Solana’s Alpenglow upgrade has moved onto both public test networks, giving developers an opportunity to test their applications before the changes are introduced to the live blockchain.
The upgrade is intended to dramatically reduce Solana’s transaction finality time. Alpenglow is targeting finality of about 150 milliseconds, compared with roughly 12.8 seconds under the network’s current architecture. Finality marks the stage at which a transaction can no longer be reversed.
A successful rollout could reduce the time exchanges take to process deposits and allow payment applications to confirm completed transactions to merchants almost immediately.
Anza, the developer of Solana’s core software, said on September 25 that Alpenglow had been activated on the devnet. The move followed the upgrade’s transition to the separate testnet on September 24. The Solana Foundation’s upgrade tracker now shows the protocol running on both public testing environments.
The devnet is designed for application developers and uses test tokens that carry no real monetary value. Solana’s separate testnet is primarily intended for testing validator infrastructure and network software under demanding conditions.
Alpenglow will also modify Solana’s consensus process. Validators will send votes directly to one another rather than recording their votes as transactions inside blocks. The revised process is expected to allow validators to reach agreement within one or two voting rounds.
The architectural change could alter how Solana’s network activity is measured. Validator votes will no longer count toward transaction totals, potentially lowering reported transaction numbers even when user activity remains unchanged. The Solana Foundation has asked data providers to account for this difference when comparing network statistics.
Blockchain indexing services will face another adjustment. They will need to keep transactions from competing candidate blocks separate until the network selects the canonical block. Combining data from competing versions could lead to inaccurate transaction records.
According to the Foundation’s documentation, applications that only submit transactions and read account balances should not require migration for Alpenglow.
The 150-millisecond figure is still a projected performance target based on simulations and has not yet been demonstrated under live network conditions. In practice, wallets, exchanges, and other applications may add their own processing time before users see a completed transaction.
Solana has not provided a confirmed date for Alpenglow’s mainnet activation. Anza’s software roadmap indicates that feature activations may resume on September 28, but the date has not been designated as the launch date for the upgrade.
