Vitalik Buterin shares roadmap to Ethereum becoming leaner and more scalable

The three reported pillars

Multiple outlets converge on the same handful of technical levers. First, the chain itself is meant to shrink — an "Extremely Lean" framing in which ZK proofs carry most of the verification burden and persistent state stays minimal. Second, native STARKs are positioned as a base-layer primitive rather than something rollups bring on themselves. Third, the explicit priority list reads as a stack: quantum resistance, scalability, privacy — with the implication that other concerns are deferred, not dropped.

That ordering matters. It tells you what the protocol team considers load-bearing for the next cycle and what is being treated as a rollup's job to solve. In practice, a near-stateless L1 changes what your L2 can safely assume about read access, archive nodes, and historical reconstruction.

What this shifts in the L2 design matrix

If STARKs become a native L1 primitive, rollups already standardized on STARK-based proving get a cheaper integration path, and alternative proof systems face quiet pressure to either bridge compatibility or accept a thinner settlement story. Conversely, a quantum-resistant L1 forces a rethink of signature schemes, validator hardware assumptions, and the long-lived contracts L2s lean on for fraud or dispute resolution. None of this is new engineering — but the incentive to do it moves.

The privacy piece is the one worth watching closely. The reported emphasis on privacy at the base layer is the same migration pattern we saw with EIP-4844 blobs: functionality that L2s ship today tends to drift down into consensus tomorrow. Teams building privacy-flavored or general-purpose L2s should be asking which of their current L1 dependencies become optional first.

What to actually do about it

The roadmap is directional, not a spec. The concrete near-term actions for production teams are conservative: keep state bloat assumptions tight, keep your proving stack portable across proof systems, and avoid tight coupling to a single L1 signature scheme. Watch the follow-up posts for specific EIPs and a clearer answer on whether "near-zero state" gets pinned to a particular aggregation strategy or remains a research target.