This article was originally published on X by luis_0x.

Preamble

Hey everyone, I’m Luis - founder at Gelato, the complete Web3 Cloud platform for Rollup-as-a-Service and onchain apps. We’ve been deep in rollup tech for over two years, launching RaaS in 2023 and going live with the first Polygon zkEVM CDK chain in production (Astar), followed by OP Stack and Arbitrum Orbit L2s, like Kraken's Ink and Animoca's OpenCampus. Along the way, I came across the concept of Sovereign Rollups through our early research and Celestia’s work, but it only really clicked for me after revisiting the topic in early 2024. A few deep chats with Mustafa (Celestia) later, I was fully Sovereign Rollup-pilled - especially around his idea of a pessimistic rollup.

From running 50+ Ethereum L2s in production, it became clear: most users weren’t using the canonical L2 bridges or fraud/zk proofs - they relied on third-party bridges like LayerZero or Across. That meant they were inheriting the downsides of L2s (cost, latency) without the intended upside (Ethereum-grade security). Sovereign Rollups sidestep all that. No native L2 bridge, no unnecessary complexity - just pure rollup logic, with decentralized consensus and DA outsourced to Celestia, which is built exactly for that.

That’s why we’re building ABC Stack - the first pessimistic, Celestia-native Sovereign Rollup stack for the EVM. Think of it as a “Rollup L1”: your own chain, but without needing a validator set or a new token. You get the flexibility of an L1, the performance and consensus of Celestia’s DA layer, and the simplicity of no bridge assumptions - just like spinning up infra on AWS, but decentralized and verifiable by default.

If your rollup use case relies heavily on the L2 bridge to Ethereum, ABC likely isn’t the right fit - it’s explicitly not an Ethereum Rollup L2 stack, but a Celestia Rollup L1 stack. While I think L2 bridges today have serious issues around trust and usability, I deeply respect the Ethereum rollup teams pushing toward trustless ZK proofs and look forward to a future where Ethereum Native Rollups eliminate L2 bridge multisigs entirely.

Now let’s dig in and make things spicy… 🌶️

Key Takeaways

Warning: You’re about to be Sovereign Rollup-pilled💊. After reading this article, you’ll realise:

• Everything you thought you knew about rollups and L2s is a lie. The dominant narrative has been shaped by Ethereum maxis, and reality has already moved beyond this framework.
• Sovereign Rollups aren't L2s—they're “Rollup L1s.” This isn't just semantics. It's a clearer view of blockchain architecture that aligns with how users actually behave.
• Users have already voted with their wallets. Canonical bridges are barely used compared to external bridges. It's already happening, with users choosing speed over theoretical security benefits that are, frankly, over-marketed and not as secure as portrayed.
• The asset is what matters, not the chain. Assets have security models independent of the chains they exist on. Instead of asking "which chain does this bridge serve?" we must ask "which asset am I moving and what's its true source of issuance and ultimate ledger?"
• Native asset issuance on rollups is the future. Most assets on rollups today aren’t bridged from Ethereum; they’re either rollup-native or multichain tokens. The L2 narrative simply doesn't match reality anymore.
• You think you're using a canonical asset, but you're probably not. Even when you bridge canonical ETH through services like Across, while that asset is in transit, you're actually using an abstraction with its own security assumptions completely separate from Ethereum.
• Celestia is rollups reimagined. They cracked the Modular Blockchain and Sovereign Rollups code. ABC is championing Celestia's vision to deliver true rollup sovereignty at scale.
• Modularity isn't optional. The ability to choose security models, bridging solutions, and execution environments based on specific needs is how we'll scale to billions of users.
• Performance matters more than the L2 bridge. The ABC Stack achieves gigagas-level throughput precisely because it doesn't carry the baggage of enshrined bridges and predetermined security models. Users don’t care about your infrastructure - they want better applications.
• Running your own full node is cool again. Unlike L2s where you still depend on a smart contract light client for state finalization, with a Sovereign Rollup you can run a full node and directly reject invalid state and maintain true decentralization because your node is the source of truth for state, not a remote contract.
• Fraud proofs are more ceremony than substance. In the future, we'll look back at today's fraud proof implementations as purely a marketing tool for decentralization larping.
• Stop waiting for ZK to magically solve everything. ZK marketers will tell you that "instant, affordable ZK proving is just round the corner", but it's not. ZK technology is promising but will introduce its own challenges. The ABC Stack is built to work today while remaining forward-compatible with ZK when it matures.
• The multisig reality. Let's be honest - canonical bridges of major L2s are secured by multisigs controlling the upgrade keys. Base, Optimism, and Arbitrum all rely on multisigs that could, in theory, change the bridge contracts at any time. Fraud and ZK proofs provide little protection against this centralized control point that will likely remain even after "Stage 2" implementations. Ethereum "Native Rollups" might be the solution to this issue - a topic that I will address in future musings.

