Solana activates Alpenglow with 150ms finality target

Solana activated Alpenglow on its mainnet on Oct. 21, replacing TowerBFT and introducing a new way for validators to agree on blocks. Anza says the upgrade is designed to reduce finality—the point when a transaction can no longer be reversed—from about 12.8 seconds to roughly 150 milliseconds. That is a protocol target, however, and early live-network reporting shows why measured performance will matter as much as the design.

In Anza’s launch announcement, the Solana engineering firm said validators retired TowerBFT at slot 459,653,000, at approximately 14:00 UTC. TowerBFT had been Solana’s consensus protocol since the network launched in March 2020. The change affects how validators establish agreement and finality; transaction execution, programs and fees remain as before, according to the Solana Foundation’s technical account.

How does Alpenglow change the path to finality?

Alpenglow replaces votes recorded as on-chain transactions with votes sent directly between validators and combined into certificates. Under TowerBFT, validators’ votes accumulated over 32 slots before a block became final; the new system can finalize a block after one or two voting rounds. The design therefore removes an intermediate layer of vote transactions as well as shortening the path to a settlement guarantee.

The distinction between confirmation and finality matters. Solana users often relied on a “confirmed” state that arrived within seconds, while the stronger guarantee that a block could not be dropped took about 12.8 seconds under TowerBFT. The upgrade aims to bring that stronger guarantee close to the earlier confirmation time. Anza describes the result as a reduction to about 150 milliseconds, but that figure should be read as the intended performance, not a guarantee that every transaction will settle at that speed in every network condition.

Does the 150ms target match early mainnet readings?

Not yet in every reported measure. A Cointelegraph report on launch day said Solana’s Alpenglow dashboard showed 6.6 seconds of transaction finality at the time of publication. That snapshot is far above the 150-millisecond target, but it does not by itself establish sustained network performance or explain the difference between the dashboard reading and Anza’s stated figure. Follow-up measurements will show whether the new protocol reaches its target under ordinary traffic and across different validators.

There is also a change in how activity appears in block data. Since validator votes no longer count as transactions, reported transaction totals and TPS can fall without a corresponding drop in user activity. Analysts and dashboard operators will need to adjust comparisons that previously included vote transactions. Solana’s technical documentation also says applications that read block streams must account for changes to how block information is exposed.

What should users and operators watch next?

For users, the practical question is whether wallets, exchanges and payment services can rely on finalized transactions sooner without changing how they submit them. For developers and data providers, the near-term work is adapting indexers and metrics to the new block structure and checking live finality readings against the prior 12.8-second baseline.

Alpenglow’s launch is the consensus change; it does not complete Solana’s broader roadmap. Anza says Rotor, a planned redesign of how block data spreads across the network, is a later improvement with no committed release date. The measures to watch are therefore concrete: mainnet finality over time, whether transaction-count dashboards are recalibrated, and whether later networking changes reduce delays beyond the consensus layer.

Source material

• Anza’s launch announcement — prnewswire.com
• Cointelegraph report on launch day — cointelegraph.com