The world of Ethereum's Layer 2 (L2) scaling solutions is undergoing a significant transformation. What was once a clear path, driven by the urgency of high transaction fees in 2020, has become a complex landscape. The original vision of L2s as 'branded shards' of Ethereum, primarily focused on offering cheaper transactions, is now being critically re-evaluated. This shift is largely due to two unforeseen realities: Ethereum's mainnet has become substantially cheaper than expected, and the decentralization of many prominent rollups has progressed slower than initially promised.
Vitalik Buterin, Ethereum's co-founder, has candidly addressed this evolution in a recent post on Ethereum Research. His message is clear: the ecosystem needs a new direction, moving beyond the simplistic notion that all L2s serve the same purpose. This isn't an abandonment of L2s; rather, it's a necessary recalibration of expectations and a sharper definition of what different types of rollups are truly aiming to build. The core question now is, what is the new job description for L2s, given that the foundational premise of the rollup-centric roadmap has significantly weakened?
The Elusive Promise of Stage 2 Decentralization
To understand the current state of L2 decentralization, L2BEAT's Stages system offers a valuable framework. This system categorizes rollups based on their progress towards full decentralization and security inheritance from Ethereum:
- Stage 0: These rollups still operate with significant 'training wheels' in place, meaning they retain meaningful trust assumptions. Critical safety properties are often managed by a centralized entity or a small group.
- Stage 1: Representing partial decentralization, Stage 1 rollups offer stronger escape hatches for users and more robust proof guarantees. However, they can still have significant trust assumptions related to upgrades or governance mechanisms.
- Stage 2: This is the 'no training wheels' milestone. At Stage 2, critical safety properties are enforced by code rather than relying on discretionary actors. This stage signifies a high degree of trustlessness and security inheritance from Ethereum.
The current distribution of value across the L2 ecosystem highlights a stark reality. According to L2BEAT's rollup scaling summary, a staggering 91.5% of the total value secured on L2s resides in Stage 1 rollups. Approximately 8.5% is in Stage 0, and a mere 0.01% has achieved Stage 2. The top three rollups alone account for roughly 71% of this total value, indicating that progress towards Stage 2 largely hinges on the decisions and developments of these major projects, rather than on smaller, experimental chains.
A primary hurdle preventing further decentralization is the ability to override proof systems and the presence of delays or constraints on upgrades. Upgrade discretion remains a common feature among the largest rollups, and moving beyond this has proven far slower and more challenging than the optimistic forecasts of 2020-2021 suggested. Some projects have even explicitly stated their intention not to progress beyond Stage 1, citing not only technical limitations related to zkEVM safety but also regulatory requirements that necessitate a degree of centralized control. While this can be a legitimate product decision for certain customer bases, it fundamentally clarifies that such chains are not 'scaling Ethereum' in the manner originally envisioned by the rollup-centric roadmap.
Why the Core Assumptions Changed
The original 'rollup-centric Ethereum roadmap,' proposed in October 2020, emerged from a very different environment. Gas prices were soaring, some applications struggled to function, and the consensus was that the ecosystem needed to go 'all-in on rollups' for the foreseeable future. The plan was for the base layer to prioritize data capacity for rollups, with users increasingly living on L2s.
However, two crucial facts have shifted dramatically since then:
- Ethereum Mainnet is Substantially Cheaper: Today, the L1 is much more affordable. Etherscan data shows a seven-day average transaction fee hovering around $0.35, with gas snapshots often in fractions of a gwei. On January 16, Ethereum even recorded an all-time high of 2,885,524 transactions in a single day. This creates a narrative of 'busier and cheaper,' a direct contradiction to the 2020 crisis that initially spurred the rollup-centric vision. While cheaper fees are generally positive for user adoption and activity, they can also be exploited. For instance, the ecosystem has seen record growth, but malicious 'address poisoning' campaigns have unfortunately leveraged these low fees to drive activity, highlighting a darker side to the increased affordability.
- L1 Execution Capacity is Rising: Ethereum's block gas limit was increased to approximately 60 million after broad validator signaling, up from the long-standing 30 million limit. With roughly 12-second blocks, 60 million gas translates to about 5 million gas per second. Community discussions even entertain aspirational targets as high as 180 million gas, which would represent a threefold increase, though this remains a directional goal rather than a committed change.
The clear interpretation of these shifts is that the 2020 premise, that 'L1 can't scale for most users,' is significantly weaker in today's fee environment. This new reality creates space for L2s to be a diverse spectrum of security and sovereignty trade-offs, rather than merely being near-identical 'shards' that compete solely on price.
L2s as a Spectrum, Not Clones
Buterin's proposed reframing treats L2s as occupying a full spectrum of functionality and security. At one end are chains that benefit from the full faith and credit of Ethereum's security, offering unique properties beyond just being an EVM clone. These could include privacy-focused systems, non-EVM execution environments, or ultra-low-latency sequencers. At the other end are options with varying degrees of Ethereum connectivity, allowing users and applications to choose based on their specific needs and desired trust models.
