Five years is a long time in Ethereum. The rollup-centric roadmap proposed in 2021 reshaped how the entire ecosystem thought about scaling: Layer 2s were the answer to Ethereum's capacity problem & the job of L2s was essentially to take pressure off the L1 while keeping users within the same trust environment. That was the right framing then. It is no longer the complete picture, & the Ethereum Foundation has now said so officially.

On March 23, 2026, the EF's Platform team published a blog post titled How L1 & L2s can build the strongest possible Ethereum, authored by Josh Rudolf, Julian Ma & Josh Stark. It is the most direct statement the foundation has made about the evolving relationship between the two layers since the rollup-centric roadmap itself.

• What Has Actually Changed
• What L1 Becomes in This Model
• What the EF Is Asking of L2s
• What the EF Is Committing to in Return

What Has Actually Changed

The original rollup-centric roadmap gave L2s one primary job: scale Ethereum. Everything else was secondary. The model worked well enough as a starting point, but the ecosystem that grew out of it looks very different from what was imagined in 2021.

1/ How L1 & L2s can build the strongest possible Ethereum

tldr: we should continue to lean into the unique capabilities of each layer, & make sure all users have a clear path to securely & seamlessly benefit from the core properties of Ethereum pic.twitter.com/jUPNscgSix— joshrudolf.eth (@rudolf6_) March 23, 2026

L2s like Arbitrum, Base, OP Mainnet, Starknet, Linea are no longer just scaling mechanisms. These are ecosystems in their own right, with their own communities, developer tooling, GTM strategies & economic activity. Some have built features that Ethereum L1 simply cannot offer: ultra-low latency, application-specific sequencing, additional privacy guarantees, compliance-oriented architecture. The value they provide to users is no longer primarily about cheaper transactions it is about capabilities that a single chain cannot provide regardless of how much it scales.

The EF's post acknowledges this shift directly. The primary objective of L2s has moved from scaling to differentiation, control & innovation. Scaling is still valuable but it is no longer the biggest lever. The biggest lever now is what an L2 can do that no other environment can.

What L1 Becomes in This Model

The blog describes it as a truly permissionless & maximally resilient global hub for settlement, shared state, liquidity, & DeFi.

This is not a diminished role but a more demanding one, because it requires L1 to be something that no individual L2 can replace: the neutral, trust-minimised foundation that every chain in the ecosystem can anchor to. The depth of DeFi liquidity on L1, the security guarantees, the decentralisation, these properties are what make it worth being an L2 on Ethereum rather than a standalone chain. The EF is explicit that a stronger L1 provides a better foundation for every L2 sitting above it.

ZK technology is what makes this model viable at scale. The post notes that ZK acceleration has moved faster than anticipated & that Ethereum L1 now has a credible path to scaling by several orders of magnitude while preserving decentralisation & security. Blobs currently sit at roughly 30% capacity with significant headroom to grow. The infrastructure for a much larger Ethereum is being laid & the foundation believes it can grow that infrastructure without compromising the properties that define what Ethereum is.