Solana cuts slot time to 250ms in third SIMD-0525 stage
Solana’s target slot time dropped to 250 milliseconds on Friday, 18 September 2026, a nearly 17% increase in how often the network produces blocks, Decrypt reported. A slot is the window assigned to a single validator to build and publish a block. The reduction is the third of four stages under SIMD-0525, the network’s proposal for shortening slot times.
Core developers at Anza said the 250ms feature would take effect one epoch after activation, and it went live at the epoch 1037 boundary around 05:01 UTC on 18 September, according to U.today. Anza’s own post framed the step as “1 more step to 200ms.” Crypto.news reported that the network now targets four slots per second, up from roughly 3.3 at the previous 300ms setting.
Also read: Revolut Hit With 6,000 XMR Ransom Demand as Hackers Threaten to Sell Customer Data
Key facts

- Solana’s target slot time fell from 300ms to 250ms, its third cut since August, when the rollout began at 400ms and moved to 350ms, then 300ms.
- Each validator’s four-slot leader window shrank from 1.2 seconds to one second.
- An expected Solana epoch compressed from roughly 36 hours to about 30 hours; epochs remain fixed at 432,000 slots.
- The change does not raise Solana’s overall processing ceiling: at the 60 million compute unit baseline, the per-slot limit sits at 37.5 million compute units.
- A final stage to 200ms has no mainnet activation date and depends on block-skip rates staying in check.
Fresher data, not more capacity
Decrypt stressed that the shorter window does not translate into a matching capacity gain, because each slot now carries proportionally less computation and data. Crypto.news put the same point in resource terms, noting that SIMD-0525 reduces limits in proportion to slot duration, and that a planned 200ms stage would lower the per-slot limit to 30 million compute units.
The applications most affected are ones that depend on fresh onchain data: oracles that bring real-world prices onto the chain, and automated market makers that price assets algorithmically rather than matching buyers and sellers directly. Both are named in the proposal because their operations can turn on the age of that data, and a stale feed can push a trade to the wrong price.
Also read: Internet Computer Tops Web3 Transactions With 134.5M in 24 Hours as ICP Stays Below $3
Infrastructure providers also take on more individual blocks to process and store even though the wall-clock ceiling is broadly unchanged, according to Crypto.news. Blockhashes expire sooner in real time, leaving less room for transactions that involve offline signing or delayed human approval. Applications that estimate elapsed time by multiplying slot numbers by a fixed duration may need to account for the faster clock.
Neighbouring upgrades
Slot timing is one piece of the wider Agave 4.2 rollout. Crypto.news reported that the client release also brought lower onchain storage rent, larger transactions and the path toward 200ms slots. Solana’s Transaction V1 raises the maximum serialized transaction size from 1,232 bytes to 4,096 bytes, supporting data-heavy operations such as zero-knowledge proofs and complex multisignature instructions in a single transaction. It is optional, and legacy and version zero transactions remain supported.
Separately, Alpenglow, Solana’s consensus redesign, is intended to replace TowerBFT with a voting system called Votor and to remove onchain vote transactions from core consensus. Crypto.news said Alpenglow entered community validator testing earlier in 2026 and targets roughly 150ms finality, with mainnet deployment tied to Agave 4.3 rather than the feature gates used for SIMD-0525.
Why it matters
Solana’s staged reductions narrow the gap between blockchain settlement and the near-instant confirmation users expect from a stock exchange or a card network, which is the practical benefit the proposal cites. For validators, the reward cycle runs on the epoch clock, so a shorter epoch means faster payouts for operating under the new setting. For traders and exchanges, more frequent block production means wallet balances and transaction status refresh sooner. The trade-off is that infrastructure, tooling and timing assumptions built around longer slots must be revisited rather than assumed to carry over.
What to watch
The remaining question is whether developers activate the final 200ms feature gate. Under SIMD-0525, that step is gated on block-skip rates, or how often a validator misses its turn, staying within acceptable bounds, and no mainnet date has been set. At 200ms the network would produce five slots per second, run roughly 800ms leader windows and shorten epochs to about 24 hours.
Sources: Decrypt, Crypto.news, U.today
