Solana reduces slot target to 350ms, edging toward sub-second

Solana reduced its mainnet slot target from 400 milliseconds to 350 milliseconds after a coordinated feature gate. The change used the protocol’s one-epoch delay and applied when epoch 1020 began on Aug. 21.

Telemetry recorded a slot-weighted mean spacing of 365.4 ms across 431,505 timed slots in epoch 1021. The network logged 331 skipped slots in that epoch, a 0.077% skip rate. By comparison, a pre-change sample from epoch 1015 showed a 420.7 ms mean spacing and 1,890 skipped slots, a 0.438% skip rate.

The feature gate flipped at slot 440,208,000, the first slot of epoch 1019. Under SIMD-0525, a one-epoch delay allows validators to apply the new timing and the reduced shred limits together.

A slot is the time window during which a designated validator may produce a block. An epoch is fixed at 432,000 slots. Solana assigns four slots per leader; shortening the slot target reduces the four-slot leader window from 1.6 seconds at 400 ms to 1.4 seconds at 350 ms.

The protocol lowers per-slot compute, account-write, vote, data and shred limits as the slot target shortens. Blocks arrive more frequently with smaller per-slot budgets, and per-second processing capacity remains similar.

Solana’s roadmap lists successive stages to 300 ms, 250 ms and 200 ms. Each stage is deployed via a feature gate paired with a one-epoch delay. The roadmap allows pausing between stages if skipped-block rates rise. The 300 ms stage remained pending as of Aug. 26.

Brennan Watt, Anza CEO and author of SIMD, expected the 300 ms stage to take effect at epoch 1024, around Aug. 28. Network participants monitor skipped-block rates as a metric for whether the timetable advances.

Skipped slots occur when the scheduled leader does not produce a block in its slot window. Finality, the point at which blocks are considered irreversible, is measured separately from slot spacing.