The "L2s Inherit Ethereum's Security" Myth and How Users Actually Behave

Metrics tell a story that the Ethereum maximalists don't want to hear:

Optimism, a poster child of Ethereum L2s, had just 404 unique users utilizing its canonical bridge last week.

Meanwhile, LayerZero alone had approximately 25,000 unique users on Optimism during the same period. That's a 60x difference for just one external bridge provider. Add in the others and we find that 90% of users are using external bridges rather than canonical bridges.

This is an overwhelming market verdict. Users are choosing speed and cost over theoretical security guarantees that matter only to a handful of academics and Ethereum purists.

The story gets even more compelling when we look at Total Value Secured (TVS) – the actual assets residing on leading L2s:

The combined value of external assets (using an external blockchain as their main ledger) and native assets (using the L2 rollup as their main ledger) exceeds the value of canonical assets (using Ethereum as their main ledger).

Think about that for a second. Most of the value on "Ethereum L2s" isn't even secured by their Canonical L2 Bridge on Ethereum. It's either created directly on the rollup or bridged in from Ethereum, or other chains, via third-party bridges.

But here's the real kicker: even a significant chunk of what gets counted as "canonical assets" isn't wholly secured by Ethereum in practice. When you use Across to bridge ETH to an L2, you're not using Ethereum's security model – while that token is in transit, it’s secured by Across - you're using Across's intents-based architecture with relayers and optimistic execution.

These intents-based bridging solutions deliver transactions in seconds instead of the days required by canonical fraud proof systems. They have different security assumptions, and users have overwhelmingly shown they prefer this tradeoff. Speed and usability win every time.

If less than half of your L2’s Total Value Secured comes through the canonical Ethereum bridge, can you genuinely claim to be “secured by Ethereum,” or even call yourself a true L2?

The Assets You Use Don't Care About Your L2 Ideology

If you create a token on Base or mint a Base NFT, Ethereum has no idea it exists. The liquidity and the asset itself exist only on Base. Ethereum can't bail you out because it doesn't even know what happened.

This is the big disconnect in the L2 narrative: the assumption that Ethereum settlement provides meaningful security for all assets on an L2 is simply false for most assets on most L2s today.

Consider the security model of these different asset types:

When external assets like USDC use bridges like Circle's CCTP, they're secured by Circle's infrastructure, not Ethereum's consensus. For these assets, the rollup is more accurately described as a "Circle rollup" than an "Ethereum rollup."

The reality is that assets have their own security models that are independent of – and often completely unrelated to – the chain they reside on. Security follows the asset issuer (bridge), not the chain.

Slow Canonical Bridges Were Doomed From the Start

The early architects of L2 rollups were mistaken when they believed that users would prioritize absolute security over user experience. They assumed users would gladly wait 7 days to withdraw funds if it meant preserving a direct security connection to Ethereum.

They were wrong.

In the real world, most users would rather pay slightly more and wait 30 seconds using a popular external bridge over a theoretically more secure but painfully slow canonical bridge.

