with Ethereum's security model. The Ethereum Settlement Score (ESS) framework quantifies these discrepancies, revealing that major L2s achieve scores indicating minimal actual utilization of Ethereum's security—a gap between marketing and reality.

The market will bifurcate along these lines. High-value financial applications, critical infrastructure, and systems requiring credible neutrality will migrate toward architectures that truly inherit Ethereum's guarantees as based and native technology matures. Applications prioritizing performance, sovereignty, or isolated economies will formalize their existing independence through sovereign rollup designs. Both paths create value, but conflating them creates mispriced risk.

L2s preparing for native rollup upgrades and ZK integration can finally capture the premium that true Ethereum security commands—without tradeoff. Those whose value propositions depend on characteristics incompatible with Ethereum alignment should embrace sovereign architectures that optimize for their actual requirements rather than maintaining costly overhead for unused security. 

The path forward isn't about forcing all rollups into a single model but rather achieving clarity about the trade-offs and ensuring teams choose architectures that align with their actual needs. For applications requiring Ethereum's unmatched security guarantees—critical DeFi protocols, financial infrastructure, or systems demanding maximum censorship resistance—the roadmap toward Stage 4 rollups lays the path. These systems will leverage Native Rollup architecture, Based sequencing, and zero-knowledge proofs to truly become extensions of Ethereum, achieving both theoretical and practical alignment with the base layer.

As modular stacks and rollup-as-a-service providers like Gelato continue to support the full spectrum of decentralization, teams can deploy applications with their desired level of autonomy, composability, security, and cost-efficiency. This trajectory signals a broader architectural realignment, with implications for how value, risk, and security are distributed across the next generation of rollups.

 

 

 

 

The information contained in this report and by Blockworks Inc. and related affiliates is for general informational purposes only and is not intended to provide legal, financial, or investment advice. The report should not be construed as an offer or solicitation to buy or sell any security, token, or financial instrument and does not represent any recommendation or endorsement of any investment or financial product or service. Blockworks Inc. and related affiliates are not registered as a securities broker-dealer or an investment advisor in any jurisdiction or country.