The L2 debate

[ernst]

You may have seen L2 discussions picking up pace on twitter.

Many of you will remember Martin’s debate with Vitalik at Devcon in 2024 about whether Gnosis Chain should become an Ethereum L2 (here) and then his talk “Ethereum needs native L2”.

Conversations have carried on in the background, culminating in Jordi Baylina’s recent eth research post and it’s not a surprise (to us at least!) that Vitalik has been thinking deeply about Ethereum L2s (here).

There’s now an opportunity for Gnosis to work alongside Jordi on a major R&D undertaking to define a new type of L2 - one that would be synchronously composable with Ethereum and the other chains that use it and that enables dApps to benefit from shared liquidity.

This could be a game changer for Ethereum.

Gnosis has always been at the leading edge of Ethereum technology. Our engineers were the first to develop constant product AMMs, which enabled contemporary DeFi including CoW Protocol. We built the Conditional Token Framework, which is used by Polymarket. And long before ‘Account Abstraction’ became a core Ethereum roadmap priority, we developed the first production-grade implementation of smart contract wallets with Safe.

We are not new to heavy infrastructure R&D, or the business opportunity they can create. The first step in this process would be a 6 month project with Jordi to shape the concept.

Thanks to careful budgeting, the Gnosis contribution could be funded from within GIP 128. We didn’t have a dedicated R&D line item in that proposal - and probably should have done - but we can cover it. We will add R&D in future proposals.

There’s a lot that is still unknown. But being at the leading edge of Ethereum is good for us. We’ll share thoughts over the next few weeks as we learn more but we wanted to raise it for discussion here sooner rather than later.

As always, would love to hear your thoughts!

Here is an FAQ with some answers to the most obvious questions – we’ll add to it as the conversation progresses:

What would this mean for Gnosis Chain?

We don’t know yet and that decision isn’t ours to make in isolation.

But Gnosis Chain was created from a very specific thesis: that credibly neutral blockspace would be in demand. In practice, that has not played out the way we expected. Blockspace has largely been commoditized.

Meanwhile, Ethereum L2s are siloed and do not meaningfully leverage Ethereum’s shared security or liquidity. This does not have to be the case.

This R&D project would enable us to explore many revenue-generating opportunities for Gnosis and Gnosis Chain.

It also opens possibilities for apps that can unlock true mainstream adoption. Apps that inherit the full security of Ethereum, that tap into shared liquidity, and build with synchronous composability with everything on Ethereum, including apps like Gnosis Pay and the new Gnosis consumer money app.

Today, simply maintaining table stakes is a disproportionate amount of work. Running a standalone domain means constantly rebuilding and sustaining a full stack of DeFi primitives, bootstrapping liquidity, maintaining bridges, operating on- and off-ramps, solving compliance and monitoring requirements, and ensuring that core infrastructure stays competitive. It often feels like swimming against the current just to remain relevant.

Synchronous interoperability changes that dynamic. Every project that operates inside a truly shared, atomically composable Ethereum domain, no longer needs to replicate the entire ecosystem just to participate in it. They can rely on shared liquidity, canonical infrastructure and common security guarantees instead of duplicating them. That frees up capital, engineering bandwidth, and strategic focus.

Instead of defending parity, networks can invest in differentiation: building better user experiences, real-world integrations, consumer-grade financial apps, and purpose-built products like Gnosis Pay and the Gnosis consumer money app. Rather than fragmenting liquidity and attention, we contribute to a single economic zone and grow usage by expanding the pie — strengthening Ethereum while building a distinct and valuable ecosystem on top.

Exactly what this could look like will become clear over the coming months but Gnosis can define it and thereby ensure that we are positioned to benefit.

What would this mean for GNO?

First, there are some key things that will not change.

GNO is the governance token of Gnosis DAO and the Gnosis ecosystem is designed so that value accrues to the GNO token.

The current GNO model for Gnosis Chain is not effective.The DAO pays out staking rewards for Gnosis Chain from the treasury. This staking yield is not from transactions, it is paid for by the DAO and simply means that those who stake GNO don’t get diluted.

This new avenue could open many new revenue options. You could have a dual token model where part of the tx fee is paid for in ETH (because you need to pay L1 tx) and part is paid in another token (that the L1 validators accept for sequencing the L2), or the GNO token being leveraged in preconf schemes like the one suggested by Vitalik (here).

What would this mean for Validators?

This question depends entirely on future decisions around Gnosis Chain and GNO.

We estimate that this initial R&D effort would take 6 months and we will have far greater clarity on our options by that time. It would be at least 9-12 months after that before the framework we would explore with Jordi reaches mainnet. There will be plenty of time for a full and open discussion.

