Home » Blockchain » Solana Cuts Slot Time to 350ms: Plan and Early Data

Solana Cuts Slot Time to 350ms: Plan and Early Data

Solana’s first slot time cut took effect on August 21, 2026. Early data shows slots running above the 350ms target, with no change to throughput or finality.

Abstract blue light-speed streaks illustrating the Solana slot time reduction and faster blockchain confirmations

On August 21, 2026, Solana’s target slot time dropped from 400 milliseconds to 350 milliseconds on mainnet. It was the first slot time reduction since the network launched. The change is the first of four 50 millisecond steps defined in SIMD-0525, a proposal that The Block reports was merged on May 14. The proposal’s stated aim is to reduce confirmation and finalization latency for users.

How the change went live

Solana Compass reports that the feature gate activated at slot 440,208,000 in epoch 1019. After a mandatory one-epoch delay, validators began epoch 1020 on August 21 with the new clock. Each later step has its own feature gate and needs a separate supermajority validator vote, and The Block reports the rollout can pause if too many validators miss their assigned slots. Jacob Creech of the Solana Foundation called the change a new era and named 300 milliseconds as the next stop, per Solana Compass.

To keep the workload per second roughly constant, the per-block limits shrink with the slot time. Solana Compass reports the compute limit per block falling from 100 million compute units to 87.5 million, which holds capacity near 250 million compute units per second. Per-slot allowances for account writes, vote budgets, and shreds also shrink, though Solana Compass notes that some of these figures come from third-party outlets rather than primary documentation. The change is not meant to raise throughput: each shorter slot carries less work, per The Block’s summary of SIMD-0525.

Leader turns get shorter too. Validators lead for four slots in a row, so at 400 milliseconds that turn lasted 1.6 seconds and at 350 milliseconds it lasts 1.4 seconds. The proposal’s authors say shorter turns reduce how long one validator can delay or reorder transactions before the next validator takes over. That limits one source of maximal extractable value (MEV), but other sources remain.

What the first measurements showed

The Block checked 1,000 slots shortly before the change and found they took 415 seconds. A 1,000-slot stretch later in epoch 1020 took 368 seconds. At the 400 millisecond target, 1,000 slots would take 400 seconds, so even the pre-change sample ran slow. These are single samples of 1,000 slots each, not network averages or benchmarks, and they measure how fast blocks were produced, not how quickly a transaction confirms.

Solana Compass later compared full epochs. Epochs running at the 350 millisecond target averaged slots of 365 to 367 milliseconds and lasted 43.9 to 44.0 hours. The nominal figure for an epoch of 432,000 slots at 350 milliseconds is 42 hours, down from about 48 hours. Solana Compass found every step in the plan overshot its target by 15 to 19 milliseconds and does not explain why. Treat 350 as a target, not an observed average.

Who Benefits and How

For an ordinary Solana user, the gain is small and indirect: blocks arrive more often, so inclusion can come sooner. The sources we reviewed did not publish a user-level confirmation benchmark, so there is no measured speedup to quote. The change is live on mainnet now, and users don’t need to do anything. Developers are the ones with a to-do item: apps should not hardcode a 400 millisecond slot, and Solana Compass advises reading slot duration from the network.

What 350ms does and doesn’t tell you

Slot time is not finality. The Block and Solana Compass both put full finality under the current TowerBFT voting system at about 12.8 seconds, and finality is the target of a separate upgrade, Alpenglow, which has not activated on mainnet. No source we opened gives user fee data or validator hardware data for this step, so claims of “no extra cost” are unconfirmed. A KuCoin analysis, which is a marketing blog and not independently verified, says faster slots raise vote traffic and give marginal hardware less time per slot, so such hardware may see more skipped slots.

The plan itself kept to schedule. Solana Compass lists the later mainnet steps as 300 milliseconds on August 28 (epoch 1024), 250 milliseconds on September 18 (epoch 1037), and 200 milliseconds on October 9 (epoch 1053). Our separate 200 millisecond post covers the final step.

Common questions

Does a 350ms slot mean my transaction is final in 350ms?

No. A slot is the time allotted to produce a block. Full finality on Solana currently takes about 12.8 seconds, per The Block.

Did transactions confirm faster after the change?

Blocks arrive more often, so inclusion can come sooner. The sources we reviewed did not publish a user-level confirmation benchmark, and Solana Compass measured slot averages of 365 to 367 milliseconds without explaining the gap.

Did the upgrade increase Solana’s capacity?

No. Per-block limits were scaled down so total capacity per second stayed near its earlier level.

Sources

Related Reading