This behavior is a perfectly rational prioritization of capital efficiency and sane user experience. A 7-day lockup of capital represents a massive opportunity cost that far outweighs the marginal security benefits for most users.

And truthfully, the security tradeoffs aren't even that significant for many use cases. For moving USDC between chains, Circle's CCTP is more appropriate than an Ethereum fraud proof system, because Circle is the issuer of the very token in question.

The Sovereign Solution

This is where the ABC Stack and Sovereign Rollups enter the picture. Instead of fighting against market realities, we're building on architecture that embraces them.

Sovereign rollups were pioneered by Celestia and its co-founder Mustafa Al-Bassam in "Rollups as Sovereign Chains" back in 2022. ABC is proudly building on and championing this innovation.Sovereign rollups – or Rollup L1s as we call them in the ABC Stack – eliminate the native bridge entirely. They handle settlement themselves, just like an L1, while leveraging Celestia for data availability and consensus.

This simple architectural shift produces massive benefits. By eliminating bridge overhead, ABC Stack achieves gigagas throughput – orders of magnitude faster than traditional L2s. Rollup developers can choose the right bridge for each asset instead of being locked into a one-size-fits-all approach. When ZK technology matures, you can integrate it – no need to wait or rebuild.

How Sovereign Rollups Actually Work

Traditional L2 rollups force a rigid architecture: handle execution on your own, but send transaction data to Ethereum as a settlement layer, data availability layer, and consensus layer. Each L2 must deploy a smart contract on Ethereum that acts as the single source of truth for determining the canonical chain state.

Sovereign rollups break this model by eliminating the mandatory settlement relationship with Ethereum. In the ABC Stack, a Rollup L1 is its own settlement layer—like any other L1 blockchain—while using Celestia purely for consensus and data availability.

This distinction is crucial. In a traditional L2, the "correct" chain is defined by a smart contract on Ethereum. In a Rollup L1, the correct chain is defined by the rollup's own network of nodes—just like Bitcoin, Ethereum, or any other L1.

Rollup L1s verify and settle transactions through their own dedicated full nodes, functioning off-chain without depending on a remote settlement layer. They maintain independent state and consensus frameworks and achieve finality within their off-chain node network.

The assumptions are fundamentally different:

• L2 assumption: Either I or someone else will maintain the integrity of the external smart contract light client that we use as the final single source of truth.
• Sovereign Rollup assumption: Every honest full node runner fully verifies the chain and will reject an invalid state update. Celestia is used for consensus finality guarantees, so all honest nodes can know the final truth by following Celestia and computing their own local blockchain state from there.

What about security? Celestia provides the consensus and data availability guarantees that traditional L1s must bootstrap themselves through validator sets. Your Rollup L1 inherits Celestia's security while maintaining sovereignty over its own state.

This combines the best aspects of standalone L1 chains (sovereignty, flexibility) with the best aspects of rollups (shared security, minimal overhead).

The True Use Cases for Rollup L1s on Celestia

Let's cut to the chase and talk about who actually needs Sovereign Rollups today.

Native Token Ecosystems

Applications like PumpFun that generate thousands of native assets (memecoins) would be strangled by an L2 architecture that assumes all important assets come from Ethereum. A Rollup L1 can easily power these token-abundant ecosystems without unnecessary overhead.

Gaming Applications

Gaming requires both high performance and would function perfectly fine in economic isolation. Most in-game currencies exist solely within the game's ecosystem—they don't need to bridge to Ethereum or anywhere else.

Brand-New Ecosystems

New application ecosystems with their own tokens, NFTs, and DeFi primitives need the freedom to evolve on their own terms. The L2 "settlement" relationship with Ethereum adds no value while imposing significant constraints.

Performance-Critical Applications

Applications that require maximum throughput are handicapped by the overhead of maintaining L2 bridge infrastructure. By eliminating this unnecessary overhead, ABC Stack delivers gigagas-level performance out of the box.