[clement]

Self quoting part of my post Gnosis Chain is Dying – Are We Going to Do Something About It? - #6 by clement

@clement

Drawbacks

Not being a L2 (or in process of becoming a L2) did harm the attractiveness of Gnosis Chain. Indeed, for the deepfunding experiment with a 50k$ grant of the EF, we were asked to do it on a L2 and we deployed on Optimism for this purpose (if Gnosis Chain was a L2, this business would have been on Gnosis).
Those days, having a bridge to Ethereum mainnet is not sufficient to be Ethereum aligned.

[e3o8o]

I read this with excitement. But I am not sure a lot of people grasps how significant this move is. So I will share my thoughts to help glean why I think this is the right move, and critically, the right response to the structural issues @N0xyGen and others have been raising.

This R&D effort is a deep architectural paradigm shift that could solve the structural problem beneath the “dying chain” symptoms (empty blocks, fragile DeFi dependencies, reliance on treasury subsidies to maintain Aave/RealT). Gnosis leads the charge when it comes to DeFi execution and innovation but the cost of ecosystem replication is brutal. Every isolated domain pays a bootstrapping tax and still, can’t out-spend mainnet’s liquidity gravity. Synchronous composability treats the disease where mercenary incentive program puts a band-aid on the symptoms.

Here is what i find most compelling compared to other L2 pivots:

1. It solves Gnosis’s specific existential problem: Synchronous composability means Gnosis doesn’t need to subsidize its own isolated Aave deployment or bootstrap a parallel DeFi ecosystem from scratch. It can compose directly with Aave on L1. This almost completely offloads the infrastructural burden that slowly kills most chains.

2. It turns Gnosis Pay into an anchor: The concern that a multichain Gnosis Pay would drain transactions from Gnosis Chain is valid. But if Gnosis Chain is synchronously composable with mainnet, Gnosis Pay running natively on Gnosis gets Ethereum liquidity by default. It removes the bridge or fragmentation worries and the multichain problem dissolves because Gnosis is no longer an island.

3. A genuine path to GNO value accrual: The current model (treasury-funded staking rewards) is replaced by a model where GNO finds genuine economic function emerging naturally from Jordi’s architecture, whether as sequencer collateral, preconf staking, or capturing fees from the cross-domain execution premium.

Technical Paradigm Shift

What makes Jordi’s execution table model architecture and the native rollup precompile, genuinely different is the atomicity guarantee. Bridges give us eventual consistency. Message-passing gives us asynchronous composability. But the ability to read and write state across domains within a single transaction changes what kinds of applications can exist. Any app that needs to reference state on multiple chains simultaneously (cross-chain liquidations, multi-venue settlement) currently has to fragment or centralize as they currently cannot operate atomically. Even state-of-the-art omnichain architectures like LayerZero OApps are fundamentally asynchronous. They require developers to fragment execution and build complex fallback logic because if the destination transaction fails, the source transaction has already finalized and cannot revert. Synchronous composability eliminates that forced choice, if one side of the cross-domain call fails, the entire transaction atomically reverts, exactly as it would on a single L1.

As this is pushed forward, there are two sharp R&D questions worth exploring in the 6-month sprint:

• Sequencer Incentive Alignment & DOS Vectors: In a based rollup model where L1 validators sequence L2 blocks, how do you prevent DOS or MEV centralization from proposers who have a head start on proof computation? As cross-domain MEV becomes the most valuable flow, keeping builder/sequencer incentives aligned will be uncharted territory
• Failure Mode Granularity: As Jordi noted in the linked post, if a rollup state update is skipped, the whole transaction fails. For a simple swap, all-or-nothing atomicity is fine. But for complex operations spanning 3+ domains, the blast radius of a single rollup hiccup is massive. Can we design graceful degradation into the execution table model, or is strict atomicity the only safe path?

Historical Risk

The honest strategic risk worth naming is historical: Gnosis’s track record of building brilliant public-good primitives (Safe, CoW, Conditional Tokens, and even pioneering the charge towards a crisis response framework - A Framework for the Future) that other ecosystems can leverage or capture the value from is both its greatest strength and a vulnerability.

Already noting that “Gnosis can define it and thereby ensure that we are positioned to benefit” is the right instinct. But defining the standard may not be enough. If this R&D produces a generalized synchronous composability framework, the value-capture mechanism needs to be designed deeply into the architecture, not bolted on after. Gnosis must be the canonical, highly-defensible implementation.

I remain excited to see where this goes as this is exactly the leadership this community and the entire crypto ecosystem needs.

