Ethereum co‑founder Vitalik Buterin has drawn a sharp line between scaling and latency in a new post, arguing that while PeerDAS and zero‑knowledge proofs can expand throughput thousands of times, reducing latency faces hard physical and economic constraints.
Buterin pointed to the speed of light as the ultimate barrier. Even with improvements, nodes must remain viable in rural areas and outside data centers, while still supporting censorship‑resistance and anonymity. He warned that if staking outside hubs like New York City became less profitable, decentralization would erode as validators concentrated in single regions.
Scaling Gains vs Latency Realities
Buterin sees room for progress on latency, but only to a point. Peer‑to‑peer upgrades such as erasure coding could cut message propagation times, and reducing node counts per slot from 30,000 to 512 could eliminate aggregation steps. Together, these changes might deliver a 3–6x improvement, bringing block latency down to two to four seconds. Faster, but not instant.
He stressed that Ethereum isn’t meant to act like a video game server. He described it as the “world heartbeat,” a base layer that prioritizes decentralization and security over raw speed. Applications needing faster responses will rely on offchain components and Layer‑2 networks. L2s already serve roles in virtual machine customization and scaling, and will remain essential even as Ethereum itself grows.
Looking ahead, Buterin tied the discussion to artificial intelligence. If AI systems process information 1000x faster than humans, their effective “speed of light” shrinks to city‑scale. That could drive demand for hyper‑localized “city chains” or even building‑specific blockchains, all running as L2s. At the same time, he dismissed the idea of staking nodes on Mars as economically unworkable, framing Ethereum as a network rooted on Earth, with L2s bridging both local and global needs.
