Ethereum node operators running their own infrastructure have long dealt with a two-daemon setup. That may change. Vitalik Buterin posted on X calling for the community to reconsider the beacon and execution client separation, stating the current approach adds unnecessary friction for those seeking self-sovereign access to the network.

“We should be open to revisiting the whole beacon/execution client separation,” Buterin wrote on X. “Running two daemons and getting them to talk to each other is far more difficult than running one daemon.”

His concern centers on user experience. The goal, he said, is to make running your own Ethereum node practical enough that it becomes the default way people interact with the chain.

One Binary. One Process. No JWT Headaches.

The timing of Buterin’s post was not coincidental. The Nimbus unified node pull request #3646 had already been merged into the main Nimbus repository on October 24, 2025. Buterin pointed to it directly in his post.

The merge, described by contributor arnetheduck on GitHub, brings execution and consensus layers into a single binary running in one process. No cross-client JWT configuration. No inter-process communication failures. No timing issues between the two daemons.

“This preview brings Nimbus full circle to where it started all these years ago,” arnetheduck noted in the pull request, “and allows running Ethereum in a single node, both as a wallet/web3 backend and as a validator.”

The shared database alone cuts storage demands compared to running separate clients. Around 500GB is needed for mainnet. Performance stays strong, and the setup command drops to a single line.

What Buterin Actually Proposed

Short-term, he suggested a standardized wrapper that lets users install Docker containers of any client and handle the inter-process communication automatically. That’s a bridge, not a solution, and he knows it.

Longer term, Buterin said the architecture itself needs revisiting once the lean consensus work under the @leanethereum project matures. That project is working on a stripped-down consensus layer design. He referenced it directly as the prerequisite for any deeper structural change.

The Nimbus unified approach already demonstrates the concept at a working prototype level. Beacon and execution chains each run in separate threads within the same process, sharing one data directory and common services. Block history going back up to 5 months is sourced from the consensus network directly.

Node Simplicity Is the Real Bottleneck

Buterin’s post lands in a broader conversation about Ethereum decentralization. Running a node today requires technical knowledge, careful configuration, and tolerance for setup failures. The unified model cuts most of that out.

The Nimbus implementation still has rough edges. Cross-thread communication currently routes through a local JSON instance rather than direct object passing, which arnetheduck flagged as inefficient. Discovery runs separately for both chains, giving the node two distinct identities. Memory usage has room for optimization.

Still, the prototype works. Syncing a testnet like Hoodi from scratch takes just over a day via P2P, without checkpoint sync. Mainnet sync pairs checkpoint sync with P2P execution chain sync. Snapshots of the mainnet execution database are available at eth1-db.nimbus.team for faster setup.

Portal and light client support are next on the roadmap, according to the GitHub notes. That would let the unified node serve a wide range of use cases, from full validators down to lightweight web3 backends.