[N0xyGen]

I start with the basic economic question that every GNO holder will and should ask : if the boundaries between chains blur, why is a dual-token use case still necessary? Currently, value accumulation for GNO holders is purely theoretical. Many projects have been spun off into separate entities, leaving GNO holders with influence only through treasury funding control.

If this transition fails to establish a clear path to value accumulation for GNO, or if its utility for validators is undermined during the shift, it will lead to further value destruction of the token. This is also a systemic risk, as the Treasury itself consists of around 75% GNO. This scenario isn’t speculative; one only needs to look at the Arbitrum chart to see the consequences of a “governance-only” token.

Regarding failed project growth
Furthermore, we must address why Gnosis has struggled to attract and support projects effectively until now. If blockspace is a commodity, it should theoretically be irrelevant which chain a project chooses. The fact that projects aren’t choosing Gnosis suggests the ecosystem is not attractive enough.

In my view: A core part of Treasury management should be managing project growth on the chain, rather than just “looting” or managing the Treasury’s own funds. If the Treasury is not used as an engine for ecosystem growth but merely as a capital reserve, the transition to an L2 won’t fix the underlying lack of adoption. (And to be clear this job doesn’t have to be only handing out mercenary capital)

So main question: Why have projects preferred to go on Base or Arbitrum so far? Because of the “developer experience”? Or the user base. Do you think the technical hurdle (composability) was really the main problem, or was it more the marketing and business infrastructure of the DAO?

[hhh]

as a user its super difficult to get funds on and off from gnosis without gnosis pay. gnosis pay is how users enter gnosis.

i think the most important thing is product and gnosis pay is a good product. the only reason why i like gnosis is because of the products mainly gnosis pay.

the chain is a tool to offer good products. Gnosis should focus 100% on cypherpunk product.

how to get value to GNO? by making gno holders owners of the product

gnosis can bring to ethereum the product first approach by holding up cypherpunk values

[ernst]

A brief update as we reach a milestone this week.

Since posting this in February, we’ve been heads down with Jordi and the team on the technical foundations. This week at EthCC in Cannes, we publicly introduced the Ethereum Economic Zone (EEZ): the synchronously composable L2 framework that this R&D effort will develop. The Ethereum Foundation is co-funding the work, all major block builders will support the execution standard from day 1, and many ecosystem players will leverage the EEZ from launch: CoW swap and Safe ofc, but also Aave, spark, monerium, xstocks, centrifuge and many others.

You can watch the talk [with demo and application examples!] here on YT. The website is live at eez.io and this is a relevant blog post.

I also want to acknowledge the questions raised in this thread directly, because they’re the right questions.

• On GNO and the validator set (N0xyGen, hhh): I hear you. The reference implementation of the EEZ rollup will use ETH as gas. But it is already clear to us that there could be compelling opportunities for our existing validator set and for the GNO token in our own imlpementation, should the DAO decide to go that route. This is not something we will decide unilaterally. It will be defined with the community over the coming months.

• On value capture (e3o8o): Your point about Gnosis’s historical pattern of creating public goods that others leverage is fair. We are thinking carefully about this. EEZ as shared Ethereum infrastructure and Gnosis’s strategic position within it are two different conversations, and we intend to have both openly.

• On products over infrastructure (hhh): Agreed. The chain is a tool to offer good products. That is exactly why synchronous composability matters to us. Instead of spending engineering effort defending parity with other ecosystems, we can focus on differentiation: Gnosis Pay, the Gnosis App, consumer-grade financial products. Shared liquidity and canonical infrastructure free us to build what actually matters.

• On why this and not better marketing (N0xyGen): Fair challenge. The honest answer is that we need both. But no amount of marketing solves the structural problem of maintaining a standalone domain in a commoditized blockspace market. This is an architectural decision, not a communicaitons one.

We will share technical specs, benchmarks, and integration details in the coming weeks. In the meantime, I look forward to the continued discussion here and at the upcoming community AMA on April 15th.

[Sora]

Deeply appreciate the detailed update, @ernst

The shift toward synchronous composability as an architectural decision (rather than just a marketing one) is exactly what caught my attention. The ‘islands of liquidity’ problem has been the biggest bottleneck for genuine Web3 adoption.

Seeing major players like Safe, CoW Swap, and Aave supporting the execution standard from Day 1 gives a lot of confidence. It moves the conversation from ‘Which chain is better?’ to ‘How can we compose the best user experience across the entire ecosystem?’

I’m particularly interested in how the deploy-once, compose-everywhere paradigm will simplify the overhead for builders. Looking forward to the technical specs and the AMA on April 15th!