Solana Expands 150-Millisecond Settlement Testing to Another Public Testnet

Solana’s Alpenglow upgrade has reached both of its public test networks, giving developers a chance to test applications and infrastructure before the changes are introduced to the mainnet.

The upgrade aims to dramatically speed up transaction finality. Alpenglow is designed to bring finality down to around 150 milliseconds, compared with approximately 12.8 seconds under Solana’s existing system. Finality is the stage at which a transaction is considered irreversible.

If the target is achieved, the shorter finality window could help exchanges process deposits faster and enable payment applications to confirm completed purchases to merchants with minimal delay.

Anza, the company developing Solana’s core software, announced on September 25 that Alpenglow had gone live on the devnet. The separate testnet completed its transition a day earlier, on September 24. The Solana Foundation’s upgrade tracker now lists Alpenglow as active across both public testing networks.

The devnet is primarily intended for application developers and uses tokens without real monetary value. The separate testnet provides an environment for testing validator operations and network software, including under high-demand conditions.

Alpenglow will also change the way Solana validators coordinate and reach consensus. Instead of including validator votes in blockchain blocks as transactions, validators will exchange voting information directly. The revised mechanism could allow consensus to be achieved within one or two voting rounds.

That architectural change could affect Solana’s reported transaction statistics. Validator votes will no longer be counted as transactions, which means headline transaction totals may decline even if user payments and trading activity remain at similar levels. The Solana Foundation has advised data providers to account for the change when analyzing network activity.

Blockchain indexing services will also have to handle competing candidate blocks differently. Transaction data from competing blocks must remain separate until the network determines which block is canonical. Combining data too early could result in incorrect transaction histories.

The Foundation’s guidance indicates that applications that only send transactions and read account balances should not require migration when Alpenglow is introduced.

The 150-millisecond figure remains a projected target based on simulations and has not been established under live network conditions. Actual transaction completion times may also be longer because wallets, exchanges, and other applications can add their own processing delays.

Solana has not announced a definitive mainnet activation date for Alpenglow. Anza’s current software roadmap suggests feature activations could resume on September 28, but the date is not confirmed as Alpenglow’s mainnet launch.