Infrastructure readers see an EVM transaction get included in a block very quickly, and the next question is usually: after inclusion, who is responsible for getting the data to the settlement layer? DuskEVM’s path separates the responsibilities. The sequencer handles the early ordering and inclusion; the L2 blocks provide execution traces; the batcher publishes the data to DuskDS; and then state commitments and fault proofs are connected to stronger verifiability. Speed only describes one part.

If you only display “included,” readers won’t be able to see whether data publication has completed, nor how the subsequent state is committed and proven. The product can provide instant feedback, but it must name it as inclusion or execution progress; when the copy starts talking about final settlement, it needs to add evidence for the DuskDS data, state commitments, and the proof conditions. Responsibilities across different layers can’t be replaced with a single time number.

This layering also makes troubleshooting easier. Waiting during the sequencer stage is completely different from having data already published but proofs not completed yet—these are different kinds of failures or waiting. Monitoring, user prompts, and settlement decisions should be handled separately. A fast path doesn’t mean skipping underlying responsibilities, and knowing the familiar EVM entry point can’t hide these state differences.

To understand DuskEVM’s speed, first ask where the data went and who takes responsibility for the next step, and only then discuss finality. By showing these stages, infrastructure teams can distinguish between ordering issues, data availability issues, and settlement proof issues. The “speed” users see is a UX metric, but the final state still has to wait for the corresponding evidence. This is also why inclusion speed cannot replace data publication and final-state proof. When you spell out each layer’s responsibility, finality won’t become an abstract “speed” phrase. Therefore, infrastructure documentation should list ordering, publication, commitments, and proofs separately, instead of using one “fast” word to stand in for the full chain of states.

@Dusk $DUSK #dusk