Solana's Alpenglow Upgrade Reaches Devnet, Targeting 150-Millisecond Transaction Finality for On-Chain Payments
Key Takeaways
- •Solana's Alpenglow upgrade aims to reduce transaction finality times from approximately 12.8 seconds to around 150 milliseconds.
- •The upgrade has reached Solana's public devnet following its testnet deployment, while a mainnet launch date remains unscheduled.
- •Alpenglow's Votor system replaces in-block vote transactions with direct validator communication and cryptographic certificates, enabling finality after one or two voting rounds.
- •The upgrade modifies only the consensus layer, leaving the Solana Virtual Machine, transaction formats, fees, and programs unchanged.
- •Faster finality could let merchants and exchanges treat payments as irreversible far sooner, bringing on-chain settlement closer to traditional electronic payment speeds.

Solana is testing a major overhaul of its consensus system that could reduce the time required for a transaction to become final from about 12.8 seconds to roughly 150 milliseconds, bringing blockchain settlement closer to the speed users expect from conventional digital payments.
The upgrade, known as Alpenglow, has moved to Solana's public devnet after reaching testnet, giving developers an opportunity to test how applications handle the faster finality before the upgrade reaches the main network. No mainnet launch date has been scheduled. In the meantime, the milestones to watch are developer testing results from the devnet environment and any subsequent announcement of a mainnet timeline.
The biggest change is not to how Solana executes transactions, but to how validators confirm that those transactions are final.
Under Solana's current TowerBFT system, validators vote by sending vote transactions that are included in blocks. Finality takes about 12.8 seconds because enough validator votes must accumulate across multiple slots.
Alpenglow's Votor system removes those votes from blocks. Validators instead communicate their votes directly and aggregate them into cryptographic certificates showing that sufficient stake has agreed on a block. A block can then become final after one or two rounds of voting.
For payments, that difference could be significant. A merchant accepting a Solana-based payment could potentially treat a transaction as irreversible in about 150 milliseconds, rather than waiting several seconds for stronger settlement certainty. An exchange could similarly release deposited funds much sooner once a transaction reaches finality, while payment applications could provide users with near-immediate confirmation that money has arrived.
That changes the practical meaning of "on-chain settlement." The transaction is not merely visible on the blockchain; the network's validators have reached the level of agreement required to make reversing it extremely difficult under the protocol's consensus assumptions. Until that point, a transaction remains provisional from the network's perspective, which is why the time it takes to reach finality shapes what payment workflows are practical to run on a given blockchain. The upgrade therefore targets one of the biggest friction points in blockchain payments: the gap between a transaction appearing to have happened and the network reaching final agreement that it has happened.
Importantly, Alpenglow does not change Solana's execution layer. The Solana Virtual Machine, transaction formats, fees, and programs remain unchanged. The upgrade is fundamentally about how the network reaches consensus and finality.
If the roughly 150-millisecond target holds under real-world mainnet conditions, the effect could extend beyond faster wallet confirmations. Payment processors, crypto exchanges, and financial applications could potentially build workflows around finality that is fast enough to approach the experience of traditional electronic payments, while retaining on-chain validation underneath. That condition is precisely what the devnet phase is meant to stress-test before any mainnet deployment.
The testnet and devnet deployments are therefore less about making Solana transactions faster to submit, and more about testing whether distributed validators can reach irreversible agreement fast enough for blockchain settlement to become effectively real-time.
This article was originally published by BitcoinKE: https://bitcoinke.io/2026/09/the-solana-alpenglow-upgrade/