The new minimum bar for L2s is straightforward: if you handle ETH or Ethereum-issued assets, you must achieve at least Stage 1 decentralization. Otherwise, Buterin suggests, you are essentially a separate Layer 1 blockchain with a bridge, and should present yourself as such. The differentiation bar is more challenging: be the best at something other than simply 'cheap EVM.'
Examples of specialization Buterin cites include:
- Enhanced privacy features.
- Efficiency tailored to a specific application.
- Truly extreme scaling capabilities that surpass even an expanded L1.
- Fundamentally different designs for non-financial applications, such as social or identity systems.
- Ultra-low-latency sequencing for high-speed use cases.
- Features like built-in oracles or decentralized dispute resolution that might not be computationally verifiable on L1.
Mechanisms to facilitate this specialization are still under investigation. A 'native rollup precompile,' for instance, could enable Ethereum to verify a standard zkEVM proof directly within the protocol. For rollups that are 'EVM plus extras,' this would mean the canonical EVM verification occurs trustlessly at the protocol level, and the rollup only needs to prove its custom extensions separately. This could foster stronger interoperability and potentially pave the way for synchronous composability, where contracts across different rollups can interact within the same transaction. However, these remain research trajectories, not yet deployed features.
Three Emerging L2 Categories
If this reframing takes root, we can expect rollups to naturally fall into clearer categories:
- Stage 2-Chasing Settlement Rollups: These projects are dedicated to maximizing their inheritance of Ethereum's security. Their core mandate is to achieve code-enforced guarantees with minimal discretionary governance, truly acting as 'scaling Ethereum' solutions.
-
Regulated or Controlled Execution Environments: This category optimizes for compliance, permissioning, or specific institutional requirements. These rollups may never progress beyond Stage 1 by design, and their marketing should honestly present this control as a feature, rather than pretending to offer full decentralization. For institutional players seeking compliance and predictable environments, this approach holds significant appeal.
- Specialized Chains: These are optimized for specific attributes such as latency, privacy, app-specific execution, or non-financial use cases. Examples include privacy rollups using zero-knowledge proofs to obscure transaction details, ultra-low-latency sequencers for high-frequency trading, or social and identity systems built with fundamentally different state models. These don't necessarily need to be EVM-compatible or even financial in nature; their justification lies in providing unique value that users cannot obtain elsewhere.
Major projects like Arbitrum One, Optimism, Base, zkSync Era, and Starknet will each need to strategically decide which of these categories they are pursuing. The ecosystem is robust enough to support all three, but the long-held assumption that every L2 performs an identical function is rapidly fading.
What This Means for Users and Builders
For users, the burden of understanding shifts significantly. They will need to grasp the specific guarantees offered by each L2. Features like escape hatches, upgrade delays, proof systems, and censorship resistance will become crucial product differentiators rather than assumed properties. Wallets and user interfaces will need to label trust assumptions more explicitly, and frameworks like L2BEAT's Stages will become even more vital in making these assumptions legible to the average user.
For builders, the era of 'cheap EVM' as a primary differentiator is largely over; it has become a commoditized offering. Differentiation will now hinge on specializing in areas like privacy, custom virtual machines, ultra-low-latency sequencing, application-specific throughput optimizations, non-financial applications (such as in social, identity, or AI contexts), or explicitly offering compliance and permissioning as a product feature, without claiming it as 'Ethereum scaling.'
In the broader market narrative, expect a more intense debate about whether L2s truly 'inherit Ethereum security' in practice, rather than merely as an aspiration. The fact that many large rollups remain at Stage 1 with discretionary governance provides rival L1 proponents with more compelling talking points.
The New L2 Reality: Re-tiering, Not Revolution
Ethereum is unlikely to experience a sudden L2 revolution. Instead, it is undergoing a profound 're-tiering.' The original rollup-centric roadmap was built on the premise that L2s would be almost identical 'branded shards' competing primarily on cost, while L1 remained expensive and capacity-constrained. This premise no longer holds true. The L1 is now cheaper and actively expanding its capacity, while L2s are diverging rapidly in their security models and use cases, despite the slow progress towards Stage 2 decentralization.
The job description for Layer 2s has fundamentally changed. The new minimum bar is to offer credible security when handling Ethereum assets. Beyond that, the key to success lies in specialization and being transparent about the underlying trust model.
Native verification primitives and ongoing research into synchronous composability signal where Ethereum aims to facilitate this easier, more interconnected future. However, these are still trajectories, not deployed features. The rollup-centric roadmap has been upgraded to accommodate the reality of a scaling L1 and an L2 ecosystem far more diverse and specialized than originally anticipated.
Post a Comment