# Solana changes its consensus on October 21 with Alpenglow

> The network debuts Votor and 150-millisecond finality versus the current 12.8 seconds, with validators required to update their client.
> By José Maldonado · Blockchain · 2026-10-20

On October 21 Solana carries out a consensus change that had been two years on the calendar. The network replaces the well-known TowerBFT consensus with [Votor](https://observatorioblockchain.com/blockchain/alpenglow-podria-llevar-a-solana-a-la-velocidad-de-nasdaq/), a mechanism that promises block finality in 150 milliseconds versus the current 12.8 seconds, and Anza has set activation for around 14:50 UTC. The validator vote is already closed and the migration checklist was published on October 13.

In this way, Alpenglow rewrites how validators reach agreement, and from that comes the figure the project sells: near-instant confirmation that brings Solana closer to the behavior of a centralized ledger. For anyone building payments or a market, the difference between waiting thirteen seconds and waiting tenths of a second is not convenience, it is what kind of product can be written on top.

Context matters. Solana reaches this change without having appeared even once in this outlet's coverage over the past two weeks, which has focused on Ethereum, its layer twos and the blobs debate. The network that boasts of speed has moved in silence while public discussion centered on the scalability of its direct competitor.

## What exactly changes in the consensus?

Votor introduces two confirmation paths that run in parallel. A fast one, which closes a block with a supermajority of votes without going through the classic rounds, and a slow one, which maintains security when the network degrades or when some validators do not respond. The design assumes that most of the time the fast path will be used and that the slow one exists for the worst days. Anza has already published the date, the time and the validator checklist.

Recall that until now, a block could be considered confirmed in around 400 milliseconds, but not finalized until more than a dozen seconds had passed, a distinction that confuses users and complicates bridges between chains. With fast finality, a centralized exchange can credit a deposit without waiting for block depth and a bridge can release funds without a reversion window.

The change has been tested in several testnet rounds and the foundation published the schedule with the specific time, something unusual in this type of migration. Validators had to update the client before the 21st and review the priority fee configuration, which remains unchanged.

## Payment and trading systems, the big winners

The first beneficiaries are products that live off confirmation time: point-of-sale payment systems, prediction markets, high-frequency trading and any application that needs to read the result of a transaction without waiting. Also bridges, which will be able to reduce the collateral locked during the journey between chains.

The cost falls on infrastructure operators. A new consensus changes voting patterns, message size and the way they propagate, and that translates into bandwidth requirements that not all small validators meet. [Anza](https://www.anza.xyz/) itself published a checklist and warned that nodes with tight hardware may fall outside the voting window in the first days.

There is a less-discussed side effect. When finality drops to tenths of a second, the advantage of layer twos over Solana shrinks in the commercial pitch, and several teams have built their value proposition precisely on that difference. The comparison between a chain with fast finality and a [rollup](https://observatorioblockchain.com/criptomonedas/bitcoin/conoce-los-zk-rollups-que-estan-llegando-para-escalar-bitcoin/) that inherits the security of [Ethereum](https://observatorioblockchain.com/en/ethereum/ethereum-blob-fees-400-l2-limites/) will no longer be framed in terms of speed and will be framed in terms of cost and liquidity fragmentation.

## Speed is not everything

Despite the major changes, Alpenglow does not touch the fee model, nor the state of the validator client, nor the fact that the network has suffered outages in the past. Finality speed does not solve the problem of validator concentration among cloud providers either, which remains the most repeated criticism and the one that has moved the least with each update.

It is worth looking at Ethereum's calendar to understand the asymmetry. This month the Glamsterdam analysis, the blob increase and the debate over burning staking rewards have coincided. Ethereum advances by layers and by committees, with years of public discussion for each change. Solana advances by client versions and a date on the calendar. Neither approach is free.

## The day the network stakes its reputation

The real test comes in the first 72 hours, when traffic normalizes and validators that did not update begin to miss votes. A consensus change that works on testnet and stumbles on mainnet leaves a scar that takes years to close, and Solana already has experience with that.

If Votor holds, the scalability conversation shifts axis: it will no longer be how many transactions per second and will become how long it takes for a transaction to become irreversible. It is a duller question and a much more useful one for anyone who has to build a product on top.
