Solana cuts mainnet slot time to 350 milliseconds in first step toward 200ms goal

Solana (SOL) saw a successful mainnet upgrade on Friday morning that cuts the network's target slot time to 350 milliseconds (ms), down 50 ms from its former target of 400 ms. A slot is the short period of time in which a designated validator can produce a block of transactions of Solana's blockchain. 

The change was the first step under SIMD-0525, a Solana upgrade proposal merged on May 14 that outlines a path to eventually cutting the target slot time to just 200 ms, half its longtime value. 

"Shorter slots reduce confirmation and finalization latency for users," the proposal explains. "Any consensus or commitment threshold measured in slots takes less wall-clock time as slot time decreases."

The change also shortens Solana's epochs, which each contain 432,000 slots, meaning an epoch that used to take about 48 hours should now only take about 42 hours. Validators, which remain in charge for four slots at a time, will now change turns every 1.4 seconds, down from 1.6 seconds. 

Solana developers plan to make three further 50 ms cuts until the slot time reaches its final target of 200 ms. Each stage will be switched on separately, and the rollout can be paused if too many validators fail to produce blocks in their assigned slots, according to the Solana Foundation's upgrade overview. 

In a point-in-time check conducted for this article, a 1,000-slot period shortly before the change took 415 seconds, compared with 368 seconds for a 1,000-slot period later in epoch 1020.

While shorter slots reduce transaction latency, the change does not mean Solana's network can automatically process more work each second, as validators handle slots more often but each slot carries less work. Shorter slots also limit how long one validator can delay or rearrange transactions before the next validator's turn, the proposal's authors argue. 

The change comes as non-voting transactions on the network, meaning transactions that have a purpose beyond supporting the network, have seen a notable rise since the start of the year according to The Block's data. 

Solana, co-founded by Anatoly Yakovenko, has in the past separately increased the amount of work that can fit into each block. The Block reported in July 2025 that developers had proposed raising the network's compute limit, which limits the amount of work the network can do to process transactions, to 100 million compute units. The change was merged and adopted into the main network on July 29. 

The network has also recently added another major validator client, written in a different language than the Rust-based Agave (and its popular fork, Jito-Agave). Jump Crypto’s Firedancer client, written in the C programming language, went live in December, improving client diversity and network resilience. 

Though slot times are now on their way down to 200 ms, full finality, or how long it takes the network to treat a block as irreversible, still takes much longer: Solana’s Alpenglow overhaul, still in development, aims to eventually cut finality to about 150 ms from roughly 12.8 seconds today.

The next planned mainnet target for slot times is 300 ms, though Solana developers have not yet set a target activation date for the change. The developers plan to monitor the network's performance with its current 350 ms slots before proceeding to cut the target time further.