Solana activates Alpenglow, targeting 150ms finality

Solana activated Alpenglow on mainnet on Oct. 21, replacing TowerBFT with a new consensus system that targets transaction finality in about 150 milliseconds, down from roughly 12.8 seconds. The Solana Foundation’s upgrade page says the change took effect at the start of epoch 1064, at approximately 13:36 UTC. For users, the shorter target could mean less waiting for payments and deposits to be treated as settled; the change is to how validators agree, not how wallets send transactions.

The switch followed a period of testing on Solana’s public development networks. The network’s previous approach, TowerBFT, built validator approvals across successive blocks. Alpenglow’s first phase, Votor, replaces that process with direct validator votes bundled into certificates. The Block reported that votes from 446 of 654 participating validators certified the starting block, and that the handover itself took about 900 milliseconds. That handover time is separate from the protocol’s 150-millisecond finality target.

What changes for someone sending SOL or using an app?

For ordinary users, the Foundation says no wallet or application migration is required: signing, sending and approving transactions work as before, and fees and transaction execution are unchanged. The potential gain is faster final settlement, which could shorten the time a merchant or exchange waits before treating a payment or deposit as irreversible.

That is a protocol target, not a promise that every transaction will settle in exactly 150 milliseconds under every network condition. Users should distinguish finality from block production: Solana separately reduced slot times from 400 milliseconds to 200 milliseconds, and the Foundation says that change completed in October. Faster finality and faster block timing are related to network experience, but they describe different things.

What does Alpenglow give up or leave for later?

Alpenglow removes validator vote transactions from blocks. That frees block space from those votes, but it also means dashboards or services that counted all on-chain transactions may report lower totals even if user activity has not fallen. The Foundation says the SVM, transaction formats and fee mechanics remain untouched, while developers who stream or parse block data may need to adjust for changes such as multiple candidate blocks per slot and the new `bank_id` field.

The rollout is also incomplete in one important respect. Votor, which changes voting and finality, is live; Rotor, intended to replace Turbine’s block-data propagation method, is planned for a later release and has no schedule yet. So Alpenglow delivers the consensus change now, while a broader redesign of how data travels across the network remains ahead.

Which signals will show whether the upgrade is working?

The first signal is observed mainnet finality over time, compared with the roughly 12.8-second TowerBFT baseline and the new 150-millisecond target. Network conditions matter, so sustained measurements will say more than the target alone. A second is whether wallets, exchanges and payment services begin using the shorter settlement window in practice.

For infrastructure teams, the near-term test is whether indexers and analytics providers adapt to vote transactions leaving blocks without mistaking the resulting drop in reported totals for a drop in user demand. Longer term, developers and validators will be watching for a Rotor release date and for evidence that faster finality holds as Solana’s network load changes.

Sources and documents

• Solana Foundation’s upgrade page — solana.com