L2s vs. Rollup L1s

What About Cosmos?

Cosmos pioneered sovereign appchains, but they aren't rollups—they're independent L1s with their own validator sets. This creates massive overhead, since every new chain must bootstrap its own validator set from scratch. This is both impractical and insecure for new projects.

Rollup L1s deliver the sovereignty of Cosmos appchains without the validator overhead. By sharing Celestia's security layer, even brand-new applications can launch with full security and censorship resistance, afforded by Celestia’s decentralization, from day one.

Cosmos's IBC remains a breakthrough in cross-chain communication, but its implementation complexity has largely confined it to the Cosmos ecosystem. Rollup L1s are not limited to any specific bridging protocol—they can integrate with LayerZero, Hyperlane, or any other bridging solution based on specific needs.

ZK Reality Check

ZK technology promises trust-minimized bridging with fast finality. But production-ready ZK with low latency and low cost is still years away, despite what ZK marketers will tell you. The current state of ZK rollups proves this point—they're either specialized for specific use cases or struggling with compatibility and performance.

Even when ZK matures, not every application will need or want it. Many will prefer other bridges or no bridges at all. And for those that do want ZK, ABC Stack’s modularity provides a clear upgrade path: run a standard ABC rollup now, then integrate ZK solutions, like Celestia’s coming ZK-based interop, when they're ready for prime time.

There Is No Single "Best" Chain Architecture

The whole point of modularity is to encompass the reality that the future blockchain ecosystem won't converge on a single chain type. We'll see a diverse landscape:

• Optimistic Ethereum L2s will continue serving Ethereum-centric use cases, but likely all ditch fraud proofs for ZK proofs once the tech is ready
• ABC Sovereign Rollup L1s will encompass all non-Ethereum-specific use cases that still want the groundbreaking properties of a global, verifiable, permissionless and censorship-resistant Web powered by Celestia
• Non-rollup L1 stacks like Avalanche or Cosmos will serve large market segments like enterprise blockchain that do not require a high degree of decentralisation
• A handful of monolithic L1s like Solana will remain, as they have reached escape velocity

Different applications have different requirements, and forcing them all into the same architectural mold is counterproductive.

Sovereign rollups will power the substantial portion of blockchain applications that need:

• Complete control over their execution environment
• Maximum performance without settlement overhead
• Freedom to issue native assets without constraints
• Flexibility to choose bridging solutions based on specific needs

Traditional L2s simply cannot meet these requirements without fundamental redesign—at which point they'd essentially become sovereign rollups anyway.

Ethereum L2s do serve a great purpose within the Ethereum-centric DeFi ecosystem. Once their bridges are ZK-based and the multisigs are gone, thanks to the Native Rollup architecture, they'll fulfill their original vision. So if you truly care about ETH the asset and want to build an application specifically around it, the Ethereum Rollup-centric roadmap provides a path forward for you.

Free Your Mind

The ABC Stack isn't building L2s—it's building Rollup L1s. Sovereign rollups represent the logical evolution of blockchain architecture. By maintaining settlement sovereignty while inheriting Celestia's security guarantees, they deliver the performance, flexibility, and sovereignty that next-generation applications demand.

It's a shift that aligns blockchain design with how users and developers actually behave in the real world. Users prefer fast, flexible bridges over canonical ones. Most assets on "Ethereum L2s" aren't even secured by Ethereum in any meaningful sense. And the theoretical security benefits of L2 settlement simply don't justify the performance and flexibility costs for many applications.

You've been living in a dreamworld, Neo. A world built on L2 bridges that nobody uses. I'm trying to free your mind, but I can only show you the door. You're the one that has to walk through it.

You take the blue pill—the story ends, you wake up in your bed and believe whatever you want to believe about L2 rollups. You take the red pill—you stay in the real world of Abundance, and I show you how deep the Sovereign Rollup rabbit hole goes.

FIN